September 5, 2007

A Response to Comments on “The Tagging Growth Curve”

Before I reply to comments from KatB & Simon Edhouse on The Tagging Growth Curve, a quick reminder that this series of postings has as it’s primary point of departure the idea that the smoothly drawn analyst’s curve for describing technology growth misrepresents the noisy and unpredictable fluctuations of reality. The outcome is that this conversation is as much about the models currently in use in the technology industry for framing how technologies grow, change, and spread, as about the specifics of tagging.

At some point, it may be necessary to branch these topics in pursuit of further clarity; on either, or both in relation.

@Simon: Incremental growth along these two paths is what we’re seeing at the moment. The apparently slow increase in the size of the total population of people and organizations who’ve adopted tagging is consistent with the idea that innovation happens during those periods when the rate of diffusion is slowed, due to the increased requirements of crossing a community boundary.

A partial list of those requirements would be:

* lower cost
* rising user experience quality
* social facilitators: advocates, ambassadors, evangelists
* conceptual bridges – like ‘horseless carriage’ – to overcome the increased friction of reconciling the new and different with an old frame of reference lagging in terms of awareness and understanding of tagging

The graph charting the curve for the rate of innovation – if it’s possible to chart something as slippery as innovation – would likely appear as the inverse of the growth curve. A curve for innovation would show spikes (for quality of innovation, quantity of innovation, or both…?) during the intervals when the spread of tagging is slower.

Anyone who wants to understand the mechanisms, rates, trajectories, etc. of technology growth should really look at diffusion and innovation as they inter-relate. Doing otherwise seems like trying to understand population sizes based only on the single factor of births or deaths, but not both working together.

@KatB: Technology innovation is definitely a social process – good of you to surface that angle explicitly. Social tagging is – well – inherently social. Taking that for granted (which says a lot about my initial frame of reference), I brought in the biological model without mentioning this given.

My knowledge of the Technology Acceptance Model is limited. What makes it a good model for the new explicitly social technologies? As opposed to the previous (non-social) technologies prevalent when TAM was formulated?

My goal in suggesting punctuated equilibrium as a better model than the Gartner Hype Cycle was to explore the fit of a different way of looking at tagging than is customary for the broad IT frame of reference. That does not discounting any other kind of model directly. There may be fundamental conflicts between one or more of the various models on the table that I’m unaware of at the moment, but then I’m unaware of them :)

Another question to ask in testing how well punctuated equilibrium fits as a model for diffusion is “Does punctuated equilibrium apply to open source technologies?”

Ravi of Luinuxhelp posted a mind map of the lineage of linux distributions in April of 2006 that surfaced in the open source and linux communities. The updated version of that map looks like this:

And these two images (from Patterns and Rates of Species Evolution by Michael J. Benton) illustrate speciation as it appears under punctuated equilibrium; first in general, and then specifically for bryozoans in the Caribbean.

We should keep two things in mind while considering these evolutionary charts: the illustrations come from completely different frames of reference; and apparent similarity – visual or otherwise – is no guarantee of genuine similarity on any level. But there is still compelling likeness in these renderings of the lineage of an open source OS, and “tiny colonial animals that generally build stony skeletons of calcium carbonate, superficially similar to coral”. Clearly, in the real world, linux distros and bryozoans are wildly unlike. Yet their evolutionary trajectories may show some of the same patterns at this level of abstraction.

Which means that punctuated equilibrium as a model might have something to say about open source software in general. [Here I have to say that informed contributions from the linux community are welcome, as I'm not qualified to discuss it's workings in any detail.] And if such is the case, then punctuated equilibrium seems to correspond to the known patterns of diffusion and evolution in the two major spaces in which software technologies evolve at the moment (until another production model arises) – commercial, and open source.

And just in case there’s any question about my professional qualifications for discussing evolutionary mechanisms, it should be clear I am in no way a sociobiologist, ethologist, geneticist, evolutionary biologist, morphologist, etc. So I’m borrowing freely from other fields to seek a new model, which means I run the risk of borrowing concepts from either (or both) in ways that don’t make sense in their original contexts.

4 Comments

  1. Joe, it seems strange to make a new post as a comment on comments, but, to each his own…

    You responded to my comment saying: “Incremental growth along these two paths is what we’re seeing at the moment.”…However in the original piece, you talked of what we can “expect to see” and how the trend “will follow these two paths to varying degrees”. In other words, you were making predictions.

    My point was that the trajectories you were predicting seemed to be incremental rather than disruptive. In observing this, I am not implying that you should have predicted a disruptive path, but rather that by not including this as a third option, it is perhaps conspicuous by its absence.

    Also, I don’t understand the linkage you are making between tagging and the Open Source/Linux development community. ~ If there are some tag-orientated projects being developed in this way, what is the significance of that to tagging? Is it incidental, or do you see a particular nexus between tagging as a form, with so called open-source methodologies?

    Comment by Simon Edhouse — September 5, 2007 @ 7:58 pm

  2. Simon,

    FYI – I mistakenly published a draft version of this posting. [Allow me to note that at the time, I was literally sitting in the middle of a five party argument about whether or not my desk had been assigned weeks ago to another group in an office re-org, meaning I needed to relocate immediately to keep my client's physical space move on schedule. Ah, the joys of consulting...]

