Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
The future of blog: The scaling barrier

By zocky in Meta
Mon Aug 26, 2002 at 12:35:40 PM EST
Tags: Kuro5hin.org (all tags)
Kuro5hin.org

Blogs have a scaling problem. Kinda like clubs. The good crowd moves in and they become these perfect little places for some time. And then too many people start coming in, and the magic disappears. Teenagers take over and after a couple of years place is converted into a bad fast food restaurant.

I propose a few improvements to help k5 avoid that.


Why do we do it?

Why do we spend as much time, thought and energy sitting behind the keyboard, reading long screens of text and then typing in our comments? More text, information, data. Is it just the basic urge for conversation? Is this the quasi-intellectual no-life equivalent of gossip? Or are we doing something more? Well, writing stories could be the quasi-intellectual no-life equivalent of readers' comments in a provincial newspaper.

But no, there must be more. There is some pure delight when somebody writes a really good article and other people contribute some really good comments and everything comes out just perfect. We make it just right. But what is it? What do we make?

What do we do?

Essentially, we're publishing a never-ending global newspaper with an overgrown readers' comment section on steroids running out of control. But not entirely out of control. Rusty has constructed it in a certain way, he has established the rules of the game. And he occasionally tweaks the rules to make the game, err, the newspaper more to his (and readership's) liking. And his and readership's liking is something like, "interesting and well written articles, intelligent discussion, good navigation and hopefully some way to be able to skip non-interesting comments".

Any problems so far?

Well, basically, blogs have the tiny little aforementioned scaling problem. Too many people start coming in and the blog goes bad.

But "too many people" shouldn't make sense in the context of a blog. More people is good because more people write more good text. It's just that finding it in the megabytes gets somewhat tedious. After a time the proportion of interesting people and text becomes too low to bother. Take a look at usenet. Take a look at The Other SiteTM.

Kuro5hin is closer to hitting the scaling barrier by the day. Some symptoms: Many off-topic comments, regular is-trolling-good-or-bad trolls, and above all, confusion. Complete lack of familiarity over time. Every couple of days, when I get onto the site, any story that I previously read or commented on, is long dead. Most stories die in a day or two. Nowadays, people hang out in the editing and the voting queue, and by the time story gets to the front page, it's dead.

Why bother to post an intelligent comment, when you know you won't get any good discussion out of it, because two days is just not enough time? Why, for that matter, bother to research and write a good article, when it'll be dead in 36 hours?

I don't want just another blog. I don't want to read endless stupid tit-for-tat debates on trivial points. I don't want to see multiple penis birds or, god forbid, links to goatse.cx. I want my never-ending global newspaper.

Rant ends here. Below, I make a few suggestions for improvements.

How to make it feel better?

P1, P2 and P3 are proposals for simple tweaks of the rules to help with the quality of articles and filtering of comments.

Proposal P1: Rework the editing

Disable topical comments until story is posted. Make editing period of 24 hours mandatory. Allow the author to add clearly marked text to the end of the story even after it's posted.

Many people who really want to post an editorial comment, do not know that they do, hence many "+1 FP" topical comments. Also, real topical comments on unfinished story add much confusion to discussion, as they generate a lot of "where did he say that?" matter.

Proposal P2: Add more comment types

At least add "off-topic". Make it easy to post different types of comments, for instance, add an "add an off-topic comment" link to the comment footer.

We sometimes just have to catch up with people, or say: man, that movie was soooo good. Or say something about somebody's sig. Sometimes you just don't want to read those. Also, there are many good reasons to not include those in the default random-visitor's view of kuro5hin.

Proposal P3: Refine ratings

Change 0-5 system to multiple choice: "I agree (+1)", "I disagree (+1)", "Badly written (-1)", "Spam (hide)"

Average of 0-5 sounds simple enough, and it is. It's too simple. What happens in reality is that "how much I agree with it" has as much to do with my rating as "how well it's written". We just have to express our dislike for other people's viewpoints. It's only natural. This way we would have simple +1 / -1 votes to use for sorting and filtering and an instant poll for every comment, plus spam hiding. "Off topic" could possibly be added. See above.

How to make it last longer?

P4 and P5 are more complex proposals for keeping stories alive longer.

Proposal P4: Merge hot-list and story ratings

Get rid of a the voting queue. On every page, show hot-listed stories, then a set number of most popular stories, then new stories (say, newer than 36 hours). When user hot-lists a story, increase the story's popularity score. Maybe hot-list a story automatically after the user posts a topical comment on it. When a user cold-lists a story, hide it and decrease its popularity score. Allow the user to choose on which page to put the story and reflect these choices in the selection of popular stories to be displayed.

This would produce a continually updated scoreboard for interesting stories. Users would have incentive to keep the ratings up to date, since the look on their pages would directly depend on their choices. This would keep the good stories on the pages for a longer time and thus produce longer discussion. This helps with familiarity and continuity of the front and section pages, but brings up the problem of hundreds of comments for one story, which makes the discussion impossible to read. Hence...

Proposal P5: Restructure the discussion

Introduce an additional level between the comment and the story, the reply. Replies can only be posted to the story and to other replies. They can be edited and their rating system works similarly to the story ratings described in P4, but possibly includes the show/hide comments option. Comments work the way they work now.

This would distinguish between the two kinds of comments we have: chat - the (1 line comments), which is fine, as long as it's new, and the genuine discussion (the full-screen comments), which is interesting to read even if it's a week old. Imposing a mandatory editing period of, say 12 hours, would ensure that the replies get some review before being included in the discussion. Users can then happily chat about the story or the replies or anything, but you can safely turn off old comments or comments to old replies and not miss out on anything. Maybe we can even get rid of comment ratings all together.

Finishing thought

Too much customization ability undermines P4 and P5.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Poll
The scaling barrier is
o years away 11%
o months away 15%
o now 22%
o it's too late 22%
o scaling as in fish? 28%

Votes: 45
Results | Other Polls

Related Links
o Kuro5hin
o Also by zocky


Display: Sort:
The future of blog: The scaling barrier | 106 comments (100 topical, 6 editorial, 0 hidden)
Another possibility (3.80 / 5) (#5)
by Anonymous 242 on Sun Aug 25, 2002 at 10:51:22 PM EST

Perhaps it is time to break up the monolith that is kuro5hin into sections on their own servers with their own communities: tech.kuro5hin.org, currentevents.kuro5hin.org, hobbies.kuro5hin.org, etc., etc.

