Jump to content

Post here for a +1


Kelandon

Recommended Posts

Interesting idea. A few kinks to work out:

- Won't work for anyone who's changed their PDN much, as you point out. That actually afflicts a number of top posters, though the impact is most obvious with Alorael, TM, and myself.

- A lot of posts no longer exist anywhere on the internet. There were actually four forum incarnations, but not the ones listed above: Ikonboard, UBB, UBB.threads, and IPB. Ikonboard posts are almost entirely gone except for some difficult-to-crawl ones on the Wayback Machine. The Pied Piper you mentioned has archived much but not all of the UBB posts; in particular, the first several years of UBB are missing, and I believe it has some other smaller gaps as well. Anything that was on UBB.threads between its inception and the last purge, no longer exists at all -- probably most relevant for currently active members. Additionally, there have been a few isolated cases of very huge threads that had to be destroyed to keep the board database from fracturing; this applied to a few large threads in UBB days, and a couple of REALLY large ones (up to 1000 posts) from UBB.threads.

- Due to inconsistencies in which forums were purged when, there are actually a fraction of posts that were archived on Pied Piper, but also survive on the current forum, largely in the game forums. Those posts will be counted twice.

- Sometimes people are quoted without their name being used, or with a manually edited variant used. In particular this happens if someone is quoting multiple people in a single post, as they are wont to manually type out the quote tags. (Before IPB there was no MultiQuote.)

 

I note that for the users you tried out, who are not significantly compromised by the PDN, the results correlate pretty closely with total postcount. Which I guess shouldn't be a surprise.

Link to comment
Share on other sites

Good idea! Unfortunately that number would be even more vulnerable to the lost-post problem. Somebody like Lilith, who posted often in both the early UBB and early UBB.threads periods, would have her denominator increased for those periods, even while her numerator stayed the same. Someone who arrived around 2008 could be even more harshly impacted.

 

OTOH, it might be possible to arrive at an accurate sample for a particular user and a particular time period. For example, someone who joined in the mid-to-late UBB period, using Pied Piper could get you a fairly accurate comparison for that era. (I think -- that's assuming PPP last stored postcounts at the end of UBB.) It really depends on the availability of outdated postcount information.

Link to comment
Share on other sites

there are actually a fraction of posts that were archived on Pied Piper, but also survive on the current forum, largely in the game forums.

That should be ok I think, quotes in the Pied Piper era went by "written by: User" which are counted only in pied-piper.ermarian, and IPB's quotes are "User, on [date]" and are counted only in spiderwebforums.ipbhost to avoid conflict.

 

Sometimes people are quoted without their name being used, or with a manually edited variant used...if someone is quoting multiple people in a single post, as they are wont to manually type out the quote tags. (Before IPB there was no MultiQuote.)

That's a problem (this post for example), but if it's too small a percentage it can be neglected. The quote-score is only a number which is indicative of your total number of posts; however, its possible that some more digging could reveal more complex crawler filters with a better precision (nearer to actual quote-count). The variation of quote-score per sub-forum including General and each of the gaming forums should be interesting.

Link to comment
Share on other sites

This is actually interesting, but on the other hand: do we really need another way to rank members? Especially, a way that weights their distant pasts as much as their presents?

 

The original idea was to rate threads, and even that was a concession to the reality of sample sizes, from the ideal goal of rating individual posts, for spamminess.

 

Edgwyn suggested some sort of rolling average of keyword overlap, as a way of measuring speed of topic drift. If your post doesn't include any of the frequent words in the preceding page of posts, that's a pretty fast drift. For interest, we could just let each post have interest 1 by default, and add to that the number of Likes the post receives. That's pretty crude, but it might not be too bad in practice, unless we were mainly concerned with perniciously pointless posts that really are not interesting at all.

 

In my fantasy world, the forum software would compute the Spam number automatically for each post, and update it as Likes come in; and show the poster (but no-one else) the score their post had reached. I imagine this serving the purpose of those little speed radars that simply show you how fast you're going, so you'll go Yikes and slow down if you're well above the limit.

 

The main problem I see is that deliberate trolls would make a point of trying to rack up high Spam numbers. I have a feeling this might do more harm than we received in the way of the benefit that unconscious spammers woke up to their spamminess and repented.

Link to comment
Share on other sites

I think the algorithms y'all are making are far too complicated. Try going back old school to doing things by hand! For instance

 

... but who wants to rate every post?

Answer: probably a disturbingly large number of the people who post here. ;)

 

4/5. Witty and slightly self-deprecating, in a humorous way. Possibly unnecessary use of smiley. No typos.

Link to comment
Share on other sites

In my fantasy world, the forum software would compute the Spam number automatically for each post, and update it as Likes come in; and show the poster (but no-one else) the score their post had reached. I imagine this serving the purpose of those little speed radars that simply show you how fast you're going, so you'll go Yikes and slow down if you're well above the limit.

 

I doubt a "spam number" would make anyone change their ways. We tell people when they are excessively spammy, first through hints and then directly if the behavior continues. If being told you are spamming doesn't convince someone to change their posting behavior, a number isn't going to do the trick either. Plus, for several years on the UBB board, user ratings were enabled, which worked a lot like this hypothetical spam number would. No one ever stopped spamming because their user rating was low.

 

Dikiyoba only slows down (if necessary) at roadside radars because they're commonly posted in school or construction zones, and Dikiyoba really doesn't want to get a ticket in those areas. Dikiyoba would also slow down without the radars, though, so really the radars aren't the relevant factor there.

Link to comment
Share on other sites

Yeah. I guess I'm just postulating a major new use for large scale AI: giving everyone an appreciative audience. In the future, everyone will be famous forever, among the bots that have been designed specifically to obsess over their every word.

 

For now, maybe we could just install a laugh track.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

×
×
  • Create New...