David Ness
Mind / Matter

Piled Higher and Deeper

By David Ness
Thursday, May 23, 2002

23 May 2002: There Ain't No Streets Mid-Ocean

... nor in the desert, for that matter. So maps aren't of much use there.

And there aren't any streets in my mind, either, so the question of whether a map of my mind will make any sense or be of any use has no obvious answer. Unless you happen to be a phrenologist, that is.

But few make the mistake of thinking that what (if anything) a phrenologist understands from his/her mind map has much to do with intellectual considerations. What we need to understand is whether maps of topics, or maps of ideas are like the view of the phrenologist or something more substantial.

Taking a Trip

Let's go on a (short) trip. Get your map.

Or do you need one?

From The TAO of Topic Maps' by Steve Pepper:

However interesting and worthwhile the experience of driving from A to B without a map might be in its own right, there can be no doubt that when the goal is to arrive at one's destination as quickly as possible (or at least without undue delay), some kind of a map is indispensable.

Are you convinced? I used to use maps a lot, but have found lately that I use them less and less. I get from my house down well into Florida, 1,000 miles or so away, with the instructions: Go South on I95. To make it all the way to Tampa, I additionally need Go West on I10, South on 301 and then South on I75. This at least gets me to the middle of Tampa. 1,000 miles and four instructions later. Usually I stop in Savannah on the way.

And, for getting around inside the town, these days I use a GPS. Even in my GPS I don't use the `map' mode much, happier instead to just have an arrow pointing generally the right direction. Once in a while the road `doubles back' and I end up going a bit out of the way, but even that happens less often with my GPS than with my old maps---which I would misread from time to time---so even there my delay is less without the map.

Similarly, if you are looking for a particular piece of information in a book (as opposed to enjoying the experience of reading it from cover to cover), a good index is an immense asset. The traditional back-of-book index can be likened to a carefully researched and hand-crafted map, and the task of the indexer, as Larry Bonura puts it [is] chart[ing] the topics of the document and [presenting] a concise and accurate map for readers..

Really? I know indexes. I like indexes. But Mr. Pepper that index is no map. Any more than a gazetteer is a map. Ever try and find your way to Florida with only a gazetteer? My gazetteer says Tampa Fla 31 H5; (A) B3 and that doesn't seem like a `concise and accurate map', at least for me. An `index' without the document is about as useful as a gazetteer without the map. You may get the names of the players, but you sure don't get much of a picture of the game.

A traditional index is in fact a map of the knowledge contained in a book; it lists the topics covered, by whatever name users might be expected to want to look them up, and includes salient (and only salient) references to those topics.

Not on my planet. Occasionally a particularly good Table of Contents sort of acts like a sketch or overview of the information contained in a book, but an index never does. Even for my favorite indexes in my favorite books.

And lists of ingredients tell you everything you need to make a cake. But it doesn't tell you how to make the cake.

So there's a bit of a disconnect, right at the beginning. But it's a provocative one. Let's see where we get if we toss it around a bit. But I'm afraid that if we already have such different views of the role of the index, the prospects for agreeing about the map aren't very good.

The Argument

Let's take a moment and try to sort out the observations and arguments that form the basis of the discussion here.

Here's the short form:

  1. Conceptual visualizations are an infertile field to invest in and likely to be waste of time and energy, but it is your time and energy, so that's ok but I advise you---as I advise my clients---against it.

  2. Maps are less useful that you'd think, and as other alternatives get better and better I expect their limitations to come to the fore.

  3. `Ideas' are (nearly) infinite dimensional and have few metrics that might suggest how to `lay them out'. Reducing them to two or three dimensions is not very likely to provide productive insights.

Or, taking a slightly deeper look.

First, there are a number of issues that relate to maps as a technology. In this part of the discussion I will argue that, for me at least, maps are of shrinking rather than growing importance as a representation. Most other discussions seem to assume that maps are a Good Thing more or less in and of themselves. While I might have believed that once, I don't any longer.

Second, a separate question is about the use of maps to represent things in a metaphorical way. It is, I will argue, one thing to use maps as an analogue for physical space, but quite another to use the kind of visualizations that maps make possible to represent things in other domains.

Third, taking the above points into account, I argue that the prospects for conceptual visualization technologies are very much more limited than might appear to be the case on the surface. These technologies are superficially attractive, but offer little prospect for payoff.

Fourth, there is one important demur. There is no way to respond to Well, I find it useful as an argument. Anybody can find (just about) anything useful. Such testimony is of little use to others unless it can be made available by following on with some indication of how the technology in question is useful in the particular situation.

The Map

First, let's take a look at maps themselves. There are at least two categories of maps:

  • analogue; and
  • metaphorical

To understand this difference and its significance, we need to look at, and understand, the role of each. If we aren't careful, sloppy thinking will cause us to misunderstand the metaphor and that can lead to some absurd and distractive conclusions.

As Analogue

The analogue map has been with us since time immemorial. Presumably our cave-dwelling ancestors may have drawn maps to show where animals could be found. Certainly by the middle ages maps were widely used to `show' (often hopelessly inaccurate) pictures of the relationship between places.

Notice that maps are generally a two dimensional representation of what has been (until very recently) largely a two dimensional world. In other words the representation of locations on a map bears a direct analogue to their location in the real world that the map represents. We take some of the important characteristics of the reality that concerns us, and then we condense it by scaling to a comprehensible and useful size.

Because the real world isn't actually two dimensional, some compromises do have to be made. We're probably lucky that there aren't a lot of people that live near the poles, as we'd have to get used to some rather different kinds of maps if we had to do a lot of navigation up (or down) there.

Fortunately for us, regular maps describe the places that we need them rather well. So we won't spend much time or effort considering them there.

As Metaphor

There are other kinds of maps. It has become commonplace to use map also as a metaphor. We talk casually about a map of thoughts or a map of ideas.

While this usage may serve some purpose, it should be clear at the outset that the use of the word map in these cases is metaphorical. We are used to maps that analogue physical proximity. A point which is closer to point A than it is to point B on the map represents a place in the real world which is closer to the place represented by A than to that represented by B. So the map makes sense. And it is consistent with our real world. Further, and perhaps even more important, conclusions that we validly draw from the map will guide us, rather than mislead us, in our reality.

Whether there is any metric that measures the closeness of ideas remains to be seen, but I am highly skeptical. And as a result I am highly skeptical of whether most of us are likely to learn much from any of these maps where we map some ephemeral world.


For it to make sense to us, most spaces need some metric to allow us to form a picture of them. As we have discussed above, physical distance is a convenient metric that lets us lay out maps of `real' two-dimensional spaces. We know a lot about how to measure distance, and given current conventional technology we can do quite a good job measuring with reasonable accuracy over a range that might run from fractions of an inch to thousands of miles. And with more sophisticated equipment we can extend this range of accuracy a great deal further.

There are some circumstances where we don't have a metric, but rather some simple form of connectivity which can be neatly represented with relevance to some particular problem. Perhaps a fair example of this kind of map is a metro map. While representations of metro stations obviously have some locality that parallels the physical placement of the stations in reality, this characteristic is often purposefully obscured with such maps. The objective of the maps is to demonstrate the connections between the stations, not to do more than hint at their actual locations.

Physical Space

We generally live in a world that might as well be two dimensional. The major exceptions to this are our time in skyscrapers and in airplanes. While our `ground' altitude varies all over depending on where we are, these variations are not, for the most part, very important to us. So we only rarely need a picture of the third dimension in our daily lives. And we lose little by generally regarding our world as pretty much flat.

Maps of real spaces have one particularly important characteristic. The rules by which they are constructed are generally quite simple and quite explicit, particularly when the maps only cover a relatively limited space---thus allowing three dimensional effects to be suppressed without much significant loss. Explicit and simple rules are important because otherwise it would prove to be difficult to interpret the significance of the map.

The Space of Ideas

But, the space of ideas is a much more vague and nebulous place. For example, I don't know of any metrics that allow one to measure the distance between ideas.

And while lines representing roads may be a very sensible way of telling something not only about the nature of the linkages between places (4-lane highways, dirt roads, ...) `lines' or other symbols don't seem to have any natural analogue in describing the relationship between ideas.

WIth ideas, or other things more ephemeral than places, representation is a very difficult issue. For example, while it seems quite clear that we might want to represent population centers on many conventional maps, what should constitute nodes on a map representing ideas is a good deal less clear. And while places, for the most part, map naturally into nodes on a conventional map, what constituted the `nodes' themselves is not at all obvious once the things being represented cease to be concrete.

Perhaps more important, the rules governing both the construction of the connections between the nodes becomes very problematical as we move away from the representation of concrete, real things. While we might want to `link' two concepts that we feel relate to one another, how should we express different kinds of linkage that might form the basis of the linkage relationship. It's one thing to show love and hate, but what about unconcern. If closeness is a surrogate for `love', is distance a surrogate for hate or for unconcern?


Some readers may be familiar with the wonderful book Flatland. This book teaches us something about the three dimensional world by looking at what things would be like if we were forced to live in what was intrinsically a two dimensional world.

What the book speaks eloquently to is what happens when we try to use fewer dimensions to present something that actually has more dimensions. It can be done in limited circumstances, but there is almost inevitably some substantial loss.

In this case it seems clear that even if we are able to use the computational capability of modern technology to let us use two dimensional screens to present three dimensional information, the world of ideas may have many more dimensions than three.

What Maps aren't good at

And maps, while they are a useful way of presenting some information, are also not very good at dealing with other kinds.

It would be pointless to list all of the things that maps aren't good at. We don't have enough paper. Maps occasionally help us with snippets of history, but history also requires lots of non-map text to tell whole stories. And maps don't have much application, at least so far, presenting poetry or music or countless other things.

Indeed, maps are actually quite bad at representing some things---and cause us to dramatically distort our views. A simple example might be the `public' perception of Greenland. Many people, if they have any picture of a map in their head at all, grossly distort the size of Greenland because of the artifact of the Mercator Projection that is generally used on the maps most people are familiar with. One characteristic of this projection is that it inflates area as we move away from the equator and towards the poles. This makes Canada, Russia and Scandinavia---among other areas---appear to be very much larger than they actually are.

Whether ideas belong in the category of `things that can be enhanced by map-based presentations' or not still remains to be seen.

Example; A Time Sensitive Map

We can even have considerable trouble using maps to represent relationships which are not far from those involved in simple distance.

Let's say that we want a map that shows physical place, but where the distance between points is more of a surrogate for travel time than just the simple mileage. If there were only one mode of transportation, then presumably this would be a pretty normal map. However, with multiple modes of transportation, the spatial distances become distorted. It is generally pretty easy to fly from NYC to LA. However, given schedules and small airports, there may be no effective way to get from NYC to central Pennsylvania other than to drive. This may well place central Pennsylvania `further' from NYC than is LA.

The argument here is that if it is this difficult to represent things which are not very far removed from the simple direct analogue of physical space, just imagine how hard it is going to be to find some other less grounded dimensions which can meaningfully be shown in the two (or with computers, perhaps three) dimensional world of maps.

If Simple tings are hard ...

... then the prospect for handling hard things isn't at all good.

This is one of the sad lessons of modeling. The kind of modeling that we do in physics and chemistry is hard serious work. But for the most part, and complex as these models are, there is some reality that we can have recourse to resolve issues associated with representation and manipulation.