Forkability (1.33 / 3) (#8)
by dipierro on Sun Aug 25, 2002 at 10:59:52 PM EST

If the posts and articles were released under the GPL that could be done by forking the site.

[ Parent ]
GPL not necessary (3.00 / 1) (#17)
by Anonymous 242 on Sun Aug 25, 2002 at 11:28:14 PM EST

The GPL or a similar license would be necessary for k5 users to fork the site, but there is no need of such for the site to go through an internal fork.

[ Parent ]
You think rusty's going to do it right? (3.00 / 1) (#19)
by dipierro on Sun Aug 25, 2002 at 11:38:47 PM EST

I'd rather allow anyone to make up their own site.

[ Parent ]
If I didn't, I wouldn't be posting (5.00 / 4) (#22)
by Anonymous 242 on Sun Aug 25, 2002 at 11:53:08 PM EST

Regularly posting stories and comments to k5 contain within the action an implicit approval of the way in which Rusty has run k5 thus far. If one does not approve of the way Rusty runs things, it seems to me to be self-defeating to keep posting.

Aside from that, IMO, Rusty has shown more maturity and competence in running an online community than any other single person in the blogging world. Does this mean he can't make mistakes? Certainly not.

Aside from that, that Rusty has advocated what seems to me to be the logical way of fighting the scaling effect, says to me that he is looking in the right directions.

And if I don't like it, I'll simply go somewhere else. That or try to arrange a hostile takeover of the forthcoming CMF. ;)

[ Parent ]

Lesser of multiple evils (3.00 / 2) (#24)
by dipierro on Mon Aug 26, 2002 at 12:04:40 AM EST

If one does not approve of the way Rusty runs things, it seems to me to be self-defeating to keep posting.

I disagree. As long as the benefits of posting outweigh the detriments of posting, then you should do it.



[ Parent ]
And believe me (3.00 / 1) (#52)
by Anonymous 242 on Mon Aug 26, 2002 at 09:39:05 AM EST

Since the k5 populace seems to be responsible for feeding Rusty's cat's, buying him a sailboat, buying him a Vespa, and buying him boxes full of jewel's and such formal regalia as a monacle and crown, the benefits of leaving k5 if I didn't think Rusty were deserving would far outweigh the benefits of staying.

[ Parent ]
Fair enough (4.00 / 2) (#61)
by dipierro on Mon Aug 26, 2002 at 12:50:50 PM EST

I certainly don't feel VA Software is deserving of my charity, but I still post there. I occassionally even post on Yahoo boards, and I certainly don't think they're deserving of my charity. It's really a simple tradeoff, Rusty (or VA, or Yahoo) gives me bandwidth, storage, and publicity, and in return I give them a nonexclusive license to use my work in their money-making scheme.

[ Parent ]
The difference between k5 and /. is (4.50 / 2) (#68)
by Anonymous 242 on Mon Aug 26, 2002 at 02:07:53 PM EST

By posting on /. you cause VA to lose money. Which sounds like an excellent proposition to me.

Yahoo, I have no opinion on. I ignore their message boards.

[ Parent ]

Gods no! (3.66 / 3) (#30)
by aziegler on Mon Aug 26, 2002 at 12:21:10 AM EST

I like the quirkiness found here on k5; I don't want to have to visit five or fifteen different k5 sites to get what is here on k5.

[ Parent ]
I propose mandatory IQ tests... (2.00 / 1) (#6)
by dipierro on Sun Aug 25, 2002 at 10:55:27 PM EST

Anyone below 120 can't rate posts and starts at 0.

In that case, (none / 0) (#16)
by medham on Sun Aug 25, 2002 at 11:26:39 PM EST

Both qpt (88) and I (105) would be shut out. Would you really want that?

The real 'medham' has userid 6831.
[ Parent ]

Well, I was joking, sort of,... (5.00 / 1) (#18)
by dipierro on Sun Aug 25, 2002 at 11:37:35 PM EST

but it wouldn't be that much of a loss. :)

[ Parent ]
Given your record (none / 0) (#54)
by CodeWright on Mon Aug 26, 2002 at 10:29:22 AM EST

Of moderately erudite prose, you must be utterly innumerate to only score a 105. On the other hand, if history bears any relevance, you might just be suffering from a debilitating inability to properly parse and solve the sort of logic problems typical to IQ tests.

--
"Humanity's combination of reckless stupidity and disrespect for the mistakes of others is, I think, what makes us great." --Parent ]
There's another couple of options (none / 0) (#60)
by medham on Mon Aug 26, 2002 at 12:43:48 PM EST

You haven't considered, master logician.

The real 'medham' has userid 6831.
[ Parent ]

Naw, I considered 'em (5.00 / 1) (#62)
by CodeWright on Mon Aug 26, 2002 at 01:01:45 PM EST

But it was more fun to taunt you. It's obvious what you were doing.

--
"Humanity's combination of reckless stupidity and disrespect for the mistakes of others is, I think, what makes us great." --Parent ]
Yes! (5.00 / 1) (#7)
by spacejack on Sun Aug 25, 2002 at 10:57:00 PM EST

I've always thought that blogs and newsgroups and so on had a lot of parallels to clubs.

And in that vein, it really has little to do with the decor, but rather with the people and the music (i.e., content). Eventually, the people you go there to see and hang out with stop going, or the music gets worse; the scene shifts in some way and you loose interest.

Then 5 years later, you talk to someone you met recently and they start going on about how this club is really great and cool. And you think that place? But it went to shit 5 years ago...

Good suggestions (3.00 / 1) (#10)
by roam on Sun Aug 25, 2002 at 11:04:50 PM EST

When I first started reading I thought this would be just another "k5 is dying" article, but you brought up some interesting points, like this one:

Disable topical comments until story is posted. Make editing period of 24 hours mandatory. Allow the author to add clearly marked text to the end of the story even after it's posted.

I agree.  Topical comments should not be allowed until the story is posted... also, I think editorial comments should only be viewable by the story's author, for all time.  Obviously this would go against the opinion of some that comment posting in the queue is used to determine if the story will generate good discussion, but I think that can be determined from just reading the story... Topical and editorial comments bias the voting, IMO.

I also think that author's names should not be displayed while a story is in the queue... but maybe that's another topic.

___
Are they like hamsters?
Specifically, can I tape up a chinchilla, slather him in axle grease, and shove him up my ass? - Patrick Bateman


Hmmm (3.00 / 3) (#11)
by crowbraid on Sun Aug 25, 2002 at 11:09:38 PM EST

That's a good idea, that only the story's author can view the suggestions. I wonder, however, if they wouldn't get flooded with 100 suggestions to fix the same thing. Dunno.

