How can code work too well, you ask? Because even the simplest game theoretic considerations can make the results of administrative actions counterintuitive for those who don't examine their intuitions closely enough. On this occasion, Webring.com failed to do just that, to the considerable hardship of many of its users.

The ssnb javascript code worked too well because having it up and running made the process of joining new rings at Webring severely easy for the prospective new ring member, once his ssnb was in place for a page. All he had to do was go to the hubpage for the ring, click on the link marked "join" in the upper right hand corner of the page, and run through a short menu by point and click. He literally wouldn't even have to type anything; just click on a button which would bring up a menu of choices of urls he already had created code for, click on the url he wanted, and then click on a button on the page that came up indicating that yes, he really did want to join the ring he was applying to join. The whole process would take five seconds - maybe 35 seconds, if the system was running slowly because of heavy traffic and one's key punching was limited by some sort of partial disability. (Yes, hi). What I just described sounded like a well-designed, smoothly flowing system, did it not? What could be so bad about that?

Because systems have to be used by human beings, and human beings are free agents, freewilled individuals who have to deal with the realities of imperfect communication and their uncertainties about the level of cooperation they can hope for out of each other. Any system whose designers fail to take that fact into account is a system headed for disaster, one which an ingrained corporate arrogance will only deepen.



When one had to cut code for each ring one joined, download each graphic individually, and then write to each ringmaster to notify him that one's site awaited inspection, one was looking at a process that took maybe a minute or two, nothing too oppressive for those who wished to join only a ring or two, but a major deterrent against the practice of joining hundreds or thousands of them. What Webring.com did was remove the deterrent, without thinking about the consequences of doing so. At first, as ingrained patterns of user behavior were still ingrained and the code was still unreliable enough that many users were reluctant to use it, the number of memberships per site one tended to find remained low. As time went on, however, circumstances changed.

Some of us who liked the reliability of the HTML code and didn't want to switch over, found ourselves getting letters from new ringmasters who threatened to throw us off of the rings that they had just taken over, unless we switched over to SSNB usage. This was a factor that a webmaster could ill afford to overlook, because Webring had become very casual about taking rings away from its ringmasters, which it now called "ring managers" to signal the fact that the one time ringmasters on their system were no longer seen as having any kind of ownership over the rings they had created, or had continued work on after taking over for somebody else. To decline to join a ring on which the hated "No HTML" policy was in place was easy. To be certain that none of the rings that one was already on wouldn't be taken over by somebody who had taken that aggressive stance was impossible.

With the amount of churning Webring.com was doing, throwing rings up for adoption as frequently as they were, the entry of a new ringmaster into one's online life became commonplace, something that was almost a daily event. If a certain percentage of those simply tossed any site with HTML code off their rings on taking over, what that would mean for the webmaster using such code would be that is ring memberships would face a high rate of attrition, once that would rise with the rate at which rings were put up for adoption. Think of it in terms of the well-known game of Russian Roulette, that entertainment for the depressed in which one spins the barrel of a revolver, having placed one bullet into one chamber (the other five chambers being empty) points the gun to one's head and pulls the trigger. The concept is a familiar one, yes? One knows that a player will probably survive any given round, and yet one also insitinctually knows not to expect to run into any veteran players. Why is that? Because we know that the odds will catch up to one.


5/6 * 5/6 * 5/6 * 5/6 * ...

will go to zero as the number of factors in that product goes to infinity, and that product is the probability of one's survival of a series of rounds - that product, with the number of factors equalling the number of rounds. The fact that we do not hold Russian Roulette championships, then, is not a sign that the games have been fixed by some malevolent interloper who has acted to turn a one in six chance of mortality into something higher, but just simply a reflection of the realities of how probabilities work, along (one hopes) with a general lack of love for the thought of somebody losing. With obviously much less dire consequences in the case of a loss, one should find that the exact same probabilistic principle applied in the case of our ringmasters, and if the lack of mortality issues should keep us from carrying what was already an established understanding over, that such a reaction is an emotional one, not a rational one, and thus in this case not an informative one.