When we model more nebulous constructs we generally one or two more levels removed from an ultimate reality. Our abstractions cannot be grounded in the same way that they can when we are dealing with real things.

Looking Forward

Maybe it would be useful to spend a few moments looking at the future for both analogue and metaphorical maps. The availability of computers has dramatically changed the nature of this world, and we should spend at least a little time considering how.

The Effect of Computers

It is only relatively recently that computers have been thrown into this mix. We now have widely available computational power that only large-scale organizations used to be able to afford. It is quite natural to re-cast our problems in terms that might make their solution susceptible to these new capabilities.

Thus, one answer to the question Why do you draw maps of topics or ideas? is Because we can. What we haven't heard so far is Because it's productive to do so, look at this ...

Learning from Randomness

The difficult thing about trying to generate some data that might tell us if this kind of mapping is worthwhile, is to separate the random bits of chance accomplishment from the regular kind of knowledge that can be acquired from some new, and productive, approach to problems.

This is not easy to do. Apocryphal stories abound, and they are difficult to separate from a more rigorous telling of potential accomplishments.

Before we believe, and pass on, such stories, however, we either ought to:

  • Make sure the data is solid; or
  • Gain some experience ourselves.
Without satisfying either of these requirements, we are simply passing along what is essentially unsubstantiated rumor---and that's not a productive thing for anyone.

Is there any `There' there?

So making maps of topics or idea maps are a superficially attractive idea, but it remains unclear if they are solid accomplishment or merely some distractive smoke and mirrors. From what I have seen so far, I tend to believe the latter.

I am reminded of the ancient joke about the thrice married virgin. Without going into gruesome detail, the punch line was And my third husband was an IBM sales rep, and he spent all his time telling me how good it was going to be. The target dates the joke, but the point is still well made. I have seen a lot of discussion of how wonderful these visual tools are going to be, but so far---even from the best and most intelligent---no examples that are even slightly convincing. And I can much better imagine someone else using the various technologies than I can see it applying to any problems I have. In the past this has been as sure sign, for me, of an idea that really doesn't fly.

The `Some Can' Argument

As I build the case that technology such as mapping topics is unlikely to be productive I have to be careful to note that this doesn't mean that the technology is profitless for all. Some may, indeed, find it useful.

However, some people find standing on their head to be useful in focusing their attention on problems that they have to solve. And it would be unreasonable to doubt their reports.

At the same time, it would be equally (or more) unreasonable to recommend `standing on one's head' as of much use in the problem solving process for most people.

So we have to be careful in extrapolating usefulness of things we find effective to others. Drinking coffee, eating a sandwich, listening to music, ... are things that some find useful---but that, on the average, we find little to recommend any of them above any other ritual practice.

And maybe `Some Will'

It is also clear that someone may, someday, discover some useful rules that allow some productive kind of idea maps to be drawn. While I would not place the probability of this happening high on my list, there is no question in my mind that it is possible. And if it were to happen it might truly revolutionize the way we look at some problem solving activities.

Then again, one can always win the lottery. And in this case that might be easier.

The Future for Maps

Even the future for maps is not so clear. Maps, themselves, are particularly useful in an age when `live' computation is impossible or very expensive. In such circumstances there is real value in coalescing a huge quantity of information in a single document which then can be published and circulated and is assumed to have a durable life measured in days, months or years.

However, these days different sorts of maps are quite conceivable. For example, I well remember driving from Florida to the Northeast during the great snowstorm of 1996. What I wanted was not a map of the US, nor even maps of Florida, Georgia, South Carolina, ... Instead I really wanted a detailed map of I95 with detail for the few miles surrounding it. And I wanted information about the state of roads in these few miles. I wasn't concerned about the rest of the world given the set of problems that faced me during those days.

Such a thing is, practically speaking, quite possible with today's technology. As recently as ten years ago it would have been a project that required millions of dollars of investment, but today it could probably be done for modest thousands. This is a sign of some things that are not only likely to come, they really _are_ here today.

But this means that there are limitations to the value of maps, even in the analogue world. And whether they represent progress when we think about applying this particular technology to the problem of mapping topics and ideas becomes even more problematical. It may be that even if maps are a stepping stone to the future they represent a step that is backwards and need not be taken.

Inferences from Metaphor

Making inferences from a metaphorical model is a very dangerous business indeed. Making inferences from a straightforward analogue model is difficult enough. Going the `extra mile' into abstraction clouds both the data and the applicability of the data to the problem even more.


I remember the long slow process of learning how to read maps. I don't mean simple road maps. I always did well in the basic skills tests that measured that kind of map reading skill. But some time in the 4th or 5th grade someone asked me to figure out the route for the shortest distance between Los Angeles and Hong Kong.

Knowing the `shortest distance between two points' rule I drew the obvious line. Then we went to a globe with a piece of string and let the string mark the real `straight line'. Of course the two routes had little to do with one another.

This is an example of a nice simple model and a straightforward piece of technology that goes horribly awry when it is inappropriately applied.

While at the moment I don't see any dangers to using the visual technologies to `map' ideas, I don't see any profit in doing so either. And my guess is that the time spent doing this will prove to be a profitless investment.

And whether we should use the capability provided by our modern computational power to present pictures or, rather, more elaborate and complex text structures still remains to be seen. For example, is it easier to understand some odd two dimensional picture of ideas, or rather some `constructed on the fly' concordance-like document that shows the appearance of concepts in context.

My bet is that the latter is more likely to be productive than the former.

Is Google a Glimpse?

Let me close with a speculation. In one view Google may just be a Live Concordance Generator with some higher level intelligence built in that allows it to do even more than is done in a simple concordance.

Google has been a success from almost the first day that it went on line. Recently there have been some clever adaptations of the Google technology to allow graphical presentations to piggy-back on this technology. Whether they will prove to be more productive than the conventional Google presentations still very much remains to be seen, but my bet goes with the more conventional text access and presentation.

In my opinion the value of visual representations has been misunderstood and overplayed ever since the ancient Apple adds which showed a child in diapers using icons on an Apple screen to play with the computer. Apparently the message we were supposed to get was So simple that even a child can use it.

The message I got, though was Great technology if you are a two-year old and your problems are that simple. That's the message I get, so far, from the attempts to use visual technologies to support the processing of ideas.

5 May 2002: ASCII Export and Import

I have a special reason to write this paper. I see systems that might interest me nearly every day,but almost always they prove to be unusable because failure to provide for some decent form of ASCII Export/Import renders them impossible to integrate into existing systems. While this paper may not be of much general interest, I hope some designers of systems that attempt to support idea management and related areas will pay attention to the needs expressed here. Then we might get some useful software of this kind.

The purpose of this paper is to argue that systems, particularly systems that deal with ideas, knowledge and the like, should have a complete ASCII Export / Import facility.

The fact that many, indeed probably most, systems---even those which are newly developed---do not have such capabilities suggests that the underlying issues are not well understood. This is important both to the developers and to the users of these systems.

These capabilities are ultimately most important for the users, of course. It is the user's ideas and knowledge that needs to be preserved across time and across the different computer systems that are bound to evolve over such periods.

The fact that these capabilities are important to users directly implies, though, that the developers of systems also need to understand them so that they can provide the needed capabilities.

Sidebar: Software Collaboration

A moment before proceeding. We are about to discuss software that can usefully be classed as collaborative, but it is possible that two rather distinct meanings for the words collaborative software will cause some confusion.

One use of these words is to describe software that has as it purpose the fostering of collaboration between people who form some kind of functioning group. This is not the use of collaborative software that we intend here.

For our purposes, we are looking at pieces of software that---themselves---collaborate well with other pieces of software. Typically this means that they are able to deliver their information to other pieces of software easily, and they are also capable of receiving information from other pieces of software in order to incorporate those pieces into their own domain.

If you buy ...?

The argument made here is straightforward. Software that purports to support ideas or knowledge management should allow complete ASCII export and import of their data base. This kind of software has a particular need to integrate with other programs that we may already be using to deal with this important world.

The core of this idea is that it makes the information being managed available to other systems that we might want to use that would relate to this task in our world. In addition it allows for information from other systems / sources to flow in to, and out of, the domain being managed. We are looking for collaborative software in the second sense that we have just defined above.

If you already buy this idea, then you probably won't learn much from the rest of this paper. However, if you don't, read on---perhaps you'll be convinced.

The Special Problems of Ideas

There are some special problems associated with managing ideas. These problems suggest that we should carefully consider new technologies in the light of what these particular characteristics demand. Our ideas and knowledge management probelms differ from other problems at least three possible ways:

  1. Their importance;
  2. The fact that our ideas and knowledge are useful many places; and
  3. Our ideas may well be long lived.

Understanding these facts is an important step on the way to seeing the importance of the kind of ASCII Export / Import facilities that we are concerned with here.

What is ASCII Export / Import?

The ASCII Export / Import capability is conceptually quite simple. We can describe it as follows:

The system has the capability to export a complete database into an ASCII file that in turn can be imported back into a clean copy of the system to reproduce the data base.

This (for reasons that will be clear in a moment) is called the weak capability.

A further capability---the strong capability---adds a requirement that allows the ASCII form of the data base to be manipulated and then reimported. A good test of this might be, for example, to replace some phrase by a phrase of differing length.

We need to require that any knowledge/idea management and manipulation system possess the strong Export / Import capability. The rest of this paper is devoted to explaining why.


First, our ideas and knowledge are particularly important---to us at least. For the most part good ideas are hard to come by, and we don't want to place any undue obstacles in the process of capturing, manipulating and displaying them.


Second, we, practically speaking, cannot be restricted to only one way of dealing with such information. Each of us needs some different kind of support to enable us to properly deal with our ideas.

Useful Many Places

Third, ideas are often useful many places. For example, some of the ideas presented in this paper belong not only here, but may also have a useful role in a paper that describes the technology that manipulates ASCII archives, as such things might form a useful part of some papers describing such technology.

Of course, there is another fairly obvious class of such uses, namely in the proposals and reports that might describe projects associated with our doing research in this area.

Long Lived

A fourth important characteristic is that our ideas tend to be long lived. It's not that our ideas are immutable, but rather that they tend to change slowly. Indeed, I suspect that the better the ideas are the less likely they are to mutate with time.

If this is correct, then we are likely to want to use the ideas in other projects, archives and databases where they complement and supplement other notions.

Erroneous Assumptions

Many of the systems built to manage ideas seem to be constructed with some underlying assumptions. The problem is that many of these assumptions are erroneous. Let's look at the assumptions of

  • Completeness;
  • No Past;
  • Permanence;
  • Separation; and
  • Universality.

Let's look at how some of these assumptions are evidenced in some of the systems that are currently available.

To be fair to some of the systems that are in the current marketplace, few exhibit all of the characteristics of the problems we are about to describe. We will use the typical artifice of discussing rather extreme examples of each of the characteristics. While this will overstate the case somewhat, it is hoped that it will help in making the distinctions more clear than they might be otherwise.

One further point deserves to be made. If we divide the potential audience for this kind of technology into two groups: those relatively sophisticated---who have used some technology to record and manipulate their ideas before---and the naive, who have not, then people with prior experience probably make a better potential market. This is only important because we should recognize, early on, that the ability to absorb prior efforts may be an important component that needs to be dealt with.

Assumption of Completeness

One bad assumption is that the system is complete. This is a very bad and very limiting assumption. Taken to the extreme it implies that the idea management system will have to implement access to its own editor, spell-checker, etc.