[ Parent ]
Comment Visibility... (5.00 / 1) (#28)
by aziegler on Mon Aug 26, 2002 at 12:18:46 AM EST

Sorry, but I think that the visibility of comments is a feature, not a bug, both during editing and during voting. If I'm looking at a story that needs certain editing, I need to be able to see if there are others pointing out that particular editorial problem so that I don't overwhelm the author with another Me Too post.

With respect to topical comments during voting, it's often the discussion rather than the story itself that helps me decide whether a story should go FP as opposed to SP, or whether it should get a zero instead of the -1 I want to give it.

I do think that topical comments shouldn't be permitted during the edit period.

-austin

[ Parent ]

Couple of things (none / 0) (#35)
by roam on Mon Aug 26, 2002 at 01:25:09 AM EST

First, in theory, an author who puts a story up for editing should be fixing the errors near the time the messages are posted.. so then other users will not be piling on posts due to errors because those errors are already fixed..

Second, I addressed that.. it seems to me (no evidence to back this up) for every 1 person who votes like you say you do, there are 10 people who scroll down to the comments section and read how they should be voting (all the posts that say -1, etc), then vote either that way or the exact opposite... either way is biasing.

I say let the story stand by itself... I'd rather have good stories that generate less discussion than bad stories that generate lots of discussion (usually those are the flamebait stories anyway).

___
Are they like hamsters?
Specifically, can I tape up a chinchilla, slather him in axle grease, and shove him up my ass? - Patrick Bateman


[ Parent ]
I disagree (5.00 / 3) (#39)
by infraoctarine on Mon Aug 26, 2002 at 04:13:36 AM EST

Topical comments should not be allowed until the story is posted.

The result of this would be that what should have been topical comments would be posted anyway, as editorial comments. People won't wait for the story to leave the editing queue, especially if it's in there for 24 hours. With this system, the interest in the story will have died out before you are even allowed to post!

I think editorial comments should only be viewable by the story's author, for all time

Either this would lead to multiple comments on the same thing, or possibly the other way around - people won't bother posting editorials because "someone else probably suggested it already". In addition, the current system allows anyone to give a second opinion on editorial comments, which I think is a good thing.

Topical and editorial comments bias the voting, IMO

It might, to some extent, but is this a bad thing? If a story is good enough, it will get posted, but there has been instances when an editorial comment pointed out something to me that I didn't know but is very helpful to know before I vote.

As an extreme example, we had some cases of plagiarism that were shot down this way, without the need for intervention from the editors. It could also be that there is some serious flaw in the story I didn't catch but someone else did. I'm not omnipotent, so I sometimes need to depend on others expertise to make an informed decision.

[ Parent ]

This reminds me so much of Northern Michigan (4.00 / 2) (#12)
by pyramid termite on Sun Aug 25, 2002 at 11:12:30 PM EST

Where people move into a nice resort town and a couple of years later are complaining that the zoning should be tightened up to not allow so many people in and preserve the nature of the town ...

Everyone wants to be the last ones to move in ...

"I forget, in a certain way, everything I write, doubtless also, in another way, what I read." - Jacques Derrida
Au contraire (2.00 / 1) (#13)
by zocky on Sun Aug 25, 2002 at 11:14:43 PM EST

This isn't about keeping people out. It's about letting more people in and keeping the site interesting.

---
I mean, if coal can be converted to energy, then couldn't diamonds?
[ Parent ]

You've stated your intentions clearly (5.00 / 1) (#25)
by pyramid termite on Mon Aug 26, 2002 at 12:12:13 AM EST

This isn't about keeping people out. It's about letting more people in and keeping the site interesting.

Lead paragraph - And then too many people start coming in, and the magic disappears. Anyway, it's not a matter of "letting people in" - they sign up and that's that.

The complaining about "newbies", "trolls", "teenagers", the size of the web community, etc. etc. etc. is getting out of hand here. We're supposed to be discussing technology and culture from the trenches here, not entrenching our culture with technology.

"I forget, in a certain way, everything I write, doubtless also, in another way, what I read." - Jacques Derrida
[ Parent ]
Anyway (none / 0) (#29)
by zocky on Mon Aug 26, 2002 at 12:20:32 AM EST

"Letting in" was misfortunate wording, granted.

I wanted to say: It's about getting more people in and keeping the site interesting.

You see, as I said in the article, I think more people is good. More people means more good text, but brings some problems. So far most "communities" (which I think is a pretty stupid name) have self-imploded, because they weren't prepared to face the problems of scaling. The whole point of this article is proposing ways to get around the problems without resorting to wholely self-defeating measures like trying to keep people out.

---
I mean, if coal can be converted to energy, then couldn't diamonds?
[ Parent ]

Don't extend the editing period (4.66 / 3) (#14)
by gbroiles on Sun Aug 25, 2002 at 11:18:00 PM EST

you suggested:

Disable topical comments until story is posted. Make editing period of 24 hours mandatory. Allow the author to add clearly marked text to the end of the story even after it's posted.
which would be OK for non-time-sensitive stories, but not good for current-event focused topics, and also subject to abuse by people who want 24 hours' worth of exposure for their crappy little troll post that currently gets clobbered pretty quickly.

There is a way around this... (4.00 / 2) (#20)
by dram on Sun Aug 25, 2002 at 11:45:11 PM EST

And that is for Rusty to enable the spam button for the edit queue. I think its already coded into scoop so it shouldnt take any work at all, just a little check box needs to be checked somewhere. This wont get rid of the time sensitive problem but it would the trolling and spam problems.

-dram
[grant.henninger.name]

[ Parent ]
Yes... (5.00 / 1) (#84)
by rusty on Mon Aug 26, 2002 at 07:32:46 PM EST

...and I will.

Incidentally, how many people considering this "no commenting in the queue" thing know that we've already tried that and it didn't work?

____
Not the real rusty
[ Parent ]

story idea (5.00 / 1) (#87)
by martingale on Mon Aug 26, 2002 at 09:53:44 PM EST

Stop me if you've had this idea already, but you could write an interesting article entitled "things we tried on k5 that simply don't work". Presumably if k5 is going to be long lived and a showcase for web communities, it would be invaluable to have a summary of wrong directions that were taken in the past. Of course, these things could be searched for individually in the k5 archive, but that's both time consuming and a research problem in its own right. Much better to get a long-time member to reminisce, before they forget it all themselves.

