Slashdot Mirror
Google Throws /. Under Bus To Snag Patent
theodp writes "Before Danny Hillis and Bran Ferren invented Google's newly-patented system for 'Delegating Authority to Evaluate Content', Google says users looking for content evaluation websites were condemned to the likes of Amazon.com and Slashdot. From the patent: 'Many sites found on the World Wide Web allow users to evaluate content found within the site. The Slashdot Web site (www.slashdot.org) allows users to "mod" comments recently posted by other users. Based on this information obtained from the users, the system determines a numerical score for each comment ranging from 1 to 5.' The problem with sites like Slashdot, Google told the USPTO, is that 'because there is no restriction on the users that may participate, the reliability of the ratings is correspondingly diminished.' Commissioning a small number of trusted evaluators or editors would increase the reliability of the evaluations, Google notes, but wouldn't allow nearly as much content to be evaluated. Google's solution? Allow trusted evaluators to transfer a 'quantity of authority' to like-minded 'contributing authorities', who in turn designate and delegate authority to additional like-minded contributing authorities. Think Microsoft Outlook 97 Delegate Access meets Slashdot Karma Points, and you've got the general idea!"
3... 2... 1... Go!
Meh. This is more like "We think we can improve on the best thing." I believe we actually had a thread about this here recently.
Help stamp out iliturcy.
Let's face it, the slashdot moderation system has been broken for a long time. That's where the term slashthink/slashdot group think comes from. If you post a comment that general user base of slashdot likes, it will be modded up. If you post a comment, even a really insightful and interesting one that the general user base doesn't like, it will be modded down. Comments that rank up? Promote free speech, removing copyrights, getting rids of patents, point out how "suits" just don't get us geeks and so on. Comments that go immediately down? Tell informative, but bad points about the current state of Linux, dislike Google, try to be reasonable about copyrights and DRM or say that Microsoft's Visual Studio still kicks ass any other IDE out there.
I can't find the old post now because it was long time ago, but it went something like this. Every user are given some amount of moderation points, that affect the moderation as a whole. In addition to that, it affects the moderation you see favorable to the likes of you. If they are on your friend lists, their moderation carries more value. If they have moderated similarly to you, their moderation weights more to you. Of course, this should be balanced so that you don't get fully one viewed comments - if some comment is generally modded very high (and forget the -1-5 scale now), it would be displayed to you anyway. If you add to that that comments where you, or similar persons to you have commented, will be fully displayed regardless of their moderation (or some adjustation of that), it would work out really well. Of course, it needs a lot more computation power on the server side.
For me, personally? I like Reddit's comment system. It has it's faults, but it's better than Slashdot. Interesting posts are on top, and you can just scroll down for more.
Still, I browse Slashdot at -1 and read what interests me. I come here for the comments, jokes and all that. I like to see it all when the subject is interesting. No moderating system can ever beat your own judgement (even if it's wrong one).
Allow trusted evaluators to transfer a 'quantity of authority' to like-minded 'contributing authorities', who in turn designate and delegate authority to additional like-minded contributing authorities.
Um, isn't this exactly what would promote the problem of politically active users donating time to keep adverse stories repressed?
...
Quality can be controlled to some extent but biases are much harder to determine
My work here is dung.
doesn't the meta-moderation system essentially do what Google is talking about - I always assumed that if your mods got marked as appropriate in metamod, your chances of modding again improved, and vice-versa.
"National Security is the chief cause of national insecurity." - Celine's First Law
Oh so long ago when it sold itself for $$$ to whoever that company was. The bus has been running it over ever since.
So who trusts the "trusted evaluators" in the first place? This could easily be abused into more of a group-think than slashdot, if indeed /. is guilty of group-think. I'm thinking more like a personal blog, where moderators must approve all comments. If the mod doesn't like it, the comment doesn't exist for the general public. Do we trust google to moderate our content for us?
I mean, I guess we (as a corporate whole) already do, based on their share of the search market, but seriously. How far should we let this go?