Either important resources will have to be devoted to developing these technologies (or at least integrating some pre-existing equivalents into the environment) or the capabilities will have to be sacrificed.

Assumption of No Past

Another bad assumption is that there is no pre-existing history of recorded ideas. As we suggested above, for many---probably more each day as time passes---the capability to incorporate and merge with some technologies that have previously been used to deal with this problem are very important.

This is not necessarily an easy thing to do. Indeed, there have been a large number of different pieces of software developed over the past several decades that we would have to adapt to individually if we were to try to build specific bridges.

If, on the other hand, a general approach---such as using ASCII data structures---is used, then we can probably build, without too much trouble, some specific bridges that would prove to be helpful.

Assumption of Permanence (No Future)

Another assumption implicit in lots of software is that it will last forever. Indeed, any piece of software which implements a hidden storage structure had better last forever or else information recorded in it will essentially be lost.

Specific examples are easy to come by. I have ASCII files from computer systems that were created twenty or thirty years ago. So long as they have been moved to media that are still accessible, the files are still readable. Other files, those which depended on strictures and markup which was not explicit, can no longer be read because the software that is needed to access the information in the files no longer runs on any machine that I have access to.

Assumption of Separation

Another important erroneous assumption is that the software that manages the ideas and knowledge can usefully stand separate from all kinds of other software that is used to manage things which might otherwise somehow be related to the idea and knowledge management capabilities.

For ideas and knowledge this is a particularly poor assumption. There are lots of correlate domains which might usefully be related to ideas and general knowledge.

Assumption of Universality

One further problem remains. Lots of the software that populates this area makes the assumption that we all want to create and manage our ideas in the same was as everyone else. Since we all seem to generate and manage ideas in our real lives in different ways, it would not be surprising to find that different people might want to handle the problems using different software approaches.

Forcing everyone to use comprehensive software that will cause them all to treat the problem in the same way is not likely to be a great service.


Systems that are implemented making these assumptions exhibit some rather obvious difficulties. First, they tend to be narrow. Systems which are narrow deal with only one face of a problem. They don't link with other systems which might exist that handle other faces of the problem.

These systems typically are constructed on the basis of a model held particularly dear by the software implementer. Often this is quite a rich and interesting view of the problem domain, but there is no guarantee that the way the developer has approached the problem will necessarily parallel our own approach. If we are not afforded the opportunity to mix our own idiosyncratic approaches into the problem space, then these systems become a limitation on our capabilities rather than something that enhances them.

These systems may also require special training. If they don't use pre-existing editors, for example, to capture and manipulate the information, then our staff needs to be trained in the characteristics of this new software. They are not able to use their pre-existing knowledge or their old ways of handling the problem. This is not only an extra expense, it is a burden that requires extra investments of the time and energy of the members of our staff.

The Broader Case

There is also a broader case to be made for ASCII import and export. If we can, and do, regularly export our idea databases into ASCII form, this will dramatically increase the likelihood that they will be able to provide some useful form of archive should we ever want to re-encounter them in the future. If, on the other hand, the data is stored in a way that requires access through specific programs in order to realize any useful result, then we may well find that our archives, even if we carefully provide for them to physically survive, will become essentially useless.

Take as an example an idea database that is only accessible (logically) through a program that runs under Windows. If at some point we convert to Linux, all of our idea databases will be inaccessible and all of the information in them will become quite difficult to reach. If we are talking about a historical archive this kind of event may take place five or ten years after our data was created, and we may find that we have no effective way of taking the `backward step' into recreating a prior environment that would allow us to make use of the data.


If we have and ASCII Export / Import facility then many of the problems that we have been discussing become quite easy to handle. For example, if we already have editing programs that are well understood by us and our staff, then we are able to continue to use them to manage our ideas. We just export our idea data into ASCII, manipulate it, and import it back into the idea database system.

Further, ASCII files are easy to handle in programs. If we can export our ideas to an ASCII format then not only can we use things like pre-existing spell checkers, but we can also write code in languages like perl and Python that can manipulate the information in the documents in ways that are useful.

Wide Range of Tools

Indeed there is a wide range of tools that are already available to manipulate ASCII information. In our own environment, for example, we have software that can help up mark up the text (enforce project standards about how certain objects should be represented when they are mentioned in documents, as a case in point).

Other tools allow the manipulation of text in an SQL database. This makes the text available not only in the form of documents, but also arranges to publish the information on our web sites.


And when in ASCII form the data in our idea databases becomes available for adaptation into concordance, dictionary, spell checking and other kinds of programs.

Fit with Existing Resources

In our case, it is also particularly important that we are able to adapt some of our existing resources into the idea management system. We have been recording ideas in a computer readable form for at least three decades. This information already exists, and we don't want to have to enter it again, just to get it into some form that is appropriate for manipulation.

The Recent Problem Case

It is perhaps worthwhile to give a specific example of some software that appears, at least as of now, to have made the mistake of being closed with respect to the kind of issues we have been discussing here. This is ADM Idea Management Software, and it was the evaluation of this particular piece of software that provided the occasion that resulted in writing this paper.

ADM's system is, superficially at least, quite attractive. I was very impressed with the initial look and feel of the screens, and the software definitely looked as though it might be a very useful adjunct to help in the task of managing ideas.

However, there was no ASCII Export / Import facility other than a rather trivial ability to `export' a file in ASCII for the purposes of printing it. This is not, of course, an adequate form of export. It does not produce a form of output that can later be turned around and re-imported directly into the machine.

Frank discussions with the implementers of the ADM product indicated that they might be considering some sort of export/import facility real soon now, but their horizon proved to be 6 to 9 months (given previous history in this area in general this usually means 12 months to 3 years) and so we dropped them from consideration until the directions of the later evolution of their product become more clear.

One Example: The Vault

It can be done. We have, so far, had one particularly happy example of a place where a piece of software was designed in such a way that made it easy to incorporate it into a collaborative environment.

This piece of software is The Vault, a straightforward and clean little outline manager developed by Eric Bergman-Terrell.

The Vault presents an outline and text in a simple two panel display, with the outline on the left and the body text on the right. I find that this format makes it easy to

  1. Create the text; and
  2. Manipulate the outline and the corresponding text.

While Bergman-Terrell's choice for storing his data is not ASCII, he has provided a complete spec of his storage format. This made writing the ASCII export / import `dump and restore' program a straightforward task that we accomplished within a day of getting the complete spec.

Now we use this program as a collaborative piece of software as we write our papers and manipulate our outlines. Members of our staff are free to use The Vault if they find it to their liking, or they can manipulate the information that is their job however they want, as it is available to them in ASCII.

Fit with TextSquibs

This approach very much fits with the approach we are taking in the TextSquibs Project at Wharton.

21 April 2002: Manhattan and the WTC

This is a complete departure from the normal computer / technology theme that is present in most articles published here. It probably won't happen often, but seemed to be appropriate given the importance of the subject.

I have worked in Manhattan on and off for many of the years of my active career. New York is a rich environment that offers many lessons. And in the aftermath of last fall's horrible events, there is some learning to be done.

I visited Ground Zero a first time not long after September 11, but it wasn't until March that I really had a chance to wander around some of my old haunts. And there were some real contrasts in the both the look and, perhaps more significantly, in the feel of Lower Manhattan.

At the time of my first visit, life in lower Manhattan was distinctly not normal. At that time there was truly a dead zone around the site, extending for blocks in all directions. Stores were all closed, and while there were lights in some buildings, they were only a sporadic sight---mostly the places were dark and foreboding. There was little traffic---other than trucks carrying rubble---and large areas seemed mostly deserted.

The March visit, however, was both a joyful and a jarring experience. And the reasons were, perhaps, instructive, and were highlighted by some contrasts. And there were some important lessons to be learned from these contrasts.


The principal joy was a result of the fact of how normal life seemed to be. True, the South Ferry subway station is closed, and there's some construction in evidence around Battery Park and the Ferry. But that's an expected, everyday occurrence in a town so often covered with Dig We Must for a Better New York. And while traffic seems lighter, it really wasn't markedly so---not much more than you'd expect from the road closures due to the construction barricades.

And everyone seems to have that same sense of haste and determination that was always so in evidence around this part of the world. The only people that seem not to be hurrying somewhere are the traders who are taking a break to smoke their huge cigars while carrying on with business on their cell. And the humor that was a part of the street scene, along with the pathos that was in evidence in daily life also seems pretty much back to normal.


What was jarring, to me at least, was how much better lower Manhattan seemed to look without the looming towers. And even more, the fact that I haven't heard this discussed anywhere over the past months.

This isn't to say that I don't miss the towers from the skyline. I do. And I keep trying to re-create in my mind's eye where they were and how each particular view used to look. But that is an effect from a distance. From a distance they are missed. But close up is another matter.

I hadn't realized how dark the towers made lower Manhattan---at least the part I frequented. And how boring and tedious a presence they formed as they dominated the scene.

As one looks towards the Hudson from Broadway today, there is light and almost distant views---no looming huge undifferentiated presence. Not a wall of regular repetitive windows, but a varying view that seems as rich as New York itself.

And I would have thought that this was something that deserved comment. Many of us are convinced that `progress is progress' without realizing what we sacrifice in terms of the ambience of daily life. Jarring events, like the sudden disappearance of these huge towers, bring us up short. We are able to see in a startling way the impact of the decisions we make about how our everyday space is occupied.

And there are lessons to be learned. There can be a slippery slope to urban development. And we may be sacrificing important values as we make changes, each of which is seemingly small, but which have a cumulative effect that becomes overwhelming.

A PC View?

I wonder if no one else comments on this because they don't share the opinion. Or, do many others share the opinion, but find it not politically correct to represent such views? Whatever the case, we are passing a valuable opportunity to learn something useful if we neglect to discuss this positive aspect to the otherwise unremittingly negative aspects of the disappearance of the towers from the scene.

Marginalism and the Slippery Slope

It is important to understand this case because it is---fortunately for us---such a rare event to have a cataclysmic `turning back of the clock' in our lives. Most of the time, things change is dribs and drabs, and our tasks make our days too busy for us to be able to notice it.

And we plan, and our plans go awry, but usually only in small ways that we are much too busy to notice. We absorb small changes, each of which only seems to have tiny effect, without ever realizing the cumulative consequence. We add a half hour to our commute in order to achieve a better life, and then five years later do it again, and again and we work ourselves into two hour commutes sort of without noticing it.

It takes some consequential event for us to be brought up short---to recognize that some of the changes have cumulative effect on us that huge. And when the opportunities for those lessons come along, we should pay attention to them---lest we miss the lessons that could inform and alter our future.

There is also, perhaps, a comment on an even higher level lesson: there are some slippery slopes. While we may argue that one or two more tall buildings won't have much impact at the margin, there are cases where differences in degree produce important differences in kind, and we find that they do have a distinct impact on our real daily life.


So that is at least one of the lessons of the loss of the these huge buildings in New York. This isn't to minimize the tragedy---and the horrible cause of the disappearance of the overwhelming presence of the towers. But it deserves to be noted that while they may have been economically effective and highly utilitarian, they were also---to me at least---an ugly and jarring presence, whose unfavorable impact I hadn't taken time to notice until they disappeared.

14 April 2002: The Outlook for Outlook (and other PIMs)

Last week the paper was very long. This week time managed to disappear, so with luck this is a bit shorter. And this time, we discuss Personal Information Managers. To do so, I'll borrow from the mythology of my Norwegian heritage and talk a little about Lemmings.