[ Parent ]
All interesting points: Culture vs Code (5.00 / 3) (#15)
by TON on Sun Aug 25, 2002 at 11:26:13 PM EST

There is plenty to think about and discuss here re: K5 and other sites. I think some of these things require "cultural" rather than "code" changes. Let's go P by P.

  • P1 Disable topical comments until story is posted.
    I like the idea behind this, even though I am violating it right now. Will disabling topical comments fix the problem? Or, will people continue to post topical ideas as ed comments? I usually try to hold back from commenting topically when something is in the queue, but as you note, more and more often the discussion has progressed pretty far by the time something posts to a page. If I want to get my two cents worth in, I have to do it early. How can this dynamic be changed? Coding the type of comments allowed will not change the basic urge. What if people started 0 rating topical comments made during the editorial/voting period? Go ahead. Somebody be my guest :) I guess this is a social convention and needs to be supported socially, rather than technically.

  • P2 At least add "off-topic".
    Again, this is a social, rather than a technical problem. I forget who suggested it, but someone posted a comment along these lines: "Everyone create a single diary called something like "Contact Me". Hotlist the diary. People can put any kind of "Oooh! Oooh! I gotta tell ya'" stuff in there. It's using the current system to create very basic messaging.

  • P3 Change 0-5 system to multiple choice: "I agree (+1)", "I disagree (+1)", "Badly written (-1)", "Spam (hide)"
    I don't care for this idea for a few reasons. Mostly, it is too simplistic. I like the 0-5 with reply for nuance. Like/don't like ratings make it even easier to mod-storm, fight, whatever, without adding content to the site.

  • P4 Proposal P4: Merge hot-list and story ratings
    Is it just me, or does this sound overly complicated?

  • P5 Introduce an additional level between the comment and the story, the reply.
    Whew! Adding more mandatory editing periods is going to fragment the discussion isn't it? Adding another level of complexity may kill the nice easy usability of K5 that I enjoy.

All that said, I do hope this makes it out of the queue. The navel-gazing (K5 discussion) should be worthwhile.

Ted
---
"Aye, aye, certainly the Roman Empire would still be standing if only

P1 is a complex topic (5.00 / 1) (#37)
by martingale on Mon Aug 26, 2002 at 02:13:58 AM EST

In principle, I like this idea too, and in practice, I find I have been violating it very often.

For physical reasons (20,000 km), I am unable to consistently view/comment on stories during the main k5 activity phases (ie lunchtime in the US, lunchtime in the EU). Due to the speed of story turnover, this means I can physically either comment on a semidead published story, or I comment very early during the editing phase, which ensures some readership and some possible responses by the next day.

The only way to fix this would be to slow down the lifecycle of a story, to something greater than 24 hours. I don't know how this can be encouraged, except that a symptom of success would be a much slower rate of story creation than the one we see now.

Moreover, if stories had a longer lifecycle, we would also see discussions over longer timespans (ie routinely several days), and if you can imagine that, you see that this requires a way of managing one's comments and being notified of others' responses in a much better way than currently possible.

Here's what I would like to see: "Your Comments" sorted by activity in the surrounding thread, eg when displaying my comments, I get a little [new] flag whenever someone commented within thread. Moreover, instead of seeing all my comments at once on the page, I'd prefer a hierarchical display of the stories or threads I've commented on (with the little [new] flag of course). If I click on such a story, I get a list of my comments on this.

Basically, it's about keeping an overview of 5-10 individual stories or threads, over a period of 5-10 days.

[ Parent ]

Unable to view/comment on stories when active (3.50 / 2) (#41)
by TON on Mon Aug 26, 2002 at 04:18:20 AM EST

Yep, I hear you. Moving from PST to JST really changed how I participate in K5. Sometimes I see something in the edit queue in the evening after work that looks good, but by the time I look again the next morning it has gotten several hundred comments. That can make for good reading, but I miss out on some of the interaction.

Ted
---
"Aye, aye, certainly the Roman Empire would still be standing if only
[
Parent ]

less nuance (4.00 / 1) (#38)
by dr k on Mon Aug 26, 2002 at 03:38:31 AM EST

"P3 Change 0-5 system to multiple choice:... I like the 0-5 with reply for nuance."

While you may like it for nuance, the fact is that the concept of nuance is totally fictional - you're just muddying the waters with highly personal justifications of a '2' versus a '3'. This kind of rating system was never intended to be used by a group of people, because on average, everything just ends up in the middle. Whenever you "nuance" your rating, you simply diminish the power of your opinion. (If you want your opinion to count - and you do, otherwise you wouldn't be rating comments - then your rating has maximum effect at the extremes.)


Destroy all trusted users!
[ Parent ]

2's and 3's (3.00 / 1) (#40)
by TON on Mon Aug 26, 2002 at 04:13:56 AM EST

I rarely, if ever, use a 2 or a 3, but 0,1,4,5 get plenty of use. I didn't mean that the numbers themselves held any nuance, but that the reply did. I'd have to go back and look, but I'm pretty sure I've given a 5 to something I disagree with because it was well written or put forward a point worth discussing, then gone on to disagree with it in the reply. Agree/disagree rating systems wouldn't allow this kind of flexibility.

Ted
---
"Aye, aye, certainly the Roman Empire would still be standing if only
[
Parent ]

the only question to ask is: (4.50 / 2) (#42)
by dr k on Mon Aug 26, 2002 at 04:28:48 AM EST

"Do you think this is a good comment?"

People are free to interpret "good" however they like. You cannot prevent them from using an interpretation that doesn't jibe with the system. Yeah, so hotly contested issues (gun control/rights) will end up with a highly split agree/disagree count... well, that just makes any such comment a somewhat average comment. Yes, it is a popularity contest. We might as well admit it.

What purpose does the rating system serve? Why, to draw out the most and least popular comments. The least popular get thrown out as spam. The most popular get more exposure - at least in theory. That's it. That's all the system does. There is no cake.

(I'd like to take a moment to repeat my request for a "Review highly rated comments" function. It is the natural counterpart to "Review hidden comments." Most Trusted Users are against this, because they know they would lose their TU status.)


Destroy all trusted users!
[ Parent ]

Voltaire is rolling in his grave. (none / 0) (#55)
by krek on Mon Aug 26, 2002 at 10:59:04 AM EST



[ Parent ]
CODE RULES HERE (none / 0) (#96)
by Swashbuckler on Tue Aug 27, 2002 at 09:27:03 AM EST

The author has described a problem. Not a problem with "code", but a problem with "culture" (as you called it). The problem he described is called The Tragedy of the Commons. That is, K5 is free for anyone to participate but when too many people do who have no real desire to uphold the environment, the whole thing crumbles.

