[CIMC-work] Re: Automate Consensus Process
ChrisGeovanis at aol.com
ChrisGeovanis at aol.com
Sat Mar 13 16:46:21 PST 2004
whoa, just when more and more of us were getting used to the awesome and
potentially dictatorial power that the new code provides for 'hide' -- a pretty
damned excellent idea grounded in direct democracy. couple things. first, i
think our code right now is pretty much set up to let us use a model very similar
to the Westgard model, at least if i'm reading admin functionality properly.
problem with same is twofold: 1) most of us who have admin capabilities don't
get how to use this functionality; 2) it takes time time time, and that is what
we are all short on.
what the current code does not do -- and i'm not sure we'd want to do -- is
hold stories in a non-visible que for a certain period of time for editorial
review. the current no-que situation is bad because crap goes up and has to come
down, but good because stuff which is not crap (and most isn't per se) goes
up right away. what tom's proposal suggests to me is that we should think about
convening a stand-alone meeting for all who currently have access to admin
privilages to really parse out the functionality of the new code, particularly
those features (which we don't currently use) that mirror the functionality Tom
describes...and start using it.
there is a huge problem -- the volunteer labor pool. right now everybody
w/admin capacity jumps on the wire or in admin and admins appropriately...and as
their volunteer availability dictates. we have an enormous problem with getting
our friends out there to even get their content on our wire, and i'm not sure
that asking them to assume the additional responsiblity of managing same
would work. seriously, alot of the stuff i post to the wire gets there via some
query from some activist who says "can you please post this because i
can't/won't/don't have full access to the internet."
however, i can personally think of at least half-a-dozen vigilant viewers of
our wire and ardent admirers of our project out there that will never, ever
come to meetings or seek to join the collective but who WOULD (and effectively
do) help monitor the wire. Bob Schwartz of the Chicago Anti-Bashing Network
leaps to mind, for example. right now he calls me at work to huff when something
objectionable comes up. is a possible solution, then, to empower a group of
'auxiliary' collective supporters who would be entrusted with helping fulfill
admin functions? we could build in both processes and code options for
accountability and transparency.
christine
In a message dated 03/13/2004 6:06:16 PM Central Standard Time,
tom at ilmechliens.com writes:
> Subj:[CIMC-work] Re: Automate Consensus Process
> Date:03/13/2004 6:06:16 PM Central Standard Time
> From:tom at ilmechliens.com
> To:imc-chicago-working at lists.indymedia.org
> Sent from the Internet
>
>
>
> I have a couple of suggestions on how to make the consensus / hide / delete
> process a little smoother and faster. I don't know if anyone else is
> familiar with the two sources of my ideas: hotornot.com and Westgard QC
> (who shares my uncommon last name but has no known relation). The benefit
> in combining these is that it would entirely automate the selection /
> prioritization process, and take a lot of the personality out of it, which I
> see as a benefit. The less this is about who likes whom, the better.
>
> First, view the process as a categorization. Each story is reviewed and
> needs to be placed in a category: Great stuff goes to the center column.
> The vast bulk goes in the vast bulk pile. Some is dirty trash and gets
> hidden. Some is so bad it deserves deletion. That's pretty discrete input
> that could be done through a computerized process pretty easily.
>
> hotornot.com is a little hard to describe. Part of it is a matching system
> for couples; the other part is a rating system in which you rate a person's
> photo from 1 (not) to 10 (hot). The important thing for this list is that
> it has "moderators" who are given a relatively complex list of criteria on
> which to judge photographs and written intros. They do a pretty good job of
> keeping out pornography and hookers. They have basically the same
> discussion on hotornot as has taken place on Indymedia: There's a place for
> prostitutes (and Klansmen) on the Web, just not on our hotornot (Indymedia).
> So they have the moderators as a filtering system, but it's simple - they
> either say that it passes, or that it doesn't, without explanation.
>
> Westgard QC permits the use of multiple-track elimination and/or
> categorization, whereas hotornot is either all-clear or dead. The Westgard
> QC multi-track model works better for Indymedia, since there are multiple
> categories to put things in: center panel, general pool, hide, delete.
>
> CIMC would have moderators who would review each story and do their
> inputting. For a certain time after filing, a story will be in the review
> pool. Indymedia moderators log on and review each submission on a webpage
> that ends with radio buttons for various categories of good or bad "votes."
> Click here if it's racially offensive. Click here if it's concise. Click
> here if it's worker-friendly. The computer would then combine the data to
> place the story, using some sort of algorithm. Drafting the rules would
> take some time, and I am not the one to draft the algorithm for CIMC, but I
> think here might be some rules, by way of example (I'm also willing to help
> with actual drafting):
>
> If the reviewer is offended by the use of a "strongly disapproved word,"
> then the reviewer clicks a radio button that reflects the level of offense,
> anywhere from "no big deal" to "godawful." "Strongly disapproved words"
> would be a defined list, including the usual ethnic, religious, and gender
> words. You might have situations like a hip-hop reporter who refers to
> Jesse Jackson as "my nigga." The sex columnist who writes about "fags." Or
> a story on guns says something about "that bitch Janet Reno." The presence
> and degree of offensiveness is contextual, and whether the particular
> reviewer takes offense would be the measurement.
>
> Center panel material is still by consensus, but there's no longer any
> emailing around to get it. There's a time limit for input, and anyone who
> gets their data in within the time limit determines what goes on the center
> panel. There's a radio button for "center panel nomination," and if any
> Indymedia reviewer doesn't click that button, then the story doesn't go
> there. You could also have a default option, in which, if no new story gets
> placed there for a certain time, then the highest-rated story gets placed
> there instead of a consensus story. Or all stories getting a rating over a
> certain number or ratio.
>
> In order to make sure the system is working according to Indymedia
> standards, you collect the data on which moderator rates what on which
> stories. Moderators agree not to be anonymous, at least to the extent of
> having their judgments recorded. Then you monitor the moderators /
> reviewers in person. This is where the email and face-to-face comes in.
> "Bob, no one else found a racially-charged word in this article. Why did
> you give it a bad vote for racial offensiveness?" It doesn't mean Bob's in
> trouble, since he may just have a bigger or more recently updated vocabulary
> than the rest. It's a chance to review the rules to see if they create the
> right result. If everyone is honestly following the same rules, there will
> largely be agreement on the ratings. It just allows you to track down
> stealthy Nazis who get in and give a bad rating to everything except Nazi
> stories.
>
> The end result would be that this selection and categorization process would
> become largely automated, leaving the current workers to step up a level -
> managing reviewer / moderators, instead of doing all that editing work
> themselves. Then there could be an effort to make sure that the moderator
> pool matches Indymedia's goal clientele. We could invite desireable people
> to become moderators, say, board members of organizations we like, who would
> commit to logging on and reviewing ten stories each day (or ten a week -
> whatever). This allows for a community outreach and publicity program for
> Indymedia.
>
> Thoughts, comments, praise, rotten vegetables?
>
> Yours,
> Tom
>
> _______________________________________________
> Imc-chicago-working mailing list
> Imc-chicago-working at lists.indymedia.org
> http://lists.indymedia.org/mailman/listinfo/imc-chicago-working
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.indymedia.org/pipermail/imc-chicago-working/attachments/20040313/7ca4d3dd/attachment.htm
More information about the Imc-chicago-working
mailing list