But that will come later.

Personal Information Managers---PIMs---have been around for a long time. They seem to be re-invented with every generation of computation. And, somehow, they just don't seem to catch on as we'd expect them to. So they die, only to be reborn in another cycle, often by people who don't apparently know much about the earlier history.

Let's see if we can figure out why.

What are PIMs?

I don't want to spend a long time describing PIMs in gruesome detail. For our purposes here it is sufficient that we focus on the hardware and software realizations of technology that allow us to manage personal information. This is principally:

  • Phone / Address data;
  • Scheduling and Time Information; and
  • Some rudimentary Project / Financial Data.

Of course we could broaden the picture, as some PIMs do, to include more types of information, but it wouldn't any better serve our purposes here to do so. We can leave that for other more in-depth treatment later in other documents.

We will also not bother to spend time differentiating the computer-based aspects of PIMs from the well-known hard copy realizations of this technology that are so popular in the form of Filofax(tm), Day Timer(tm) and other paper-based instruments. These all relate---and to some extent at least cover the same territory.

Picking on Outlook?

For the most part, while I think Outlook is a lousy piece of software, I really don't want to pick on it with a particularly heavy hand. Outlook is only the current candidate for most common core piece of software in this problem area. It is positioned to be in the center of the whole set of products that form The Microsoft Office, and since that is the most widely-used software suite in the commercial world, Outlook assumes the role of what is probably the most widely used PIM.

Outlook probably isn't the worst PIM in existence, and it safely isn't the best either. And other than the disadvantages caused by the fact that it is the most common PIM (we'll track down and discuss some of these problems later), it's problems are generally shared with almost every other PIM.

So, where necessary we'll use examples drawn from experiences with Outlook to give focus, but not to pick on the particulars of it's design or construction. It is, for most purposes, quite typical and what we about PIMs in general applies to it, was well as to a lot of other software.

Some History

My first recollection of PIMs predates the world of PCs. I can't recall any software or hardware devoted to PIM-like purposes back in the days of the individual large-scale mainframes, but as soon as we began to get personal utilization of computers it was natural that some part of that technology would end up being applied to the problems of managing personal information.

My first encounter with such technology was some crude programs which we developed as soon as we got our hands on a time-shared computer in the middle 1960s. It seemed to be a natural, at the time, for passing around information about our schedules, as we had to schedule project meetings in the face of lots of other demands on our time, so we created an early form of EMail as well as crude schedule management technology.

The first publications that I can find in my personal archives that deal with this kind of problem are from our Office Automation Project in the 1970s. By then we had a well established EMail system as well as basic scheduling tools. I have in hand, for example, a copy of a Working Paper of the Department of Decision Sciences at Wharton. It is dated 15 July 1975 and is titled: Office Automation Project: Personal Scheduler Version: 2-1C(6); Since I remember our numbering scheme, I can conclude that the version information relates to the version of the software, not to the version of the paper. So I can say that we were obviously at least six versions into our software as of the indicated publication date.

I won't claim that the paper is very interesting---other than as a historical artifact---by modern standards. But it does detail the purpose and use of three modules The scheduling system consists of three modules:

  • SCHCRE---used to create new items on a schedule;
  • SCHSCH---used to print out a daily listing of events; and
  • SCHFST---used to format input in order to schedule a sequence of events easily.

The point of mentioning this is just to establish age. The use of computers to help in scheduling tasks not only pre-dates the rise of the net, it precedes personal computers by at least a decade. We have at least 25 years of experience (actually 35 is more like it) with this kind of technology.

And it seems to me that it has had remarkably little impact on our lives, and has lived through birth and death cycles with each new round of hardware and software advance since the 1970s.

Two Streams

There are two separate streams that deserve analysis. First, there is the issue of the evolution of problem domain that PIMs deal with. Second, there is an evolution of hardware from:

  • Time-shared computers; to
  • Personal Computers; to
  • Networked Computers; to
  • Handhelds
Each stage of this hardware evolution has presented different opportunities, and yet each has also proven to fail---ultimately---to gain the kind of hard won acceptance that allows them to function as a central part of our daily lives over a longer period of time.

The Changing Problems for PIMs

Modern PIMs exist in a hardware/software world that is changing quite rapidly. And our problems are changing as rapidly as the technology is developing to help us deal with the change.

Phone numbers and addresses are nice examples of things that have been relatively easy to manage for decades that now are subject to considerable change and require a more active strategy to manage them.


For example, the nature of telephone numbers has changed rather dramatically since the advent and widespread availability of the cell phone. Area Code is becoming insignificant or at least less and less related to a physical location. I currently carry two cell phones, one based at home and one based in my daughter's calling area. It's substantially cheaper this way. And the fact that one of my cell phones carries the same area code as my home phone means less and less every day. I have essentially free (at the margin) long distance service, so I no longer pay much attention to where I am calling. And when people call me they are as likely to reach me in some remote city as they are to find my in my home region, even though they call my home area code.

All of this means that my phone number is a much more volatile piece of information than it used to be. I had only one phone number for the first seventeen years of my life. Today I have four personal phones, each serving different places or dealing with different problems, and it is not uncommon for each kid in many families to have a phone, and sometimes a cell phone as well.

So managing phone numbers is a bigger deal than it used to be. We have more numbers and they change faster than ever before. And less and less we rely on our memory to bridge the problem.

Machine Readable Information

The fact that cell phones increasingly contain their own directories has two important effects on PIMs. And they have potentially differing effect.

In order to understand this effect, it is important to note that increasingly our lists of phone numbers need to be in some machine readable form.

First, the phones themselves increasingly store the numbers. This means that we have less and less occasion to use the numbers other than through the phone. So this decreases our need for having access to the numbers live. In fact many commonly report that they no longer have a clue about what real numbers are, `only their phone knows for sure.'

Second, and pulling the other way, is the fact that increasingly we need to be able to load the numbers in some automated fashion. Every time we change cell phone, or have to live through the horror of severe battery failure, we have to reload our numbers. This can be a daunting task if the numbers have to actually all be entered again by hand.

So PIMs, insofar as they can be used to manage the lists of numbers for uploading into the phones themselves, are of increasing importance, but insofar as they are needed as direct sources of the numbers themselves they are of decreasing importance.


And a similar thing has happened to addresses. It used to be that addresses only represented real, physical places, and as a result they tended to be rather stable over time. We changed addresses only when we moved. And we didn't move very often.

But in the last few years, we have started to create virtual locations: EMail addresses and Home Page URLs are two well-known examples. And while physical addresses are generally quite stable, the is certainly not the case with virtual addresses.

For example, my ISP went bankrupt not long ago. This is common occurrence these days in the computer business. And my ISP housed by EMail address. So even though my business was transferred from one company to another without much change on my part (it took a couple of hours, and didn't involve anything other than software changes), my EMail address had to change to accommodate the new ISP.

So managing this more volatile and complicated addresses also presents new challenges.

The Changing Hardware

Hardware is also changing dramatically. The past decade has witnessed a decrease in cost. When combined with the usual improvement in speed and capacity, coupled with vastly improved small computer screens this has changed the nature of handheld computers. Whether this change will, in the long run, materially effect the way we deal with computers remains to be seen.

The Material Changes

While everything changes with considerable rapidity in the computer business, some of these changes make more of a difference to personal information management than do other changes. Let's track down at least four threads of possible importance:

Processor Speed

First, processor speeds have increased dramatically. When the first PC's came out in the late 1970s and early 1980s they had speeds in the range of 2 million operations per second. Now we have wide availability of machines that run 1,000 times that fast.

However, for the purposes we discuss here, this kind of speed change really doesn't have a very dramatic effect. None of the tasks that are commonly performed when managing simple personal information requires a great deal of processor capacity.


Second, storage capacities have changed in similar magnitudes. When buying my HP200Lx a decade or so ago, the big decision was whether to get 1 million or 2 million bytes of storage. And a 5 million byte extra storage cost several hundred dollars.

A year or so ago, I bought a little disk about the size of a quarter that holds a billion (US) bytes. It cost just about as much as the 5 million bytes bought roughly a decade before.

Unlike processor speed, this does make a difference. A billion bytes is truly a lot of storage. A book of 1,000 pages might be a few million bytes. So it would now be feasible to carry hundreds of books worth of information. And this allows us to tackle and solve some potentially interesting classes of problems.

A good example of the significance of this is the way digital photography has swept the photographic world. Modern cameras take pictures that might approximate a million bytes in size. This is quite a bit of space, and memory becomes important. With current storage devices we can now store hundreds of photographs on a simple small device. On my last trip to Europe, I made about 200 photos, and had room for 500 or 600 more on my CF hard disk memory.

Communications Bandwidth

Third, we are right in the middle of a parallel transition in the speed of communications. My first teletype allowed remote use of my computer at a rate of information flow approximating 10 characters per second. My current cable modem allows me to communicate with the net at more than 100,000 characters per second.

This change in bandwidth capacity is particularly important because it is a good example of a place where a difference in degrees is large enough so that it becomes a difference in kind. At some rate of speed, probably in the neighborhood of 5,000 characters per second, it becomes practically feasible to transmit graphical information as well as text. At low speeds graphics simply aren't feasible.

However, in the US at least, cellular phones don't connect at anything like those rates. My cell, when it works, allows me about 500 characters per second. That is slow enough that I rarely bother to use it, preferring instead most often to wait until I can have my high speed line. Whether I'd use it more if it were dramatically faster is an evaluation that will have to await some experience.


The fourth factor is, of course, cost. The cost of carrying sophisticated computing equipment has dropped so dramatically that many things which weren't feasible become quite practical. But it will take some time before we gradually develop a notion of which of these feasible things will turn out to be of practical interest.


So, some aspects of the technology now allow us to manage our personal information in different ways. Whether these will actually prove to be effective or useful, however, must await further experience and experimentation.

Indeed, I rather think there is less here than meets the eye, and that the zealous projections of computer based PIMs gradually dominating more and more of our lives are way too optimistic with respect to the likely uses of the technology over time in the future.

Why doesn't it work better?

I have suggested that we have seen cycle after cycle of PIMs being born in an environment of excitement and with all kinds of optimistic projections about their future, only to gradually wither and die on the vine.

And I think it's happening again. And further, I think there are reasons for it that we can understand. Let's discuss some of them.

The Problem is Time

First, and probably the most important thing to understand is that the there is a fundamental problem with managing Time. There, simply put, isn't enough of it. It is this fact, and not the fact that we have difficulty accounting for our time that is at the core of the difficulties.

No PIM creates time. They simply allow us to manage it easier (if we're lucky). This is an essential core fact of our situation.

I learned a parallel lesson one time long ago when I developed an on-line registration system for an educational program. We spent a lot of time and energy trying to make the allocation of seats in our popular classes more fair. But the real core problem was not actually the fairness but rather the fact that we simply didn't have enough seats in the popular classes. We could make---with luck---small improvements in fairness, but they were of truly marginal importance relative to the real problem. So we worked very hard and still had little positive effect on our real problems.

Disappointed Expectations

Second, this leads to disappointed expectations. We are subtly sold on the fact that we are wasting huge quantities of time. But, for the most part, only a little of this time is wasted because we fail to account for it in some appropriate way. And thus, when we get our fine new PIM we not only have to use some of our valuable time to learn how to use it (increasingly, it would seem, these devices get complicated), but we are also then disappointed to find that at the end of the process we actually haven't actually found any huge quantities of time released by our dependence on the new technology. And in the process we run the risk that in addition we may actually become dependent on the technology in a way that is detrimental.

