Addendum, March '07: This piece has been garnering some interest again lately, primarily due to a ruckuss kicked up here and here. Unfortunately, I now find this write-up a bit cringe worthy, with a pseudo-intellectual veneer that is downright unpleasant. I won't have it nuked, because it says something I think e2 still needs to hear, and I completely stand by the main thrust of the argument, but dear god I wish I'd written it in a different way... So - this write-up contains elements that could be interesting, if done differently, by someone else, better.
How to make a slime mould sick
A structural interpretation of E2’s current malaise
This writeup is written at a time when E2 appears to be in the midst of disputes over its purpose. It is the impression of the author that these arguments may be missing one of the fundamentals of E2's nature: a forum where anything an individual chooses to contribute will be published and judged. The question has come up continually in the recent past: what should be the purpose for E2? It is the contention of this author that these arguments orginate in a fundamental misunderstanding of how the E2 site currently functions.
The author assumes that the site architecture allows virtually anything to enter E2 at the front end (as it does now and will not do under The New E2); this essay is aimed at discussing the incentives and disincentives that apply to users and their contributions, the structure of the database and how these combine to shape the E2 contribution base.
With his hat now firmly in the ring, the author's admits to an agenda. I believe the problems E2 is currently experiencing are produced by a combination of increasing database size, incumbency within the userbase and database, a rigidly structured voting/experience system, XP inflation, and the resulting effects these have upon the userbase. I will here attempt to justify this argument in terms of some basic systems theory concepts: incumbency, contingency and Emergence.
Other commentators have discussed how E2 can be viewed as a self-organizing system. In this model, much like a slime mould or life on this planet, a chaotic system of disunited simple components can grow, through common underlying biases, into an organized system with order lacking in its parts. Steven Johnson discussed in his book Emergence, how E2 self-organizes through the nature of its Voting/Experience system. The voting mechanism has produced an editorial body of thousands from the disparate components of a global userbase. It’s an inspiring demonstration of anarchy in action, and I am in love with it.
E2 is not only managed by this userbase, although it numbers thousands, there are administrative staff with powers above them. A quick comparison of the powers of users and administrators should demonstrate that the administrators fufill a role more akin to “gardeners” of the database than managers. There is no fundamental difference to users. The Blab button means that editorial commentary comes as easily from users as administrators. The vote system is the primary organizing mechanism of the site, and forms the main incentive, especially through merit, for users to contribute. The user recommendation system, The C!, forms the main database access mechanism other than searches, the softlinking system forms the main cross referencing system. All the above are user driven.
Site administrators can wield the knife, but the verdict is the key to the E2 entity's response, the knife is punctuation. The administrators are more active than users in moulding fresh blood, but in this they act as value added users, members of the user base with the benefit of rank. Further, site administrators are so numerous and disparate that they act as a slime-mould themselves, governed more by underlying biases of the system than policy. It is therefore suggested that the current sport of venting at the site admins for E2’s malaise is like blaming the janitor when a building falls down. The issues run deeper, if E2 is failing, its a problem with the underlying structures, this slime mould has caught a disease.
Looking for the problem at E2, the slime-mould, the first thing to look at is the Voting/Experience System, since it is the key mechanism of self organisation. This system forms the main incentive for contributing to E2. Votes form the roughest index of how your work has been viewed by others, and accompanied with C!'s and direct feedback are the essential functions of the E2 forum.
The voting system was erected when E2 was first formed, and much like any currency, its value has changed over time. When the voting/level system was instigated E2 was empty and needed content more by volume than quality. Short, low XP writeups were the purpose, particularly since E1 had a maximum writeup length, giving the whole database a built in short writeup incumbant. However, the voting experience system was erected to absolute values rather than relative indexes, much like the gold standard, this was an error. As the database filled the incentive to add content changed from contributing short writeups, filling in E2's gaps, to writing contributions that stood on their own merits. This resulted in a devaluation of the vote currency, and an attempt was made to compensate for this with the merit system. Again we see the action of incumbency as a system built for a specific purpose stayed long past its time. The merit system was a contingent cludge, allowing only a doubling of progression rate, and the system remained linked to contribution by volume. It should be apparent to all users of E2 that this site is no longer about volume by user, so the retention of this system is an anachronism. Potentially worse, the merit system was in part based on average merit per writeup, actively dissuading new users from writing material that could reduce this average.