You have suggested that his idea(s) are misguided because the problem(s) can't be fixed with "code"; they are social problems and they need to be dealt with by changing the norms of this community. I think that you are missing the point.

The idea behind K5 is that we can architect, with code, the norms that we want. Of course the problem is a social one and not a tech one; but, we believe, as K5ers, that we can mend social problems with innovative code. That is why Meta is my favorite section. Its that last technological romanticism that I'm going to hold onto; I (want to) believe that truly reflective journalism can be architected with code.




*Note* - this comment contains no inside K5 humour because inside K5 humour is only for/by K5-wankers. Media does not = "community."
[ Parent ]
Comment Pagination (3.00 / 2) (#21)
by pb on Sun Aug 25, 2002 at 11:50:27 PM EST

Comment pagination is one of those nifty features slashdot has that helps the site scale.

I'd really like to see it as on option on kuro5hin too.
---
"See what the drooling, ravening, flesh-eating hordes^W^W^W^WKuro5hin.org readers have to say."
-- pwhysall

Nooo.... (4.00 / 2) (#44)
by mozmozmoz on Mon Aug 26, 2002 at 04:32:52 AM EST

I hate that. Broadband makes single, big pages easier to get, and I much prefer to read that rather than paging constantly between bits of the comment stream. Commercial sites need it to get their adstream up, but (hopefully) kuro5hin doesn't. If this has to be, at least make it optional.

There's lots of comedy on TV too. Does that make children funnier?
[ Parent ]

I second that (none / 0) (#48)
by martingale on Mon Aug 26, 2002 at 06:18:17 AM EST

I like to keep a minimalist page which contains *all* the comment headings. Then I navigate by searching for the [new] tags and jumping directly to them, or searching for a keyword, or searching for an author. With pagination, half or more of the content would be lost and I'd be reduced to the stone age - scanning the discussions by eye. Ugh.

[ Parent ]
hmm? (none / 0) (#50)
by pb on Mon Aug 26, 2002 at 09:26:38 AM EST

What part of "optional" don't you understand?

In any case, even on broadband, it takes me at least 30 seconds to load up a 700+ comment story in nested mode, whereas it would take a fraction of the time with pagination as well as likely taking load off of k5, and I could just read a page at a time.
---
"See what the drooling, ravening, flesh-eating hordes^W^W^W^WKuro5hin.org readers have to say."
-- pwhysall
[ Parent ]

Comment modes (none / 0) (#85)
by rusty on Mon Aug 26, 2002 at 07:41:21 PM EST

I seriously recommend you review your comment mode choices. Reading a 700-comment story in nested mode is just wrong in deep and meaningful ways, not to mention pretty brutal on the servers. It's much smarter to have varying thresholds, like "nested up to 100 comments, threaded up to 200, and minimal thereafter." This way, you get nested mode when the page is small enough, and when it isn't you can pop open new threads in their own window, which will individually show up as nested. Check your comment prefs to change settings.

Myself, I use nested up to 30 comments, and dynamic minimal thereafter. Even stories with a ridiculously huge number of comments load relatively quickly.

____
Not the real rusty
[ Parent ]

I do. (none / 0) (#88)
by pb on Mon Aug 26, 2002 at 09:58:20 PM EST

I have it set up that way, but sometimes I just want to read the comments without having to expand them one by one.  Pagination would be one answer to this; another answer would be having some more options for expanding subtrees in the dynamic nested modes.

Basically, the less burden you can put on the user to do the right thing, the better, because we won't.  :)
---
"See what the drooling, ravening, flesh-eating hordes^W^W^W^WKuro5hin.org readers have to say."
-- pwhysall
[ Parent ]

Yeah (none / 0) (#90)
by rusty on Mon Aug 26, 2002 at 10:23:59 PM EST

To be clear, I'm not really opposed to pagination. I can see it being useful. I just wanted to suggest currently-working alternatives in case you (or some other reader) weren't aware of them.

____
Not the real rusty
[ Parent ]
P4 is a horrable idea. (3.00 / 2) (#23)
by dram on Sun Aug 25, 2002 at 11:57:51 PM EST

If you go to e.thePeople you can see what an implementation of P4 would be like. If you go there try to find the latest story without using the seach feature. I doubt you will be able to do it easily. Doing something like P4 would mean many good stories get lost, burried behind other older and more 'popular' stories.

Also this would allow troll and other juveniles to post pointless, crappy articles that would quickly get voted down in the current system. I for one dont want to see everything that goes through the queue on any one of the section pages, even if it was burried under better articles.

Instead we could try to get Rusty to add the most_hotlisted_stories box and then allow users to disable it if they wished.

-dram
[grant.henninger.name]

Agreed (none / 0) (#43)
by infraoctarine on Mon Aug 26, 2002 at 04:29:17 AM EST

An unpredictable user interface would be highly confusing. For instance, posting a comment to a story would mean it gets moved around somehow, but you really don't know how since it depends on a number of factors you cannot really keep track of.

The chronological system is predictable and makes a lot of sense to people.

I like your idea of a box with most hotlisted stories too.

[ Parent ]

My comments (4.00 / 1) (#26)
by zocky on Mon Aug 26, 2002 at 12:12:56 AM EST

My comments on my proposals:

P1: What I should have maybe added is: delete the editorial comments once story is posted. And maybe they should really only be visible to the author.

P2: This would allow us to continue on commenting on misspellings and catching up with friends and family while keeping both archive pages and random reader's pages (which IMHO are very important for the future of k5) free of a lot of junk, which I think is a Good Thing.

P3: I think some people misread what I suggested. Voting "I disagree" would increase the comment's score, not decrease it.

P4: I think this one is the most interesting and not hard to implement.

P5: It would in fact create stories within stories. another way to look at it is to have relationships between stories, i.e. this story is a reply to <a></a>. But I think that grouping replies inside a story will help to keep the frontpage from changing to quickly, which is on of my basic intentions.

---
I mean, if coal can be converted to energy, then couldn't diamonds?

Comments on P1 (4.00 / 1) (#33)
by bodrius on Mon Aug 26, 2002 at 12:39:49 AM EST

P1: Making the Editorial comments only visible to the author is a Bad Idea (TM).
    Not only would the author be probably swamped by redundant comments (misspellings particularly), it would mean no "Editorial Threads", which I believe is the main advantage of the Edit Queue vs Individual Editors.
    Threads of Editorial Comments allow the author and the criticizing editors to exchange ideas on possible changes and corrections to the structure of the piece. Nothing stops other editors from giving alternative suggestions, or even opposing a criticism.
    Deleting these after the story is published is, perhaps, a good idea. I would prefer making it customizable in the Display Prefs: "Show/Hide Editorial Comments", and make this "Hide" by default.