Pandering to the God of Organization

Third, PIMs pander to the God of Organization. This is the (often unfounded) belief that better organization will lead to our being able to more effectively deal with our problems.

While it is likely true that some people are able to improve their organization of activities in a way that is effective for them, I regard the case as unproven with respect to most of us.

Some of us have a compulsive need for organization. Indeed I have known some truly pathological cases of organization. I know someone who got up at 4am every day to make lists and organize the day. It took hours. And, if my observations as an outsider were any indication, few of the tasks on the list were ever accomplished. The objective became generating and managing the list not getting the tasks on the list done.

So that's one of the downsides of PIMs. If we are not careful keeping the PIM up-to-date and organized becomes the objective, and replaces the original objective which had something to do with improving our life, not improving our PIM.

The Technology can be Exploited

PIMs also are a technology which can be exploited, sometimes to the detriment of its users. This is particularly true if the PIM in question is very widely used. This happens because a PIM can become an attractive nuisance to people not unlike those who are so interested in SPAM.


Since PIMs are used to maintain a lot of the addresses that are used---particularly for EMail---on the net, they are a natural focus for spreading a virus if someone wants to do that.

This is particularly true if a PIM is widely used. Then it is even more attractive as a focus for such efforts. Presumably those who spam and spread virii are interested in reaching (and causing havoc among) as wide a community as possible. By understanding how to do this in Outlook for example, one can have a much larger (negative) impact on the community of users than if one targets some much less widely used program.

The Technology is Superficial

Another aspect of the technology which contributes to its lack of staying power in dealing with these problems is that the technology is often rather superficial. For example, it was very much a vogue for a while to exchange business cards by `shaking hands' with Palms or other PIM hardware.

This already seems passe. While there were many PIMs in evidence in most meetings a couple of years ago, I have noticed a distinct decline in their presence in more recent days. After all, while it is possible to use a PIM to take notes or exchange information, it is also possible to do so with pencil and paper, and after the initial flush of excitement associated with the absorption of any new technology, after a little time passes we often return to older, more comfortable ways.

Some Examples

There are lots of examples. We have already given some. Let me close this section by telling a little of the further history of the Office Automation Project Scheduling System that I introduced above.

That particular system reached a high point of about two dozen users. This was a fair percentage of those of us who had access to the technology in our department. (Remember, back then computers were rare, and there was no widespread use, so not many people were in the community of potential users.)

After a few months, though, the interest dwindled, and I well remember noticing that about a year into the project we had dropped back to only a couple of users of the scheduling system---the author of the system not among them.

It was an early omen.

Witness the history of widely used products like Agenda for example. Agenda had (indeed it still has) a dedicated band of adherents. Many of them claim that no `modern' software can hold a candle to its basic capabilities. And yet even a company with the market impact and strength of IBM/Lotus couldn't (or at least didn't) make a go of it. And follow on products, notably Notes, have been less than overwhelmingly successful.

Similarly for Ecco, also conceded to be a fine product. A very active community of Ecco users was deeply chagrined when Ecco was simply abandoned several years ago. Again, many Ecco users would claim that even today there is no product that compares with its flexibility and capability. Yet even with and active user community and a good product, sufficient interest couldn't be sustained to keep the product available.

So the long run history of PIMs is not replete with optimistic examples.

Lemming Problems

All of this leads to a an observation about the nature of some problems. It bears repeating here in this context as it is relevant to the problems we are discussing.

Some problems are what we came to call Lemming Problems, named after those little furry Norwegian creatures that mythologically engage in an odd form of ritual suicide. (Best evidence is that real lemmings don't do this, by the way, but we are interested in the lemming of `myth,' not the real animal.) Lemming problems have just the same characteristic. They are suicidal---consuming time and energy far beyond any return that is obtained from dealing with them. And they kill not only systems, but careers as well.

Lemming problems have some characteristics that are relevant here. The are:

  • Falsely attractive;
  • Have little upside potential; and
  • Have huge downside potential.

In an earlier generation, payroll was just this kind of problem. Indeed payroll was both the driving force and the bane of the early stages of computerization in the corporate world.


Think about payroll. First, it looked attractive. It was easy to understand, and conceptually simple. And management could positively taste the savings involved in being able to get rid of hordes of payroll clerks.

But think about the problem a minute. It was, indeed, conceptually simple: Pay = Rate * Hours described a lot of it. But, as they say, the Devil is in the details. And for payroll the details were bewilderingly complex. Not only did we have many kinds of federal taxes to worry about, there were state taxes, city taxes and many different sorts of withholding. Getting it all straight took a lot of time and energy.

And there was little upside. Sure, some money could be saved in the salaries paid to payroll clerks, but for the most part these were not expensive employees anyway. And other than these savings there was no upside to payroll. You can't really imagine anyone talking about a good payroll.

But there sure was such a thing as a bad payroll: one that was late, for example. Don't bother to eat this weekend, we'll get you your money on Monday doesn't make for happy employees. The downside potential was huge.


I'd argue that the calendaring aspects of PIMs are just such a Lemming problem. Calendars are superficially simple, but again when you get into the details things can get very complicated indeed.

One example. What should a calendar say when it is asked if I am `busy' at a particular time? The real answer depends, surely, on who is asking. I might well be `busy' with some low priority event, and therefore quite ready to make time available if it were to fit with some other agenda. And the relative importance of events can shift as circumstances change. If I haven't seen someone in a long time, I may be more likely to cancel an existing appointment on their behalf.

So even something seemingly simple like `is a particular time available' has a very complicated answer.

And there are more details. One of Outlook's most disastrous mistakes, for me at least, has to do with the way it handles appointments in other time zones. To be fair, this problem may have been corrected---I gave up using Outlook a long time ago for this and for other reasons so I'm no longer au courant with any changes---but simply put it scheduled appointments effectively in the time zone that was in play when the appointment was made, not those that were relevant when they were actually going to be kept. No person would make this kind of mistake. When I call from New York to make an appointment for lunch in Paris, we can set a time without discussing time zone. Neither of us will be confused.

The downside of mishandling a calendaring problem is substantial. It is measured in missed appointments and missed planes. And perish the thought that your battery runs out at an inopportune time. Appointments can disappear with the loss of a few electrons.

And what about the upside. Aren't meetings, for example, easier to schedule? Perhaps. But this saves my secretary time, not me. And actually, if the truth be known, I'm really not very interested in making it easier to schedule more appointments. I go to plenty of meetings already. Making it easier to schedule more is necessarily helping.

The Sound of Music (and Video)

We might want to regard both audio and video as a new form of personal information. This would certainly be reasonable, and there is a lot of evidence that this form of personal information use and exchange has become very important.

This is something well supported by the technology. The MP3 players, for example, allow us to store from dozens to thousands of songs that we can carry with us and play with considerably more convenience than is the case with devices of earlier generations.

These devices differ from the radios of past times because they are under our control and will play whatever interests and amuses us---so long as we are willing to pay the cost of the effort involved in adapting them to our wishes.

Whether these devices are popular because of this freedom and flexibility or simply because they allow the piracy of otherwise copyrighted information remains to be seen. That particular battle is being fought out in the marketplace, the courts and in Congress right now.

But, in any case, management of this form of personal information is quite different from that handled by conventional PIMs. We'll have to leave that for later discussion after some more time has passed.

The 1939 World's Fair

Moving to the beginning of the end of this discussion, it's worth mentioning the view of the world implicit in the futuristic promise of the 1939 World's Fair. A lot of the current discussion of PIMs has the same flavor of over-optimistic hyped futurism.

The vision of the world embodied in the 1939 Fair was essentially that of three-dimensional life. People floating around in individual little helicopter-like devices. Streets that far overarched the ground. Airships `landing' and `taking off' from the high floors of buildings.

The pictures have proven, so far at least, to have an odd, and now since September 11, eerie quality that suggests they probably won't be realized for many decades at a minimum or perhaps ever.

Life today, in fact, looks a lot more like it did in 1939, than it looks like the futuristic picture portrayed back then. New York certainly has more tall buildings than it did in 1939, and the 3rd Avenue El is gone---leading to some clear improvements---but people alive in 1939 would have little trouble recognizing the City of today, even though many of the details may have changed.

Differential Impact

PIMs certainly have differing impact on people. For some, particularly perhaps those who bill `by the hour' they are an effective and easy way to capture initial data. They avoid the stage of committing time utilization notes to paper and then needing to have them entered into a machine readable form.

But there is lots of room for negative impact as well. For example, while it is true that one can lose a physical PIM, I have only done so once in the last thirty years (at Heathrow in the 1970s). I have `lost' my PIM four times within the past year---due to battery failure, not due to an increased sloppiness about the physical object.

And while in the latter case, I am able to restore my records, it generally costs me several hours to restore everything to my PIM, as many pieces of software have to be reloaded, and have their keys re-entered.

So the technology has a different impact. Good for some, though I think the number is small, and not as good for many others.


All of this makes me not optimistic about the futuristic pictures painted of the PIM connected world. I do see the opportunities for some dramatic changes---GPS are one example, being `lost' is pretty much a thing of the past for me given that I carry my GPS with me all the time. And the implications of our ability to keep track of our location remains largely uninvestigated.

And there is room for a lot of improvement in personal information that relates to important things like our medical condition. For example, as a diabetic I can now measure and carry blood glucose information that was essentially impossible as recently as a decade ago. As we develop more devices that allow us to capture historical data, significant opportunities may emerge.

But, while I see the opportunities for many of these things to have some dramatic impact on our life, I remain unconvinced that PIMs and their related technologies will provide much of this impact.

I am also pessimistic about the related issues of the world of the wireless connection, but that will have to await a later paper in this series.

7 April 2002: Instant Outlines

Sidebar: This piece got long. It was supposed to be a straightforward description of some of the less attractive aspects of instant outlines in an attempt to counterbalance what I regard as overly optimistic views presented by some others.

If you already buy the point that instant outlines are

  • Overhyped; and
  • Unlikely to be of much impact;
then you might want to skip this paper. If, on the other hand, you `think there might be something to it', read on before making any commitments.

Another vogue topic is Instant Outlining. For some (to me unknown} reason this is now being abbreviated as I/O---perhaps to hearken back to a past age when this was a common abbreviation for Input/Output---but the appearance of the slash seems to be an odd conceit, having nothing to do with the subject at hand. However, we'll accept the convention and use it here.

What are they?

Instant Outlines are supposedly a new way to manage the process of building knowledge about some particular topic. Presumptively they allow us to participate in content management by working in a collaborative work group. The term is being popularized largely in the Userland community. So maybe a short tour of Userland would be appropriate, as an understanding of where the idea came from can help us understand its strengths and weaknesses.

Userland and Winer

Userland is the creation of Dave Winer. Apparently it is a combination of:

  • A community of users;
  • Some software;
  • Support Facilities; and
  • A Web server.

The purpose of Userland, in addition of course to being a successful commercial enterprise, is to provide the community of users with software that helps them create and maintain web pages and to support them while doing so. In addition, if they want web services, Userland will also assume responsibility for the physical delivery of the pages to the net.

Other Ways

Other organizations deliver various parts of the same set of capabilities and services that are promoted by Userland, so there's nothing unique in their particular services or language. Like any capability / language / service some customers will prefer one and some will prefer others.