    The discussion of fit between open source software’s evolution and the punctuated equilibrium model was intended to address KatB’s comments rather than yours. I’ve moved it accordingly. That and some additional updates should make this correctly published version clearer overall.

    On to your comments.

    My contention is that rather than being seen as an over-hyped new technology with uncertain commercial prospects, as is typical when using Gartner’s Hype Cycle frames, tagging is currently in a period of apparent quietude in growth that reflects genuine innovation ‘below the waterline’. (See Gene Smith’s update on recent changes at LibraryThing, titled Is Tagging Stuck? Hardly. as an example of ongoing innovation that supports this view).

    In fact, tagging is currently at one stage of the natural growth path for a(ny) technology that we can understand more completely using punctuated equilibrium as a model. This model, borrowed from biology, seems to apply to both the commercial path of software evolution, and the open source path. The model allows us to make some general predictions about how the growth of tagging will continue. And so I wager a vague guess about continued incremental change over the next 12 months, but wouldn’t put my retirement money on the line. (Maybe someone who reads this blog can set up a prediction market for us…)

    To the extent that I’m using it right now, punctuated equilibrium as a model for the evolution of tagging only discusses the rate of growth and change for tagging. I’m not using punctuated equilibrium to address the question of whether tagging is or will be disruptive. In predicting incremental trajectories, I’m following the model based on my understanding of where tagging is at the moment.

    The question of the current state of growth and development of tagging is – at least at this stage of the conversation – independent of the question of whether tagging itself is disruptive, or is part of a larger disruption. The two questions address different levels of the environment; tagging in specifc, vs. the overall state of the ecosystem. As I see it, whether or not tagging is a disruptive innovation depends on changes at the level of the larger IT ecosystem, and not current state (fast growth or slow, innovation or diffusion) or the path that tagging follows. Tagging could follow the open source path, the commercial path, or both, and end up being either disruptive regardless.

    You (seem to) mention the possibility of tagging following another disruptive path here:

    My point was that the trajectories you were predicting seemed to be incremental rather than disruptive. In observing this, I am not implying that you should have predicted a disruptive path, but rather that by not including this as a third option, it is perhaps conspicuous by its absence.

    What would a third / disruptive path look like – one neither open source, nor commercial, but something else? Dion Hinchcliffe talks about the mashup as a possible new model for software creation in Mashups: The next major new software development model?. Is this an example of what you had in mind for a disruptive path? (Briefly returning to punctuated equilibrium as a model, it seems the same pattern is in effect with the growth and innovation of the mashup; this is a stage of rapid innovation, and very small diffusion / adoption.)

    Regarding whether tagging will prove to be disruptive to the IT ecosystem and how we might determine that, I have another posting in progress that posits three major characteristics of disruptive innovations. I’ll try not to release a draft by mistake…

    Comment by joelamantia — September 6, 2007 @ 1:39 pm

  3. Joe, understood. ~ I agree with you that tagging is not stuck, as such, but perhaps in gestation, to return to hype status yet again at another time. However, I think its a mistake to refer to tagging as a ‘technology’, it is not.

    It is in fact many things to many people, but really is a form of ‘markup-language’. As I said in an earlier comment to another post, I think the tagging phenomenon on the web really began with html tags, but prior to that, I think the ‘head of the stream’ really occurred around 1967 in Ottawa with William Tunnicliffe’s presentation called “The Separation of Information Content of Documents from their Format” which led to ‘markup-code’ in the publishing and printing industry, then evolved to html in 1989, courtesy of Tim Berners-Lee.

    So, tagging is a highly dynamic variant of markup language, therefore it is essentially ‘linguistic/semantic’ in origin (and is still) and not a technology as such.

    As to how the ‘disruptive path’ for tagging may manifest itself… I don’t see that as necessarily being non-commercial. I think tags can be part of a disruptive innovation in the commercial sphere.

    Comment by Simon Edhouse — September 7, 2007 @ 8:42 am

  4. Joe, understood. ~ I agree with you that tagging is not stuck, as such, but perhaps in gestation, to return to hype status yet again at another time. However, I think its a mistake to refer to tagging as a ‘technology’, it is not.

    It is in fact many things to many people, but really is a form of ‘markup-language’. As I said in an earlier comment to another post, I think the tagging phenomenon on the web really began with html tags, but prior to that, I think the ‘head of the stream’ really occurred around 1967 in Ottawa with William Tunnicliffe’s presentation called “The Separation of Information Content of Documents from their Format” which led to ‘markup-code’ in the publishing and printing industry, then evolved to html in 1989, courtesy of Tim Berners-Lee.

    So, tagging is a highly dynamic variant of markup language, therefore it is essentially ‘linguistic/semantic’ in origin (and is still) and not a technology as such.

    As to how the ‘disruptive path’ for tagging may manifest itself… I don’t see that as necessarily being non-commercial. I think tags can be an integral part of a disruptive innovation in the commercial sphere.

    Comment by Simon Edhouse — September 7, 2007 @ 5:30 pm

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.