 
Freedom is the freedom to say 2+2=4, everything else follows...
[ Parent ]

Simple scaling? (4.33 / 3) (#27)
by ph317 on Mon Aug 26, 2002 at 12:15:17 AM EST


Perhaps the blog problem of scale is really a simple matter of scale.  Perhaps the right answer is that as the population grows, the standards have to raise.  If the site has three front page stories a day at 200 users after all the modding is done - when it has 2000 users it's likely to have too much crap at the same standards level.  So you raise the bar for the front page, make it take a larger percentage of yes votes - and lean more weight in favor of what is believed to be a good voter (age of account, previous posting history).  Try to make sure that at 20,000 users, you still have 3 FP stories every day, just of higher quality from a larger field.


A few problems (5.00 / 6) (#31)
by bodrius on Mon Aug 26, 2002 at 12:28:44 AM EST

I guess the cat is out of the bag and the topical discussion started, just the kind of thing you claim should be avoided. But since you're also right in that the valuable discussion will be dead by the time this gets to FP (hopefully), I'll add my insignificant portion of money:

- P1: I support this without reservations.
  It should eliminate the editorial comments under "topicals" and force us to start discussion only after the story is published.
  There are some losses: some discussions actually get interesting in stories that are never published, and they will be lost. But if it will stop the trolling, perhaps it's for the best.

-P2: Agreed with reservations.  
     I'm not sure it's such a good idea, but if it is implemented, it should be just formalizing some [offtopic] tag to the comments that could let the user identify it at sight (append it to the subject), or even choose not to display it at all (through user preferences).
     Adding more comment types than that would be overcomplicating the system.

-P3: No. NO NO NO NO NO.
     To be more specific: comment ratings should be kept as simple as possible so as to make rating a polling system with high participation, not an editorial system with low participation or, even worse, an editorial system with high participation.
     The fact is that, although there is constant abuse of the rating system (1s and 0s for diferring political opinions), the overall effect is under control because most of the population ranks by overall quality.
     Qualified Rating will not solve the problem. People already know they should rate by comment literary/logical quality, not by subjective bias. I'm skeptical that a qualified system will work any better: look at how it "works" at That Other Site (TM)!
      You'll get meta-comments on the moderation's merits ("that comment was not 'Funny', it was 'Insightful'!"), comments that don't fit into your categories ("we have to add a new, 'Sarcastic but Not Funny', category"), and trolls pushing the system.
      How exactly would you rate, say, a typical medham's comment? It's not "Badly Written", it's rarely "Offtopic", almost never "Spam", and "I disagree" doesn't exactly express the purposeful misinformation of the comment, as it doesn't distinguish from a well-informed opposing view.

   Perhaps it's worth noting that Kuro5hin's seems to be very effective at controlling commentary trolling when compared to many weblogs, including That Other Site(TM), that use qualified comment rating.
   If we're going to try to "fix" whatever is "wrong" with Kuro5hin, we should look at where the trolling has become the biggest problem: the Edit and Vote queues.

- P4: This looks like a massive re-working of a system that is relatively functional into an experiment that may or may not be as functional. And not a very convincing experiment at that.

- P5: Ugh. No. Agh.
   There is no reason to distinguish between the "chat" and the "real discussions".
   That's not how interesting discussions develop in real life, and we have no reason to think it's so different in kuro5hin.
   Many "replies" develop into "chat", and many "chats" with one-line questions/comments that seem simple enough develop into "replies" and full-screen analysis, that is, "real discussions".
   Nothing stops a one-line comment or joke from being more insightful, and more worthy of being read months later, than a full-screen sub-story posted as a "reply".
   The disctinction is unnecessary, because the rating system is working. If you are really desperate to let your comment live forever as fully-developed analysis of the topic in its own right, submit it as a story. It would not be the first time.
Freedom is the freedom to say 2+2=4, everything else follows...

Add bells and whistles, listen to the noise (4.20 / 5) (#32)
by FlipFlop on Mon Aug 26, 2002 at 12:30:38 AM EST

At least add "off-topic". Make it easy to post different types of comments, for instance, add an "add an off-topic comment" link to the comment footer.

The trouble is, if you add a new feature, people will use it. Over on Slashdot, I'm guilty of this. I post all kinds of stuff anonymously that I would never post under my phoney username.

If you add comment types like rant, off-topic, or funny, people will post more rants, off-topic and funny comments because the system gives those types of comments a stamp of approval.

Personally, I like the high-rated funny comments at Slashdot, but for every funny comment, there's a dozen other failed attempts at humor.

AdTI - The think tank that didn't

editorial (4.00 / 2) (#34)
by delmoi on Mon Aug 26, 2002 at 01:02:08 AM EST

Disable topical comments until story is posted. Make editing period of 24 hours mandatory. Allow the author to add clearly marked text to the end of the story even after it's posted.

Actualy, thats the way things used to be. It didn't work. People just want to comment on stuff in the queue.

Really there are two kinds of editorial comments the "fix this editing error" and "vote this down, because:" stuff. The 'vote this down because:' should probably stay. Maybe we need another comment type.
--
"'argumentation' is not a word, idiot." -- thelizman
24 hours! good start (2.00 / 1) (#36)
by turmeric on Mon Aug 26, 2002 at 01:56:13 AM EST

the limit should be like 48 or something. it should also be clearly easily stated after u 'submit to editing queue' how long you have to rewrite it before it goes to voting. that would be 'like butter'

Mandatory editing? (4.00 / 1) (#45)
by Mwongozi on Mon Aug 26, 2002 at 05:36:03 AM EST

Make editing period of 24 hours mandatory.

Why?



abuse of (4.00 / 1) (#46)
by guyjin on Mon Aug 26, 2002 at 06:13:16 AM EST

"
Change "0-5 system to multiple choice: "I agree (+1)", "I disagree (+1)", "Badly written (-1)", "Spam (hide)"
"

The voting system would still be abused. People who disagree would choose 'badly written' or 'spam'. I think that a system like this could work if the user's votes should be weighted depending on their mojo - trolls who mark things as spam that arent loose mojo, and their votes become less relevant.
-- 散弾銃でおうがいして ください

Voting System (4.00 / 2) (#49)
by Kamui on Mon Aug 26, 2002 at 06:38:52 AM EST

I think such a voting system would at least reduce the number of "careless" votes from people that have never really thought about the fact that voting something down to "1" because they disagree isn't at all what was intended with the voting system in the first place.