However, from the viewpoint of an outside reader, there is more intellectual pretense associated with Userland than with most of its competitors. Some find this attitude productive, and think that it contributes to the sense of community. Others find it cloying and self-aggrandizing. I find some truth in both views, but don't think that it matters much in either case.

Relevance to I/O

Why detour from talking about Outlines to spend time at Userland and their competitors? Well, I think it's important because it should be understood that some of the push behind the ballyhooing of these technologies may come from an attempt to beat the bushes to stir up interest in the commercial product. Not that there's anything wrong with that mind you, other than mistaking a self-serving veiled sales pitch for a useful intellectual contribution. But in the interests of disclosure the fact that some of the ideas may be being promoted because of a financial interest in the outcome is a fact which should be known.

A Sidebar on EMail

Before getting into the analysis of the material specifically about instant outlines, there are some misconceptions about EMail that appear to cloud some of this discussion. It might be useful for us to straighten out these matters right at the beginning, then we won't have to do so each time the misconceptions are introduced in the material which we will be discussing.

While the basic model of a piece of EMail as a message sent from a single writer to a single reader is certainly the prototypical case, EMail on the net has several other modes that allow it to assume quite different roles in different circumstances.

For example, we have at least

  • The Newsgroup;
  • The List-Server; and
  • The Archive.
Each of these represents quite different ways of transmitting information.

The Newsgroup

The Newsgroup generally behaves like a classic old-time bulletin board. Instead of sending an EMail message to an individual, you can post it to a Newsgroup that deals with whatever the topic of your message is. Typical newsgroups deal with (there are literally thousands, so this is only a tiny sample):

  • particular programming languages (perl, visual basic, python, ...);
  • hobbies (cars, hifi, ...)
  • technologies (GPS, satellite TV, ...)
  • history (computing history, civil war, WW2, ...)
  • ...

Once you find newsgroups that deal with things that interest you, they become quite effective information exchanges. For example, people with perl programming problems often write comp.lang.perl.misc and very knowledgeable perl programmers will usually respond with helpful (if occasionally derisive) suggestions. This information is of tremendously variable quality, ranging from the high-powered super-professional help one gets on newsgroups like the perl group mentioned above down to completely crazy hidden pitches for bizarre ideas.

Newsgroups can be either moderated or un moderated. Un moderated newsgroups generally publish all of the mail they receive. Usually this works well, but occasionally it can degenerate into destructive communications if some participants should decide to have a public argument, for example.

Moderated newsgroups have moderators who exercise control over what gets published. This may result in trimming some unproductive conversations, but can also impose some delays and dangers of censorship.

Moderated and un moderated newsgroups have different characteristics, but each type can be very effective in some circumstances. In any case, Newsgroups are one form of EMail that is distinctly much broader than pairwise communication.

The List-Server

The List Server is another structure for EMail communication. Here all of the participants agree to direct mail to the Server which has two major functions:

  • It maintains a list of subscribers; and
  • Copies all of the mail it receives to all of those on the subscriber list.

Generally list servers, like newsgroups, can either be moderated or not. In addition, list servers often offer the option of mailing out periodic digests of the incoming mail, as an alternative to individually copying each arriving message.

The Archive

Another alternative is to read mail when it is stored on an Archive. There are several different archives. Probably the best known of these is the very popular Google Groups which has taken over from Deja News, which served as a web-wide archive for several years.

An archive simply accumulates the mail which is sent to Newsgroups and makes it available in a historical record. There's about a decade's worth of mail currently available.

The purpose of this discussion is to point out that any notion that EMail is necessarily some form of interrupting. person-to-person communication is not necessarily correct. We can, and do, use EMail many ways.


Before we tackle some of the details, it is time for a general caution. One should be wary of instant anything, particularly if the matters at hand are intellectual. There is little evidence, I think, that people are prone to have their best ideas quickly. Quite often good, intelligent, consistent thoughts are a result of some reflection and contemplation, and not available to most of us immediately as we first hear of something.

So I would not expect instant outlines to be particularly intelligent or effective. Of course we are all growing up in an age of instant potatoes, instant rice, and instant news analysis, so it is not a surprise that there is some attraction to instant outlines. But, like instant potatoes, instant outlines may not necessarily be the most satisfying.

Yet it seems to be an implicit assumption of most of the people who are touting this technology that there is some positive value in the very speed with which this can all take place. As far as I can see, no one advances any evidence whatsoever of this proposition, and yet this assumption seems to pervade all of the discussion without any real justification.

I have had access to these kinds of facilities for a long time now, and I can unequivocally state that the ability to get such speedy feedback to change proved to be a hindrance to me, not a help. Of course, it would be presumptuous to assume that this is true of others. We are all different, after all, but while I can't offer any evidence that these kind of facilities are counter-productive for everyone, I surely can claim that they are counter-productive for some, at least me.

Why it's easy to misunderstand early experience

Probably the worst problem with the focus on instant things is that it is very easy for early experience to be quite atypical. To give an example, early Weblogs might be interesting not because there is anything particularly effective about the technology, but rather simply because the people who opt into the technology early may be particularly interesting people, and thus they have something unusually worthwhile to say. Early adopters aren't necessarily better, but it is quite probable that they are different in what may prove to be very important ways.

This was surely the case with many of our early judgments about EMail, for example. We got EMail in the middle 1960s, and thought it was the greatest thing since the proverbial sliced bread. What we misunderstood was that early EMail was interesting largely because it was written by us and our friends and we were all (to ourselves, anyway) real interesting. As time passed, and the community of people using our EMail facilities broadened, the quality of our EMail dropped quickly to approximate that which was roughly normal for our regular mail.

What do Others say?

There have been several commentaries written on the subject of I/O. Some of them take a view not far from the one that I am expressing here. I don't think it's necessary to use these opinions to pile on. A good example is the opinions expressed by Eberhard Lutz.

However, there are also several commentaries that suggest very favorable prospects for I/O. In this note we will treat some of these specifically. For the most part, I find the optimistic arguments untenable for reasons that will be made clear.

Robb's Analysis

John Robb wrote a brief analysis of Instant Outlining on his Weblog. I believe Robb works for Userland, so I guess it isn't surprising that he walks his party line, clearly erring on the side of favorable assumptions about the technology and its impact on our processes. But just because he has a `salesman's stake' in the outcomes, that doesn't---in itself---invalidate his message. So let's give it a critical look.

He begins with Here's my thinking on why Instant Outlining (I/O) and weblogs provide value beyond what's provided by e-mail and instant messaging. Both IM and e-mail are great tools for conversations between consenting individuals. Beyond that, e-mail and IM break down, and weblogs and I/O take over.

As we will see I find his analysis deeply flawed, and I think he rather completely fails to make the case stated here. But his thoughts have been regularly liked and favorably pointed to by other bloggers (mostly, not surprisingly I guess, those who work for or with Userland). So let me treat each of his points in the order he raises them.

Scalability and Overload

First, Robb treats scalability and information overload. He apparently believes that the solutions he is discussing scale nicely. Since I believe the opposite, some work is needed to reconcile / contrast our disparate views.

Everyone is facing information overload. There is too much information that the average person needs to know to function effectively. So how should you get this information? Right now, most people get it through e-mail.

Maybe this is just badly expressed, and he is referring to some particular kind of information. But taken as written, this strikes me as, at best, an odd view of the world. At least it is an odd (and incorrect) view of my world. I get a lot of EMail, and have for some time now, and yet my EMail counts for only an infinitesimal portion of my daily information intake. My typical EMail message is a few hundred bytes. I get a few hundred on a busy day. This is a total load of 10K to 20K bytes, or equivalently a few thousand words. A two minute segment of TV News is thousands (millions?) of times as much raw information. And I spend a lot more of the day with television, radio, newspapers, magazines and the like than I do with EMail. So the notion that most people get [this information] through e-mail seems quite wrong to me.

Robb continues: However, for those of us on the leading edge of online workflow, the volume of informational e-mails has exceeded our ability to parse it.

Being on the leading edge of online workflow is a dangerous business. I rather suspect that I may well have been receiving my first EMails (in the mid-1960s) when Mr. Robb may have been still in Grade School (In the interests of full disclosure, I should add, I suppose, that I don't know Mr. Robb, or have any idea how old he is, so this is pure surmise on my part, but the remark doesn't sound like deep experience to my ear).

So some of us have spent a lot of time with an awful lot of EMail, and have not still found myself half as overwhelmed by my EMail as I am, for example, by Annual Reports, Account Statements, Proxy Requests, etc. which often end up in piles whose thickness is measured in feet, not inches, and which arrive with alarming regularity in deliveries from the US Post Office.

Onward: Why? E-mail is a terrible one-to-many publishing tool. Not because the technology can't do it, it can, but because the volume of information published by an increasing number of publishers crowds out its basic functionality: conversations.

I suppose this could be true, though even in such a case I would find the point irrelevant. But it's not true, in any case. My EMail is, if anything, easier to filter than my physical mail. A fair portion of my `bad mail' disappears without any action whatever on my part directly from my inbox to the trash bin. And conversations are no more the basic functionality of EMail than they are of our regular mail. How many of the letters that you received this week that were delivered by the US Post were conversations? During the holiday season the number might rise to 5% of my mail (friend's Christmas Cards, greetings, etc.), but normally most weeks go by without any conversations though I do have to process piles of mail.

Finding a valid conversation in the stack of inbox spam from friends, co-workers, and nameless hawkers of "penis enlargers" is frustrating and increasingly futile. In contrast, weblogs and I/O provide publishers a place to put relevant information where it can be found by interested parties. It rationalizes the flow and allows it to scale. It is a parallel processing environment for the mind.

Finding valid conversations in life in general is both frustrating and increasingly futile. But that doesn't have anything to do with particular aspects of this technology. Across one's life's span, good conversations---in the sense discussed here---largely occur in inverse proportion to one's importance. When we are young we are all---except in the rarest of circumstances---unimportant enough that no one was to influence us through conversation. They occur without any hidden agendas or axe grinding. As we become more important, though, we not only acquire more money, but also more power and influence. And as a result we become the target of those who now view us as worthwhile targets for influence. We even have words for them: influence peddlers. So the quality of the conversations declines just as their frequency increases as others attempt to exploit and influence us.

So none of this, it seems to me, has anything to do with weblogs and I/O. Weblogs and I/O may offer a new place to put relevant information where it can be found by interested parties. But whether it is some intrinsically different kind of place than much older technologies: bulletin boards, newsletters, etc. still very much remains to be seen. And whether it rationalizes the flow and allows it to scale in any different way than many ancient technologies certainly isn't proven by the ad hominum statements made in Robb's piece. And how it is a parallel processing environment for the mind is quite incomprehensible to me.

But now it's time to turn to the next specific points.

Passive vs. active

Robb next develops the dichotomy between passive and active.

E-mail and IM demand my attention and my time (a dwindling resource) when I am least able to provide it. The tools force me to read something I am not prepared to read (granted, e-mail is more passive than phone calls).

In contrast, Weblogs and I/O leverage my time. They put me in control. I can batch process my interactions with individuals and groups.

This is certainly contrary to my experience. Of course I read weblogs when I want to. But I also only process EMail when I want to. My EMail arrives as it arrives, but it doesn't demand that I pay any particular attention to it unless I want to. For me, it's no different than someone else's Weblog.