The other key feature in my model of this site's change is the increasing size of the database. Note should be taken here that I am mostly disregarding administrative policy as a driving force behind E2. Since all users have the potential under the site architecture to add content on any issue they choose, including editorial policy, what the slime mould doesn't like it will delete. What has happened as E2 filled? Each user encountering the database is bound to look at pre-existing content to influence their style, and pre-existing editorial comments for primary influence, the voting incentive gives a built in desire to conform to the median in order to achieve the rewards. As the database expands in size there is a clear pressure for stylistic conformity, driving the emergence of the E2 "house style". Further, as the database fills there is less incentive to contribute short works towards the project and more to write extensively on indivual's areas of expertise, or fiction. Also, under the slime mould, quality floats. The best material written keeps coming back to the front page, causing the database to have a permanent influence on new users. I propose that the biggest cause of the ever rising bar is the near permanent presence on the front page of the best works of the fled masters. It should be pretty clear to any user looking at the front page that E2's content has very little relationship to everything now, showing a strong bias to emotional content and in-house humour. I believe this is driven by the interaction of a specific type of userbase and a huge database.
Taking these 2 processes as the only significant changes to E2 over time (constant architecture and "gardener" admins, leaves only a changing userbase), what has caused the disease? I cannot overrate the importance of the voting/experience system in my interpretation of E2, and therefore when this system began to fail I believe E2 began to collapse. The requirement for volume posting means that E2 requires a massive time contribution from users in order to progress. With the increasing bias towards length and quality, this meant that E2 users were judged on their time contribution. This mechanism selects for users with disproportionate amounts of free time: the young, students and similar. By leaving anyone with less than 25 writeups without the franchise, (which grew to mean a minimum of 25 hours contribution) active human beings less committed to the project had no influence on the direction of the slime mould. E2 therefore came to increasingly represent the interests of those who could manage the necessary time contribution. The voting/experience system meant that a successful professor or judge, world authority on their field, who wished to contribute on their area of expertise would be unable to influence the E2 direction for 25 hours, when in all likelihood they had lost interest in the project. Assuming a perfect merit (at least 1 hour per wu) it would require 75 contributions to acquire the C!. Since the Cool list on the front page is the primary introduction of new users to the database, E2 effectively became a dominion for its established users, who had been here when contributing required less time.
Wikipedia, without E2’s user hierarchy, has free access to all the experts of the world, those unable to manage the time contribution E2 requires. With Wiki's emergence E2 changed again. As the userbase became more biased towards those who, for whatever reason, had more than average spare time, this actively selected against time-poor users and more emotional content became dominant on the front page. Slowly E2 has begun to lose its established userbase. I believe E2 has now declined sufficiently that, It no longer acts as a slime mould. Without enough users with sufficiently varied backgrounds, new factual contributions can no longer be judged on their own terms, thereby not getting the votes, and leaving the user without the incentive to contribute. This is a very dangerous time for such a project because there are now so many selective forces pushing for mediocre, emotional, teenage content that E2 is almost incapable of breaking out from its database's biases. The new users are not rising up the voting system to maintain self organisation because the time contribution required has become too high. It has regularly been voiced that low level users feel disadvantaged, but the effect this has on the database has not been discussed, E2 needs to have as many voting users as possible in order to function!
It is not the intention of this piece to suggest changes to E2, I realise there is a heirarchy and this is viewed as disruptive. However, if the analysis provided above is truly the cause of E2's difficulties there are many ways to correct it. The author would like to propose that these changes should be structural. The pivotal change would be to make the voting system a more efficient currency. There are a multitude of different ways that this could be done.
This is my most outlandish suggestion: there has been an admirable increase in the non-contributing user base over the past 6 months, but these do not play a role in the feedback system and therefore remain invisible. If E2 were to give the franchise to everyone who logged in, this would fundamentally change the nature of E2. When has giving more people the vote ever been a bad thing? The XP peaks would get higher, the troughs lower, but nothing would fundamentally change. In one simple step the slime mould would go from an editorial body of hundreds to an editorial body of tens of thousands, drawing on the experience of the whole world. This is a kill or cure solution to the malaise.
Fundamentally, any tinkering with the XP system would improve the situation by increasing the energy of the slime mould. Thus, may i suggest annually indexing the merit system: established users mostly disregard this feature although it is vital to newcomers, since it averages out over a large number of contributions. By making it annually indexed, established users would have to maintain a continuous stream of high quality writeups in order to maintain their level. Or lowering the franchise to 10 writeups (a dirty cludge to a system that needs a total overhaul), or even making XP a real currency, and setting levels indexed to just XP without writeup volume.
In brief commentary on The New E2. The introduction of publishers is a potential threat to the slime moulds nature, this should be considered before this is acted upon or else the database will become a square peg in the round hole of The New E2.
E2 is not dead, nor is it likely to be soon, but we are reaching a dead end with our current userbase structure. E2 must open up in order to grow.