However, I think that Mojo doesn't reflect the user's level of reliability as accurately as it should - some comments get lost (many really good comments never receive a single vote, or at least don't for a long time), many people abuse their ability to vote and so on, it's just too "blurred".

There should, however, also be a way to judge a comment "very well written" or "extremely good".

[ Parent ]

Fix the search first! (4.75 / 4) (#51)
by wiredog on Mon Aug 26, 2002 at 09:27:34 AM EST

Until that is fixed, nothing else should be messed with.

Can't sleep. The clowns will get me.
really, i don't get it.. (none / 0) (#56)
by infinitera on Mon Aug 26, 2002 at 11:26:05 AM EST

why is the search so important? Are you all so vain that you must search for occurances of your name? The only time I've ever had a use for with the search was finding links for my own posts, and I manage to do that just by remembering the story they were in and the poster's name, and googling - that's always worked better anyways.

[ Parent ]
To allow authors to see previous stories (5.00 / 1) (#57)
by HidingMyName on Mon Aug 26, 2002 at 12:06:08 PM EST

Sometimes a topic is revived, e.g. I recently posted a story about Dr. Al-Arian at USF ( Tenure, Terrorism and Academic Freedom) which was also discussed by mehdam in December 2001 . Not being able to look up the prior story meant missing a citation, which lead to a misunderstanding . Although there was no plagiarism, I should have been able to easily find his work and to cite it as part of the routine literature search.

[ Parent ]
To find old stories, duh (none / 0) (#77)
by Josh A on Mon Aug 26, 2002 at 04:39:56 PM EST

Sometimes I want to refer to old K5 stories in chat or email, etc. Relying on Google is not the answer. As Jakob Nielsen never tires of trying to brainwash people: search is the most important function on a large site.

---
Thank God for Canada, if only because they annoy the Republicans so much. – Blarney


[ Parent ]
About search [a tad off topic] (none / 0) (#63)
by bob6 on Mon Aug 26, 2002 at 01:02:31 PM EST

Does anyone know what's going wrong with the search? Indexing? Ovrload?
I thought we could index stories manually with a few but very relevant keywords. Authors and editors for academic litterature have been doing this successfully for a long time. It would be a relatively low-cost and efficient way to index (therefore to search for) stories.

Cheers.
[ Parent ]
mysql (none / 0) (#66)
by pheta on Mon Aug 26, 2002 at 01:30:09 PM EST

The most likely cause is that mysql doesn't have a (good) full text indexing mechanism (this may have changed in the last year or so, anyone know better?). So whomever choose to code it would need to write their own indexing mechanism and that doesn't seem to have been done yet.

[ Parent ]
Sort of (none / 0) (#82)
by rusty on Mon Aug 26, 2002 at 07:16:40 PM EST

MySQL does have a functional (though not ideal) fulltext index, in MyISAM table types. It does not in InnoDB tables. Our problem is that we need to use InnoDB for the speed advantages it provides.

The solution in the works is mirroring some fields of comments and stories in MyISAM tables, with fulltext indices, where the actual searching is done. Since we can update these tables in batches from time to time, it should be possible to overcome the speed limitations that would be a problem if we used them for all posting and reading functions.

If anyone with a serious yen for search knows something about perl and Scoop and wants to fix this yourself, please contact me. Otherwise, it's will get done when I get it done. It hasn't been abandoned or forgotten though.

____
Not the real rusty
[ Parent ]

a faster, simpler and better solution - imho (none / 0) (#86)
by kraft on Mon Aug 26, 2002 at 08:17:36 PM EST

I don't know anything about scoop or how you structure k5.

HOWEVER I do have some experience with this, and I wouldn't recommend full-text indexing. It's waaay too slow, even with innodb. I have tried it on a mere 8,000 documents, and I'm sure you have much more here on k5.

What I have done with my site is pretty simple.
1. I made two simple script. One generates a link list in html of all the documents I have (script 1). Each link points to script 2 by the document's id. What script 2 does is simply show this document in the most simple way. No banner, header or footer, just text.
2. install htdig.
3. start indexing, but make sure you point the start url to script 1.
4. make a new search script. It will use the htdig index. When htdig returns a result, it points to the wrong script, namely script 2. Now take the result, and strip everything but the id (ie. strip the url to script 2).

And voila - you have a search engine that returns the ids of the documents you want.

Maybe I didn't write this very clearly, just mail me if you need to: jens@cheathouse.com

After I had this idea (inspired by how they do it over at php.net) I implemented it in two days. It's wicked fast and just works.

The documents are indexed once a day.

--
a signature has the format "dash-dash-newline-text". dammit.
[ Parent ]
Probably not (none / 0) (#89)
by rusty on Mon Aug 26, 2002 at 10:22:16 PM EST

It's waaay too slow, even with innodb.

InnoDB has no fulltext index capability.

About your htdig idea: how would that cope with searching individual comments? Right now we have a total of 672963 comments and 52978 stories. So htdig would have to plow through 725941 documents the first time through. Notwithstanding the time it would take to pull each of these out of the DB one at a time in the first place (nearly a nonstop 24 hour day, at a rather optimistic 10 pages per second estimate), and assuming it would index from then on in a "what's new since last time" fashion, there's still the disk space issue. Htdig's estimate for space is (# of docs * 12000) bytes, which means we're looking at an index right now of maybe 8.7 Gb, which would require a new hard drive in each Scoop machine just to index.

I think doing it in the DB would be smarter.

____
Not the real rusty
[ Parent ]

Why full text? (none / 0) (#94)
by wiredog on Tue Aug 27, 2002 at 09:04:28 AM EST

Why not keywords? I assume you don't index on 'and' 'or' 'what' <Other_common_words>.

Can't sleep. The clowns will get me.
[ Parent ]
Fulltext (none / 0) (#95)
by rusty on Tue Aug 27, 2002 at 09:17:49 AM EST

"Fulltext index", in this context, means an SQL mechanism that MySQL provides of searching for words within text. It basically is a keyword index, just a lot less work for us because MySQL does most of it internally.

____
Not the real rusty
[ Parent ]
The hell you say.. (3.00 / 1) (#53)
by ignatiusst on Mon Aug 26, 2002 at 09:56:32 AM EST

If the problem really were, as the author suggests, the impact of mere commoners moving into places they were not wanted then the problem could be simply solved. Rusty could just restrict access to all the bells-and-whistles of his site and leave the common trash out in the cold looking in, but never participating.