And I think `instant messaging' (I assume that's the IM) can be left out of the picture entirely. It seems to me that IM is just a low-bit-rate form of conversation that might just as well take place on the phone.

Personally, I would much prefer the phone to IM. Not only is the bit-rate much higher, but I find considerable information in intonation and vocal shading. For me the only advantage of IM is that it offers the opportunity to masquerade, and that just isn't my thing.

There are many different modes of access which are allowed by some of these technologies. For example, I read several newsgroups by having them send me EMail. Others I choose to read by periodically scanning the newsgroups in Google Groups or EScribe or through other similar archives.

I can expand my circle of personal interactions and collaboration with little fear of being overwhelmed by the resulting interactions.

Why? Again, I don't get the differences and saying `it so' doesn't convince me at all.

For me, the ability to time-shift in a passive collaborative environment makes me infinitely more productive. Thinking in a massively active and interruption driven environment is like wearing a thought inhibiter[sic].

Again, this point seems to me to only be relevant if for some reason one chooses to treat EMail and other similar media as interruption driven. If you don't then the point is irrelevant.

So, while I do believe one can make an argument that the phone is intrinsically different (unless one is hearing impaired, or is wearing ear plugs) because sounds demand attention, it is incorrect to view EMail as different on the active / passive dimension from Weblogs.

Quality and complexity

Finally, Robb deals with quality and complexity.

Weblogs and I/O allow me to construct and publish complex thinking. Further, it archives that thinking so it isn't lost. The conversational nature of e-mail and IM make sharing complex thoughts difficult and more time consuming. It's hard, if not impossible to build a body of work that conveys a complex idea or plan. Additionally, I can't easily leverage previous thinking or the thinking of others to create a more complex work. The ephemeral nature of e-mail and IM is like thinking in quicksand.

Again, this is not a solid point. First, in the academic community, information has been exchanged in the form of papers for several centuries. On-line delivery of papers has been going on for at least several decades. We generally don't bother to exchange developed ideas in EMails. For the most part, EMails are used to carry attachments that contain papers, or that contain URLs or other pointers to the papers. And so it's untrue that, as Robb says, It's hard, if not impossible to build a body of work that conveys a complex idea or plan.


OPML is the Outline Processor Markup Language. It seems to be largely a Userland invention. By agreeing to this standard, outline information can be passed between processes. This makes it possible for processes to pass information about outlines around the net.

This standard is published and, by now, quite widely shared. It remains to be seen whether it is a good standard or not. My initial impressions are that it is not a good standard. But that conclusion will require a little more time to process before writing an appropriate critique.

Udell's Commentary

Jon Udell has written about I/O in Instant Outlining, Instant Gratification

Udell starts the discussion on I/O by looking at structured editing: -- but I think it's fair to say that structured editing is the exception rather than the rule. Certainly that's true for the application through which the vast majority of our keystrokes flow, namely email.

I'm not sure where this comes from. It is certainly contrary to my intuition about the state of things, but I must confess it is only a feeling---I don't have any real data to support my demur. I concede, of course, that lots of keystrokes flow through EMail, but I'm not at all sure that it's really all that many in comparison to the number that flow through other sorts of documents that we prepare. Further, again speaking for myself without claim on others' experiences, I would suggest that only a very few of my EMail keystrokes are spent dealing with my ideas. The vast majority are spent dealing with the trivia of everyday life.

Udell continues with Last fall at UserLand, the core development team began to use the outliner in another way: to communicate and collaborate. This instant outlining technique, now becoming available to Radio users, blends instant messaging, outlining, and blogging to create a new synthesis that Winer claims (and I hope) is a way out of the email hole we've dug ourselves into.

For some of us, communicating and writing from outlines is a time honored experience. In the 1970s during the heyday of the Office Automation Project at Wharton we made heavy use of outliners both in preparing our EMail and in helping with some of our automated writing tasks. It was quite useful, particularly for some of us (interestingly, it didn't work at all well for others), but in the long run it proved to be of no overwhelming consequence. Merely another small useful tool available to be carried in the bag of tricks that some found useful in getting their tasks done. And no contribution, at least as far as we could see, to the quality of the information that was being passed back and forth.

Picking up on a theme in other writings, Udell continues with In my book and elsewhere, I've cited a litany of email woes, which I would sum up in a single phrase: email is not a context-preserving medium. I'll have to track down the references to Udell's earlier stuff, as I must confess that I don't really understand this at all. It seems to me, probably naively, that EMail is just as context preserving as most of the other media that we deal with, but I'll have to do the research and report later to make sure. At least it is common practice when responding or forwarding to `mark up' the incoming document in such a way that the comments will make sense. That seems adequately context preserving to me, at least for most situations.

Is it important that: Communication in this environment is by invitation only, and two-way communication requires mutual invitation. Can't I easily effect the same procedure, should I wish to, with my EMail? Isn't that what unlisted phone numbers are all about. If you have an unlisted phone you are pretty much guaranteed to be able to control interactions by only giving out the number to people that you want to be able to reach you. Clearly, the same applies to these kinds of communications. If I want communication by invitation only it's easy enough to accomplish it with EMail. I just arrange to have an automated process throw away whatever I receive that comes from sources not on my invited to communicate list.

So while it might be Sayonara to spam, that's something that is easy enough to do with other technologies. It doesn't mean that there's anything at all new here.

After providing a useful and succinct word picture of the I/O process, Udell continues with: Transcluding content in this way is a long-overdue feature of the Web. What's especially stunning, though, is that the feature is here woven into an authoring tool that aims to replace email as the primary mode of communication in closely-knit collaborative teams.

This isn't the place for a long discussion of transclusion. However, I think it is the place to mention that before we attempt to replace email as the primary mode of communication for collaborative teams, we should take a moment to consider just how much of the activity of such teams actually has to do with processing ideas as opposed to the more mundane daily aspects of everyday project management. In my experience the core of the ideas are involved in about 10% of our project activity, and the process of managing budgets, arranging for speakers and their rooms, developing, handling meetings, etc. accounts for the rest and is a much larger share of the effort. This is likely to continue to be done by EMail anyway. So the question becomes not only:

  • Is the new technology effective in contributing to the solution of real problems; but also
  • Does the new technology contribute enough added-value to make it worth developing and learning.

Winer and his team don't email one another any more, and they claim radical productivity gains as a result of the switch to instant outlining...

Radical productivity gains are often claimed for new technology. But these claims are most often chimerical. There are too many possible reasons to go into in this context---we could easily write a whole paper just on these kinds of claims---but a few pointers might make clear that this is the case:

  • Productivity claims are often profoundly influenced by the Hawthorne Effect, particularly if they are made early in the use of a new technology; and
  • It is very often easy to `improve the productivity of' tasks which really shouldn't be done at all.
I'll leave this discussion there for later work.

Suffice it to say when someone claims huge productivity gains you should be suspicious. And if they make such claims in the first year or two of using a new process or procedure, you should be downright skeptical.

In one of the more interesting side remarks, Udell suggests There are subtler conventions that relate specifically to outlining as a way of organizing and communicating information. Outlines are context-preserving. When you save an outline in a certain state, that's the way it opens for your teammate. So it's good form to leave your outline in a state that establishes context not only for yourself, when you return to it, but also for the team with which you are sharing the outline. While Udell is apparently quite happy with this feature, it strikes me as a bad one, not a good one. One of the earliest lessons of teaching is that everyone sees things in different ways, and that if I communicate too much of my view of the problem to you, I hinder your problem solving capabilities, not enhance them. It may be helpful for me to give you some of my knowledge, but I don't want to give you so much of my view unless I have already solved the problem, and there's nothing for you to do. It is a phenomenon well known to the psychologists as problem set. And it hurts problem solving rather than helping it.

Udell concludes with What does all this portend for instant outlining? He is optimistic that this technology will provide some opportunities that are not, in his opinion, directly present in EMail: The key ingredients of that framework are structured storage (for content and metadata) and structured messaging (for data transmission, and publish/subscribe/presence notification). It's all XML, and really simple XML at that. If that's true, then it won't matter much whether the specifics of I/O prove to be hype or not. The other consequences of providing some form for organizing the material will win the day in any case.

A Touch of Curry

The more things change, the more they stay the same. an olde [sic] adage perhaps, but it still holds true with the Instant Outliner software the early addicts have been raving about.

This is Adam Curry's lead into a discussion of I/O. The problem is that the discussion really doesn't go any further, unless you want to count the re-telling of an odd story about Gopher, Marc Andreesen and a `headless Sun3 card in a rack above a chinese resturant in [sic] Virginia'

The one try at something useful: Intant [sic] Outlining is sharing thoughts. It's what you want but can't get from email. email went the way of the web...context is drowned out by html, flash, and lack of structure. All show, no go. My I/O has already replaced my email conversations with 4 people in one day

Again, different people, different experience. After early use of HTML in mail, for example, many of us gave it up years ago. Indeed I think it's still Usenet standard to send plain ASCII text because HTML contributes so little to solving the problems of communication by EMail, and it gets in the way of so many others. And the great majority of my EMail messages are one-liners requesting or delivering information. But then, maybe a single line message is just a very simple outline---if you want to look at it that way.


Wikis are a curious invention, and probably count as one of the few genuine innovations in the use of the web. The concept is simple, and the implementation seems to be effective. It remains to be seen whether it will also prove to be durable technology, but current indications are quite favorable in that direction.

A Wiki contains pages of information. Each of the pages has what is essentially a one word title, where that word is generally a run-together phrase. For example, `TheWikiWay' would be a legitimate title. Notice that also by convention there is unusual capitalization in the phrase.

This convention has a point. It is possible for an automated process to recognize such items, and to classify them as indexable. In this way any references to this title in other Wiki pages can be automatically handled. And we can easily have an automated process that makes lists of which Wiki pages refer to other Wiki pages, thus providing us with a map of the information structure of a whole plex of Wiki pages.

This makes writing a Wiki a rather straightforward task. You just write, and whenever to encounter a term that should (or might already) form another Wiki page, you simply give it the indicated form. The automated technology will then make the automatic links, or if there are none it will append a question mark to the phrase as a reminder that a new page should be created.

Wikis and Outlines

Wikis are an alternative to outlines. Actually, a Wiki might contain an outline, or, as we will see in a moment, a Wiki might be based on an outline-like structure. But, in general, a Wiki can represent a much more complex structure than the outline, which generally represents a strict simple hierarchical relationship. The relationship expressed in a generalized Wiki can be almost arbitrarily complex, and most certainly is not restricted to tree structures or any other simple paradigm.

Bowden's Remarks

Tony Bowden gives an interesting view of the relationship between Wikis and instant outlines. For the most part his remarks are quite sensible, but they do suffer a little bit from perhaps being a bit too instant themselves, not having been given sufficient reflection.

In a perceptive observation he notes: Whilst on the surface Wikis and Outliners seem related, they're subtly different. A wiki seems to have a much more fluid and circular structure than an outliner, leading to problems in trying to automatically represent an entire wiki as an outline.

And Wikis are great for complex thoughts. Lots of life isn't easily described by hierarchical trees, and Wikis are much better for them.

Perhaps certain sections of a wiki could be automagically translated to (quite bad?) OPML, but it seems to me, at the moment, that it would really need to be a manual job, slicing pages up and constantly narrowing the focus.

I think the main differences between Wikis and Outliners is deeper than the technology problems. It's to do with mindset and culture.