Even if the new ringmasters who were willing to penalize users for refusing to switch over to SSNB were nothing more than a noticably large minority, the minority would rule if churning was rapid enough, because just as the Russian roulette player would eventually have the bad luck to roll up that one chamber with a bullet (as outnumbered as it was by friendlier vacancies in that gun), a webmaster would eventually get blindsided by a new ringmaster who was not going to adopt a live and let live attitude toward the use of HTML code on his ring. "No sir!" More than a few of them were just that evangelistic about the whole silly business, some of them quite strident and selfish about the whole matter, even when SSNB was proving highly unreliable, with navbars mysteriously refusing to appear on site on which the code was correctly installed, those sites then failing both visual and automatic check on the part of the new ringmasters. SSNB offered the promise to some ringmasters of being able to update their code as needed without having to deal with foot dragging from individual members, and if that meant that some members ended up greatly frustrated by malfunctions beyond their control from time to time, so be it. SSNB was the latest new thing, and holdouts against modernity weren't to be tolerated.

Sighing, many of us broke down and installed the new SSNB code thinking that at least we would be saved the periodic chore of cutting out code for yet another ring we'd been thrown off of for using HTML, and at least SSNB did have the virtue that the navbars for the rings one was thrown off of due to system failure would vanish on their own. Someday. Eventually. Heartened by this though not one bit, we cut, pasted, posted, and hope to get back to doing something productive now that we were just going to be left alone. We hoped in vain.



With the spread of ssnb, as some of us fogeys were dragged kicking and screaming into an unwanted modern age of javascripted ease, one had not a handful of webmasters finding that they could join new rings more quickly than they could microwave popcorn, but hordes of them, hundreds of thousands of them, the population of a major city. Now, getting back to the subject of statistical experiences life has left us to identify with - if you find that your wallet has fallen out on the street, as you head back hoping against hope to retrieve it, which thought makes you more nervous - that it fell out when you were walking down that shady, quiet residential street that you saw practically nobody on, or that maybe it's sitting somewhere on the North Michigan Boulevard bridge, just down the street from the Tribune Tower? Why?

Because your money is a lot less likely to avoid theft on that shady side street, that's why. Is that because the people on shady side streets are individually more honest? Obviously not. We're looking at the same segment of the population going down each type of street - everybody, more or less. You've got a fighting chance of getting your money back if your wallet fell out on that side street, because most people are reasonably honest and will leave what is not theirs alone, but if you have something sitting out where the tourist trolleys are running past, that's a lot of people to trust. Just like the would-be Russian Roulette champ, you only need to lose once, right? As the number of ssnb users climbed, as the system had more and more users set up to have the opportunity to add rings almost as quickly as they could blink, a similar principal began to take hold.

In the beginning, there was an informal understanding that one would exercise some sort of restraint as one joined rings, this understanding being a matter of aesthetics, not morality. To have a single page on a hundred rings might not have been seen as being wrong, but it was seen as being ugly, and very few people wanted their work to be ugly. But there were always a few, and as the barriers to mass webring subscription weren't just dropped but were battered down, ugliness became easy enough that some people started to notice how profitable it could be, in a variety of ways. Megausers such as the "Rainbow of Spirituality" gentleman I allude to elsewhere in my pages could be seen signing individual pages up for hundreds of rings at a time, on many of which their sites would be grossly off-topic. Most users, in the beginning, didn't approve of this, but there was a corrosive side-effect of the practice. Because individual ring memberships had just become noticably more numerous, without the total number of visitors to Webring going up proportionately, the memberships one had were, in some sense, watered down.

With more users out looking to join more rings, now that the practical restraints against doing so were removed, on creating new rings, ringmasters found that they were likelier to get new members. New members who were likelier to have belonged to an absurdly high number of rings, but because SSNB is usually set to display just one navbar from one's stack, the ringmaster wouldn't know that unless he took the time to open the stack. Picture yourself opening your e-mail in the morning, before heading out to work, and seeing the notice that somebody is in queue. You can take the time to look over that site in detail, maybe you should do that - but many won't and many didn't. The number of rings soared, with the result that a large amount of time and work went into designing rings that in the end, were hardly visited by anybody at all, (0.5 hits / day) per site on the ring being the rate at which most of the more successful rings were being visited. At that point, having one membership on one ring translated into nothing.