Let's face it, the slashdot moderation system has been broken for a long time. That's where the term slashthink/slashdot group think comes from. If you post a comment that general user base of slashdot likes, it will be modded up. If you post a comment, even a really insightful and interesting one that the general user base doesn't like, it will be modded down. Comments that rank up? Promote free speech, removing copyrights, getting rids of patents, point out how "suits" just don't get us geeks and so on. Comments that go immediately down? Tell informative, but bad points about the current state of Linux, dislike Google, try to be reasonable about copyrights and DRM or say that Microsoft's Visual Studio still kicks ass any other IDE out there.
There's a difference between being "unpopular" and "wrong." I disagree with you and find that well written -- though unpopular -- posts will be moderated highly. I, myself, have participated in receiving such moderation. You can make valid points about the current state of Linux (without having to be apologetic) as long as you know what you are talking about. Here's one of my own posts where I rip on Google's tax evasion and it's moderated +4. That's just a quick one, if you need more, I'd be happy to spend some time to provide you counter examples do your claims. As a developer, however, I must say that your Visual Studio statement is completely without merit and will always be modded down. I come to Slashdot not because I'm afraid of debate but because I thirst for it. The most valuable comments are those that put me in my place.
I can't find the old post now because it was long time ago, but it went something like this. Every user are given some amount of moderation points, that affect the moderation as a whole. In addition to that, it affects the moderation you see favorable to the likes of you. If they are on your friend lists, their moderation carries more value. If they have moderated similarly to you, their moderation weights more to you. Of course, this should be balanced so that you don't get fully one viewed comments - if some comment is generally modded very high (and forget the -1-5 scale now), it would be displayed to you anyway. If you add to that that comments where you, or similar persons to you have commented, will be fully displayed regardless of their moderation (or some adjustation of that), it would work out really well. Of course, it needs a lot more computation power on the server side.
That sounds like a really sheltered solution. All I can think about as a comparison is people who live in -- and I'm not picking on them specifically -- a Mormon community only holding their immediate relatives as valid sources of comments. This can be said for any number of things, however, but this proposed "lensing" of Slashdot would just allow people to turtle into their sheltered bubbles. Eventually any contradictory points that I might have been exposed to are safely locked away and I am never challenged. What a horrible, repressed, unenlightened, biased, polarized existence! The website will be a therapist -- telling you only what you want to hear. Disagree with something? Delete the offending friend.
For me, personally? I like Reddit's comment system. It has it's faults, but it's better than Slashdot. Interesting posts are on top, and you can just scroll down for more.
Then go back to Reddit. Why are you here? Go back there where you can delete or modify what you just said when someone wants to engage in a debate with you! Never have I been so exasperated as with my brief foray on Reddit. Valid counterpoint? Deletes his post. Now what?
No moderating system can ever beat your own judgement (even if it's wrong one).
I think you're hung up on wrong/right versus unpopular/popular opinion. It's not so black and white and there is a blur there but I feel that Slashdot 1) presents a decent mix of stories and 2) the subsequent moderation gives you a good idea of what is popular and generally correct/informed.
My work here is dung.
Someone with brains AND A VOICE finally speaks up against Slashdot's miserable organization, wait.....what.....hold on a sec:
Okay, so what am I missing here? Where's the article? I see a link to a patent, a link to a pointless JPEG, and some kid's anecdotal evidence (if even that) that Google hates Slashdot. C'mon, theodp. This is the Internets. If you're going to make some absurd comment, at least have the wherewithal to link to someone else's page where someone else actually came up with or cited the idea. Even if it is completely bogus. It looks to me as though you waved your hands, threw some pixie dust, and declared that Google just insulted Slashdot. Where's the beef, sir?
--TSP
captcha: smoked
I think the /. community mods accurately, the good out weighs the bad, and I have had more than one of my comments modded out of existence, and frankly some of my comments deserved to be (we all have bad days) but the thing to keep in mind here is we, we being the members of /., aren't modding for the outside World we mod for the community here on /. so it works well even with the trolls and hopeless pontificates.
No changes needed in my view.
"If any question why we died, Tell them because our fathers lied."
In case you hadn't noticed, this place has been invaded by a succession of increasingly dumber editors, which are probably rejects from Boingboing.