Wikis are great for expressing unstructured thoughts, and for discovering hidden connections. The first time you make a WikiLink in something you're typing and unexpectedly discovering that that page already exists is an enlightening moment into the Way Of The Wiki.

Outlines on the other hand require you to structure everything up front. An outline is great for reading, but not so easy to write.

And in what is some rare discussion for this topic, Bowden properly suggests that the process may be strongly influenced by the way that particular individuals think.

Perhaps software developers find it easier to think in structured ways? Good developers break things down into lots of little subproblems, each of which in turn is a series of other subproblems, etc. Some people can do this naturally. Others write in wider scope, and refactor into smaller chunks later.

A true knowledge sharing system needs to cater for both approaches. There should be the ability to jot down random thoughts (a personal weblog). To narrate your work (a personal outline). To record the structure of a project (a collaborated outline). To have a project calendar (a collaborated weblog). And to connect all these in interesting ways (a wiki).

Bill Seitz

Bill Seitz has taken an interesting, if not entirely successful, approach to this problem by merging outline technology into the world of the Wiki.

First, while I don't like the look of his pages at all, Seitz collects a lot of good links and references and is therefore always worth reading. Typography be damned, in the final analysis it is the content that counts.

Second, Seitz is particularly interested in both outlines and Wikis. As a result he is experimenting with merging the two cultures, and his web site shows it. So far I fail to have anything about his linking the two approaches strike me as particularly productive, but I am happy that he is experimenting with this approach.

What Wikis Have Taught Us

Wikis have taught us that collaboration with shared documents really works pretty well. As a result the aspect of I/O that requires `messaging' is not as important at it might appear at first. Further, the richness of the Wiki can cover some problem areas much better than the more strict hierarchical orientation of the instant outline.

All of which leads me to wonder if Wikis that contain outlines, without any further special fuss, aren't a far superior solution to the problems being discussed here than the rather more restricted facilities of I/O.

Problems with Outlines

And there are serious problems with outlines. They are not a superior technology, rather they are simply an alternative technology. Good for some, and not so very good for others.

Let's look at some of the problems.

Outlines as a Panacea

First, in many situations outlines are treated as a panacea. Indeed for some of the people who use them they may seem to be.

But, quite plainly, they don't really work well for everyone. Think through literature as an example. Some great works can be neatly and rigorously outlined without distracting from their beauty. Other works would absolutely and completely defy any such attempt at organization.

That's just a reflection of the way that it is in the world of ideas. Some good ideas are hierarchically organized trees of knowledge. Other good ideas are complex messes that form a rich tapestry.

Outlines as a Placebo

Second, in some circumstances outlines are a placebo. My scattered and vague ideas may look much more profound if I have expressed them in a seemingly organized structure. The very expression of organization can lend weight to the appearance and therefore to the effect of ideas.

Outlines as Authority

Outlines are rather a rigid structure. Therefore, we might at least hypothesize that they might be better for expressing authority than for promoting collaboration.

Again, only time will tell. But I do find it tantalizing that in his own descriptions of using I/O to manage project development at Userland, there seems to be a fair authoritarian air about what is going on. Winer's own descriptions of how interactions with the staff are managed leaves me little doubt about how authority flows at Userland, and I wonder if this is separable from the fact that he finds I/O so productive in this situation.

Outlines for Everyone

One of the most consistent mistakes made on the net is the assumption that what works for someone will necessarily therefore work well for others. This simply isn't the case.

Some people think in outlines. Others don't. There's no right way, just different ways. Hierarchical thoughts fit into outlines nicely while non-hierarchical, non-linear thoughts might well not.

Why not Wiki Structure?

Would Wikis solve most of these problems? It's hard to say. I must say, however, that at the moment, the prospects for finding a useful solution via Wikis is much better than the prospects for finding one using the technology of instant outlines.


I am very pessimistic about much coming from the whole notion of instant outlines. It's a two stage problem. First, outlines strike me as relevant and helpful technology for only some portion of the population. How small this portion is still remains to be researched and understood, but it would not surprise me at all if the number is quite small.

Second, even if outlines were useful in these situations, there is little---I think---to be gained by sharing them with haste. I like and use outlines a lot, but while I am forming my outlines they are highly volatile, and communicating them all too quickly would not serve either my interests or those of my collaborators very well.


Winer's original: http://radio.outliners.com/beta

Udell's comments: http://www.oreillynet.com/pub/a/webservices/2002/04/01/outlining.html

Tony Bowden: http://radio.weblogs.com/0106165/2002/04/04.html

Bill Seitz: http://webseitz.fluxent.com/wiki/InstantOutlining

John Robb: http://jrobb.userland.com/2002/04/04.html

Eberhard Lutz: http://www.e7l3.com

31 March 2002: Live Blogging

Climbing mountains because they are there is a time-honored human activity. But while the view from most mountain tops is at least stunning, doing other things just because we can may not fulfill such a lofty purpose.

There's been some discussion of live blogging recently. The thought appears to run this way:

  • We have had laptops that could be used in the context of conferences and meetings for some time. They are unobtrusive enough that they can be actively used while events are taking place without detracting from those events much.
  • The advent of wireless access points now allows live linking to a LAN and/or to the web. No messy cabling or other direct link via wire is needed.
  • The existence of the Blog means that documents can be published directly on the net in a matter of minutes or seconds.
  • People who read the blog can respond to it via EMail.
  • The EMail can be delivered through the net and the wireless link directly to the laptop of the person sitting in the meeting. Of course, this completes the whole circle. Information has gone from inside the room to the outside world and back again.
So now we don't even have to wait for the afternoon edition of our newspaper for the latest information. Not only can we get our hands on it quite immediately---something we have been able to do with modern radio and television coverage for quite some time now---but we as readers/viewers are able to respond, and provide feedback with a speed that would have been effectively impossible in an earlier age. Thus we have now climbed the mountain of immediate response. But just because it's there, that doesn't mean that climbing it is a useful or edifying experience. Let's see what we can see. Take an unkind description of the same sequence of events:
  • Start with an event which will make no useful news in any case;
  • Place amateur stenographers at the event with live computer equipment;
  • Make sure they are so busy struggling to write down what is happening that they can't have a useful thought of their own; and
  • Claim this is some kind of new journalism that represents a `great leap forward'.
Same events. Different descriptions. Purpose served: not much.

You're Dissimulating When

Poker players know about tells. Tells are aspects of behavior that reveal the truth which underlies the appearance, even when someone is attempting to keep that truth hidden, or when they are actually dissimulating even to themselves. There are several tells that indicate live blogging is not likely to be either useful or effective.

No `Real' Press

First, most of the events that are live blogged are unimportant enough that few members of the regular press are likely to show up. That should be taken as a clue that there isn't much news to be made at the conferences, and that it is at best at least unlikely that someone needs to report any of it. Thus trying to get news out quickly when there's not really very much news being made is swimming upstream. Notice that for some short while, live blogging itself will be the news. But that's just because any new technology that shows up adapted to a new problem will provoke some sort of comment. Unless some real value-added is delivered, this interest will not last long.

Need Immediate Distribution

Second, the material presented is not likely to be so time sensitive that breaking the news anytime within the month is important, much less any time in the week, day, hour or minute. Presentations at most such events either tend to be thinly veiled ads for a product or service or exercises in self-promotion, neither of which need be brought to the attention of the broader community with undue haste. It doesn't generally matter much to us when we hear most ads, or when we listen to self-promotion. The impact isn't very time sensitive.

Don't Let any Useful Reflection Interfere

Third, by occupying the attention of the bloggers with the task of recording the proceedings we keep their minds from being devoted to processing any of the information that is being presented. Few people can keep their minds on several different things at the same time, and most live blogs show every sign that the more faithful the bloggers are to getting the words being uttered down in the blog, the less time they have available to give any thought to the questions that the material raises. Indeed, they often do not seem to have time to exercise even the most rudimentary of judgments about what material is worth passing on. Perhaps it is worth picturing, just for a moment, the extreme blogging event where all of the audience was actively involved in the task of blogging. One can easily imagine that the interactions left in the room would not be very interesting.

An Audio / Video Record

Fourth, if an accurate record were of any importance, we have had technology available for many years that could handle the problem. Simple audio recording is adequate for NPR, and presumably if the density of information presented at conferences were anywhere near that important the same technology could be employed. Indeed, if it were useful one could even have a video record of the proceedings. Of course, most conferences are so visually dull that this would probably have only soporific value, but occasionally a performance like Ballmer's Dance, Monkey Boy, Dance or one of Jobs' absurd Dog and Pony shows might have sufficient value to make a video record at least entertaining if not useful.

Where's the Beef?

What are we supposed to gain from these efforts? Surely if we look at what we have seen so far, there are no substantial insights. If we needed to be at the conference, we'd probably have gone. If we couldn't go, and still wanted the information, perhaps an audio record or some conference proceedings might be nice. Most of the people making presentations have an interest in getting you to pay attention, so generally they are willing to go to considerable trouble to get you up-to-date information about whatever was the subject of their discourse. So you're unlikely to miss much.

Who's In Control

There has been a suggestion that somehow the quality of the event will be directly influenced by the presence of the technology in the room. The technology can mediate between those on the dais and those in the audience. And in this way the audience will be able to exercise some new influence over the course of the presentations. Of course we have always had paper available to perform this task, and it only rarely seems useful, so that doesn't exactly augur well for the technology being important. But that's only part of the story.

Who's in control anyway? Is it the audience or presenter. The few times that the formal part of conferences have been useful to me are distinguished by the fact that they have invariably been situations firmly in the control of the presenter---generally paying no attention whatsoever to the audience. If the audience had much to contribute, they should be presenting.

It is a situation not unlike the classroom. The classes I found useful had teachers who didn't ask the students to vote on what they wanted to learn. Lots of good learning may have taken place in much less structured, more informal groups, but these groups generally met in cafeterias at the Student Union or in Fraternity Tap Rooms. And none of the interactions would have been facilitated, as far as I can see, by some elaborate linking technology. The best technology I know of for such kinds of purposes is that available in the US Congress, and we all know how profound the floor interactions are in our House and Senate.

Pander to Populism

So, I find it unlikely that increased communications capability in the audience is much of a contribution to the forward progress of intellectual discourse. Rather, it strikes me as being much more tales told by idiots, full of sound and fury, signifying not very damn much. And the more the audience is able to inject their interests and focus into the discussion, the less effective I'd expect it to be. It seems to me that there is very little evidence to suggest that playing to the audience ever makes things much better, at least when the purpose of the activity in informative or intellectual rather than pure entertainment.

The Solitaire Measure

Up to now, live bloggers have had their hands full simply getting their live networks to work. We now seem to be moving past that set of problems into an age when regular and reliable communication is mundane. My prediction is that what we will now see is a gradual increase in the live machines at conferences which are devoted to the time-honored tasks of modern day computing: reading EMail and playing solitaire.

The fact that an increasing number of machines at these conferences are now linked on line should not be misread as an indication that the technology makes an important contribution to the business of the conference. Instead, it should be read as a sign of just how boring and inconsequential the usual material presented at the conferences is.


Live blogging, like the new journalism from which it grew is a flash in the pan. It is a technology in search of a problem. It is looking for our lost key under the lamp rather than where we lost it. Of course if Hemingway happened to write the blog it would surely still be worth reading, even if the technology had nothing to contribute to the process.

David Ness' summary of work can be found at http://mywebpages.comcast.net/dness

This document collects the articles that earlier appeared in Top of the Heap.