X-Message-Number: 27165 Date: Sun, 02 Oct 2005 04:12:53 -0400 From: "Kevin Q. Brown" <> Subject: ratings, risks, and redemption In message # 27152 Kennita Watson expressed concern that CryoNet messages accessible via the web, but filtered from the digests because of low ratings, would not be seen and rebutted by CryoNet readers. The archives thus would not provide sufficient and obvious clues to a casual web surfer that such messages should be regarded with suspicion. Furthermore, if an outrageously bad message slipped through like this, it could cause Trouble. That's an interesting point, Kennita, and, now that you've pointed it out, I cannot claim to be ignorant of the possibility. Darn. Now I'll have to do something about it. First, for messages that CryoNet readers _have_ read and rated, I just made it easier to see the ratings. For example, the bottom of the page for viewing your message # 27152: http://www.cryonet.org/cgi-bin/dsp.cgi?msg=27152 now includes the link: Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=27152 Clicking on that link will show the current rating for that message as well as provide an opportunity to add a new rating. Although the rating system makes it easy to view the ratings for your message, it does _not_ make it easy for anyone but you to find the overrall rating for your email address (""). That is by design. (See message # 24033.) For messages that are filtered from digests, due to low reputation, that may be outrageously bad, yet also unread and thus unrated, I have not yet decided upon the best action to take. So far, this has been a low-probability event, and remaining oblivious to the hypothetical risk has worked quite well. But now that I've been warned, it would be negligent of me not to _do something_ about it. :-( Charles Platt asked in message # 27153: "I wonder if the Cryonet system for barring people whose reputation is too low includes any provision for gradual reinstatement." First, I must point out an incorrect assumption in your question and then I will try to answer the corrected question. The rating system does not really bar people from CryoNet; it just filters messages from the daily digests that are from email addresses with low reputations. That is an important distinction, not just a semantic quibble, since a person is not an email address. In fact, a person can have _many_ email addresses. Furthermore, as I have pointed out before, only the daily digests are filtered, not the various other means for accessing CryoNet, and even the digests tell explicitly how to retrieve the contents of the filtered messages. Now, to answer your question, my original plan for the rating system included a time-based discounting scheme that would weight the ratings of old messages less than the ratings of recent messages. That would convert a low rating into a temporary "penalty box" for bad behavior, accomplishing essentially what you are suggesting. That aspect of the rating system never got developed, though, and now I doubt that it would prove useful. While a low rating can be improved by posting high-quality messages, that is much more difficult to do (especially when most people never read those messages) than to get a new email address and immediately start over. We saw that happen recently with Despres. No waiting for eventual redemption or hard work writing well-considered messages was required before he could resume posting unfiltered messages to CryoNet digests. Fortunately, enough CryoNet readers responded unfavorably to those postings that the reputation of his second email address dropped quickly. Yes, there are ways to hack the rating system. Kennita Watson has pointed out some weaknesses, too. But it was not designed to foil any formidable adversary; it was designed to discourage disruptors. Getting a "reputation too low" is like a public spanking. It shows that the CryoNet readers place little value in what one has to say. Filtering the postings from the digest is kind of like shunning. Usually that is sufficient. If and when stronger remedies are needed, then I'll escalate as necessary. Kevin Q. Brown (include "cryonics" or "CryoNet" in the subject line) Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=27165