Google didn't throw anything under the bus, they just pointed out what we /.ers have known for 15 years. They're not patenting Slashcode, they're patenting "weighted moderation" or something along those lines, where each user has a certain numeric authority assigned to them, which affects how strongly their opinion is weighted in the scoring process. Still, boo urns on Google for patenting such a trivial algorithm, but I'm pretty sure they repealed "Do no evil" a long-ass motherfucking time ago.
-Billco, Fnarg.com
I think Slashdot eds are being a little too sensitive. They didn't sue Slashdot or harm it, they simply claimed in a patent that they devised a better system. While I think software patents are dumb, I don't think creating a different system and saying why you think it's better is much of a problem.
Democracy Now! - your daily, uncensored, corporate-free
The minute /. starts to "Allow trusted evaluators to transfer a 'quantity of authority' to like-minded 'contributing authorities', who in turn designate and delegate authority to additional like-minded contributing authorities." because that is too much like the current system of media control and politics, or in other words go with the flow or fuck off.
Again, leave it alone it has worked just dandy all these years.
"If any question why we died, Tell them because our fathers lied."
I think I may want to contest this patent.
The patent cites Slashdot comment moderation as an example of how not to assign importance to user actions. Its authors were apparently unaware that the algorithm they described in November 2010 is virtually identical to the way Slashdot has actually assigned importance to user voting on Firehose stories since May 2008 (give or take). I know because I wrote it.
What this patent calls "authority," we call user "clout."
Multiple clouts, actually. Each Slashdot user has a number that describes how valuable the system believes their up/down votes in the firehose are, and it's separate from how valuable their descriptive tags applied to stories are. (Up/down votes are simply tags with special names, making vote-scoring and description-determination very similar under the hood.)
It's been a while since I looked at this code -- I work for sister company ThinkGeek now -- but scanning over our public repository here are some of the interesting parts:
plugs/Tags/tags_updateclouts.pl - the tags_peerclout table is the way that each type of clout is built. It has fixed entries at gen=0, the zeroth generation, which would typically be the Slashdot editors or other users considered reliable and definitive. To build gen=1, the code looks at how many users tagged or voted on the same objects as the gen=0 users did, and assigns the gen=1 users scores based on similarity (or difference). Then from the gen=1 users, gen=2 users are assigned scores similarly, and so on.
The gen=0 entries in that table "designate one or more contributing authorities by delegating to each a specific quantity of authority." I don't think I could describe that better myself.
plugins/Tags/Clout/Vote.pm process_nextgen() - here's where each new generation of user clout is successively determined, for firehose votes in particular. Line 194 invokes the algorithm and line 203 assigns that user their new voting clout. This iterative process is the automated method through which "each contributing authority may in turn designate and delegate authority to one or more additional contributing authorities."
plugins/Tags/Clout/Vote.pm init() - sum_weight_vectors totals the change in clout for each generation, and possible weight decreases exponentially. If you're in gen=1 the maximum weight you can have is only 60% of the maximum from gen=0, etc. The fraction is smaller than 100%, which helps ensure "that the total quantity of authority delegated does not exceed the quantity of authority the contributing authority was itself delegated." When the clouts are used to determine firehose item ratings, "the ratings are combined in a manner that affords a higher priority to the ratings provided by contributing authorities to which a greater quantity of authority was delegated."
All this may have changed since it was written. I don't actually know what's running on Slashdot at this moment. I'm just going by the public repository that I knew was on sf.net, and I don't even know if there's a later version of the code available anywhere.
But I suspect that this system would constitute prior art.
Also, looking over my code from 2008, boy, I really wish I'd put in more comments.
Negativity is always more powerful than positivity on positions. What plus only systems do, such as on Facebook, Twitter and Google+ is rank things purely on their positive ratings, though Twitter's ranking is obscured through the Trending Topics system.
Plus only systems don't let anyone actively destroy content, but simply choose to promote or not. Slashdot could use the same system using uncapped mod points per post, and allowing to see top X posts, instead of setting which score to see. Of course, this doesn't change the fact that only random users can mod.
I8-D