Quite understandably, users weren't happy to settle for having nothing, so more and more users started applying to more and more rings, and the problem snowballed, producing perverse incentives. As the number of ring memberships skyrocketted, the dilution of the value of each individual membership became more exteme, creating a strong incentive for the individual member of the Webring system to increase the number of ring memberships for his sites, as he found himself running at high speed just to stay where he was, scrambling to find new rings to join just so that his traffic wouldn't diminish more than it already had. This labor saving innovation, in the long run, hadn't proved to be labor saving at all.



What had been the abberation was now becoming the norm. More and more, the members of Webring were tending to join large numbers of rings on that system, even as they found themselves saying "people really shouldn't join as many rings as they're joining. But if you were the odd man out who refused, or, like me, the straggler who started doing as others did with reluctance, you'd suddenly find that your traffic from your rings had dropped through the floor. What Webring.com had created was something known as "a prisoner's dilemma"; a rational response to one's best interests did not produce good results in general, and given the large number of players in the game, the suboptimality of the results became a statistical inevitability. One might be able to reasonably count on five or six or even twenty people to hold off for the greater good, but statistically speaking, the likelihood of people doing so by the thousands becomes remote. Where one once would typically find a single ring membership, maybe two or three in the case of a very large site, one now found a dozen or two memberships on a page, sought out by members who found that in a manner that some might find ironic, the labor saving nature of the code ended up saving almost none of the membership any labor in the short run, and in the long run?

One day, Webring management decided that its system needed cleaning. There were too many sites on too many rings, they decided. Perhaps so, one might concede, but who is responsible if upper level management sets perverse incentives in place, churns ring ownerships so often that the incentives can't be ignored by site owners, rendering the incentives even more aggressively perverse, and then keeps on doing so for a few years with the result that a site owner can not, as a practical matter, simply wait for better days to come? Webring's position was that the blame for the logical consequences of its own freely made choices lay with the evil ringmasters and ring members. Webring's management announced the creation of a new membership policy that was causing more than a little concern on the part of the membership. As explained on the benefits and transition plan pages, there would be three tiers of membership (Free, Premium Level 1 Membership and Premium Level 2 Membership), each level of membership allowing one to manage a certain number of rings and hold a certain number of ring memberships:

Free. Hold up to five ring memberships, manage up to two rings.

Premium Level 1 Membership. $12/year. Hold up to fifteen memberships, manage up to ten rings.

Premium Level 2 Membership. $36/year. Hold up to fifty memberships, manage up to thirty rings.

At this point, Webring decided to get a little devious. While the new membership plan would offically begin on October 1, 2006 according to the letter we were sent, one would have until January 15, 2007 to choose a membership plan. That sounds nice and forgiving, doesn't it? Not really. Think about it. What was going to happen on that deadline? According to Webring, as common sense would suggest, one's memberships would be trimmed down to the limit set by one's membership plan; if one still had a free membership in Webring, all but five of one's ring memberships would be deleted, reportedly the top five appearing on one's management page, and all but two of the rings one managed would be put up for adoption. A considerable theft of intellectual property from those who refused to pay for the privilege of doing volunteer work for the for-profit corporation that Webring now was, one could well argue and many did, but there's a bigger question that was left largely unaddressed.

When would one start running up a tab? To read this plan literally, what Webring was doing was the equivalent of having a cabbie walk out of his cab with a passenger sitting in the back seat, while he left the meter running. If the membership plan began at the start of October, this change being announced in a letter dated Tue, 26 Sep 2006 (giving not much notice), then the implication would be that when January first came around, one would be billed for three months of service at a much higher than free membership level; in fact, one could be billed for multiple memberships, because even at level two one wouldn't be alloted as many memberships as a great many people already had, in response to the prisoner's dilemma situation detailed above. How many level one or two memberships would one need to get up to a few hundred memberships, and how much would all of those memberships cost over a three month period? While we're at it, what would happen when the three months were up? Would the members recieve bills for those memberships, finding even that which was allowed them as holders of free memberships suspended because their memberships had been locked for nonpayment? Would they find themselves in small claims court if the bills were large enough? After all, a noncontested and properly delivered bill does become automatically collectable even if otherwise without merit, after about sixty days in many locations.

Between the thought of being billed, and of the sudden invisibility of one's site should all of one's ring memberships be suspended, an understandable panic took hold, especially after Webring responded to an expression of these concerns with dead silence. Many, finding the task of deleting that many memberships to be an oppressively time consuming one, just cut code on a great many of their pages, hoping that the automatic suspension and deletion process would downsize their ring memberships to an allowable level before the bills ran up. As these suspensions and deletions were carried out automatically by Webring's software, this would seem to be a reasonable solution to the problem, but it was one that the Webring staff was unwilling to be tolerant of, as the system's response would show, delivering a message the staff would soon reinforce.

For each membership suspended, one would get an automated message sent to remind one that the code for that specific membership was missing from one's site, and that said membership was in danger of deletion. Fair enough, perhaps, though maybe a little bit on the overkill side when the same message was sent for each ring that a given page was dropping out of - if you knew that you had cut the ssnb for one ring on that page, how could one not know that one had cut it for all others, if one was using the same code fragment for all of those memberships? But it went further than that. Let's say that you've removed the code for a particular ring membership, and that after the first reminder, you get another, and another, and another, until a few dozen reminders had arrived for that membership, alone, at times several in a day. Now let's say that this response was multiplied by the number of rings the page you had removed the code from had been on - would you consider that to be a friendly reminder, or would you consider that to be harassment? Countless man-hours were wasted as individual users would literally clear thousands of the nagging reminder e-mails that the system would not stop sending, until in sheer exasperation some started filtering all e-mail from Webring. Some, deciding that they just wanted to be done with the entire system, decided to delete their Webring accounts altogether, and discovered that the system had been configured in such a way as to make that impossible. There is no delete button for a Webring member to hit. This fact was defended by one of Webring's many anonymous support staff members with a defiant show of attitude in the Webring members forum

For far too long we have been catering to the mega users. That has stopped. If you create 500 memberships and decide one day that you can't be bothered...we don't care. Honestly, we don't. You created the mess, you clean it up. We're focusing our energy on the other 99%, those that have one or two memberships.

Rings are COMMUNITIES folks, not pages of links like ringsmurf. Have the courtesy of leaving a community gracefully if you decide it's time to leave. COURTESY. We're demanding it of the managers and we're expecting it from our members as well. We're busting our humps to build a great system (currently traffic delivered TO sites outpaces traffic FROM sites 2:1) and we're simply asking everyone to act like they are in a community, not a me me me free for all.

brushing off reasonable complaints with an arrogance which the Webring administration has become known for, blaming its hapless outgoing former users for the predictable results of its own bad choices, along the way excusing itself for working a needless hardship on a number of the very users its bad choices had driven off. One can pardon them for their naivite about the matters which we have discussed. Stupidity is forgivable. Stupidity acting in tandem with callousness is not. Perhaps the best moment came after the webring system ended up being cluttered with abandoned and deleted rings back from the dead, courtesy of intervention by the Webring staff, and the policy was altered in such a way as to make the deletions unnecessary before the January 1, 2007 deadline - but not until they had been made by the countless thousands. Somebody wrote in to ask whether some of the ringmasters who had given up their rings in response to the old new policy would be allowed to adopt them back, now that the policy had been changed, and Webring's response was to say that they saw no need to protect those from the natural consequences of their poorly thought out actions - actually daring to imply that compliance with its own stated policies was an irresponsible practice calling out for a punitive response, or indifference to the user's concerns at the very least.

"How dare you obey us. Now you must accept that you will pay the price for your actions". That took nerve; even the word "chutzpah" doesn't come close to being adequate to describe such a response. If compliance with policy is itself declared to be an actionable offense, how does one avoiding offending anybody who takes such a stance? Left with reasonable grounds for questioning the sanity of the support staff, I found myself with no choice but to relocate this ring to another server, which is why you now see it on the World of Webring system. Getting it to appear in the search engine results has been difficult, but the peace of mind that comes from working with a more stable group of people more than justifies the trouble. My only regret is that I didn't move this ring sooner.



[ continue ]