But, really, that is nothing more than elitist bullshit. Serious consideration of creating an environment that propagates the kind of intellectual snobbery that is being proposed here should be shown the door with as much scorn as we can muster for the occasion.

I am not saying that there aren't improvements that can be made to k5, and I am not suggesting that there is no place on the internet for discussion devoid of juvenile commenting and trollers. I am certain, though, that we should never consider turning k5 into the equivalent of a WASP-only country club.

When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. -- Jonathan Swift

I don't understand (none / 0) (#78)
by Josh A on Mon Aug 26, 2002 at 04:54:04 PM EST

How do you see any of the author's suggestions "turning k5 into the equivalent of a WASP-only country club"?

Where did the author imply that the problem was "commoners moving into places they were not wanted"?

As far as I can tell, the stated problem is too many people (hence the use of the word "scale">... regardless of what kind of people they are... you seem to be looking at the author's quantity problem and denouncing it as a quality problem.

---
Thank God for Canada, if only because they annoy the Republicans so much. – Blarney


[ Parent ]
My observations.. (none / 0) (#81)
by ignatiusst on Mon Aug 26, 2002 at 07:07:05 PM EST

To begin with, the author has already begun building the metaphor of a country club with Blogs have a scaling problem. Kinda like clubs..

Then, s/he builds most of the dichotomy between the elite and the commoner with, The good crowd moves in and they become these perfect little places for some time. And then too many people start coming in, and the magic disappears. It seems to me, especially in the context of the rest of the intro that "too many people" is a pleasant way of referring to the "commoners".

S/he then completes the nose-thumbing of (what I have termed) commoners by equating the new (presumably less desirable) atmosphere to that epitome of common life, the fast-food restaurant: Teenagers take over and after a couple of years place is converted into a bad fast food restaurant. And we won't even mention the eltism that goes into such a loaded statement about teenagers.

But, this is all just my interpretation. Obviously there are others.

When a true genius appears in the world, you may know him by this sign, that the dunces are all in confederacy against him. -- Jonathan Swift
[ Parent ]

Still doubt it (none / 0) (#106)
by Josh A on Fri Aug 30, 2002 at 11:01:55 PM EST

All right, I see where you started from at least... but I am still not sure how you got there. I guess we just each read it differently.

If the author had instead written Blogs have a scaling problem. Kind of like real life groups. (which is how I took it) we may have never ended up on this topic. Real life groups do have a scaling problem. But a night club is hardly a country club, whether or not night clubs also have scaling problems.

In the second bit you quote, the quantity issue continues to be explicit (then too many people start coming in) and any quality issue that may be there remains, at best, covert. All I can say is that if the author did have the intentions you state, s/he didn't express them very clearly. Or, s/he covered her butt well ;)

The "teenagers" crack did bother me, but I don't think the author meant to equate the new atmosphere with fast food restaurants... I think the author was trying to say because of the new atmosphere the club goes out of business and someone opens a fast food joint in its old building.

---
Thank God for Canada, if only because they annoy the Republicans so much. – Blarney


[ Parent ]
More people (2.00 / 1) (#58)
by jmzero on Mon Aug 26, 2002 at 12:16:56 PM EST

Also, there are many good reasons to not include those in the default random-visitor's view of kuro5hin.

You seem to assume that kuro5hin should grow.

I don't really want it to grow.  Right now it's a bunch of intelligent people, me, and E   R  I C.  If lots more people come on, then there's a real good chance we're going to get a lot more stupid comments.  Then we'll be SlashDot, where nobody reads comments (because there's too many and they're too stupid), so everyone just sort of shouts into the void.  And everyone complains about the rating system.

I say we make K5 as intimidating as possible to new users.  The next logical goal would then be to drive off me and E R  1      C.

.
"Let's not stir that bag of worms." - my lovely wife

Re: Slashdot (5.00 / 1) (#97)
by FourDegreez on Tue Aug 27, 2002 at 02:56:13 PM EST

I suppose the Yogi quote applies: "Nobody goes there anymore. It's too crowded."

[ Parent ]
Mandatory Edit Queue... (3.00 / 1) (#59)
by br284 on Mon Aug 26, 2002 at 12:17:55 PM EST

... As I've stated before, I hate the edit queue. Therefore, I also hate the idea of a mandatory edit queue. I figure that when I post an article, I might as well do the community a service and post a finished one that has been spellchecked and grammar checked. I prefer to do my writing all by my lonesome and without the community's involvement. I like the feedback, but not before I'm finished.

Interesting suggestions otherwise.

-Chris

Why don't you just rip off all of Plastic's ideas? (3.00 / 1) (#64)
by Genady on Mon Aug 26, 2002 at 01:03:56 PM EST

Is it just me or do most of these proposals sound like the way Plastic does things? You know what, they don't work over there either.

This is a cultural battle pure and simple, between the people that want to change their community to suit their desires and the people that don't.

--
Turtles all the way down.

the problem isn't solved by technology (4.25 / 4) (#65)
by 5linky on Mon Aug 26, 2002 at 01:18:21 PM EST

but by fragmentation.

In any society / club / group, as they grow then people won't like what happens to the group. The more people the less personal it becomes and the more mainstream it becomes. This is what is happening on K5. The solution off-line is that a small group of people will leave the group and set up their own group, which will start off small and then probably grow as well, and it starts again.

It's a human / social thing which won't be fixed by technology. No matter how you change the voting rules or anything, the same problems will still be around. All you are doing are masking the problems not addressing them. You are blaming the technology. This is like politicians using violence in films as a defletion to the real problem of violence in society.

Like in the UK the big dance clubs are starting to lose customers because they are now the mainstream and the public want to goto smaller clubs which aren't the mainstream. Something which started small, the rave scene, became large and corporate, and has now fragmented again into smaller underground factions, which in their turn will become the mainstream ad infinitum.

Fragmentation is the answer. Some people will stay at K5 because they like it, others, if they really don't like it here, should either start their own Scoop sites or find other smaller ones. But technology won't solve your problems. The web might not be "real" to some of you but the principles of community are still the same.



But for the evidence (none / 0) (#74)
by khallow on Mon Aug 26, 2002 at 04:17:36 PM EST

I hate to break this post up, but you seem to neglect the social effects of such technologies as television or the telephone. While I don't see any advance of technology helping you get "personal" with a billion other humans, it seems likely that technology does aid in the social processes you describe. For example, is it easier now to find a few dozen people with common interests than it was fifty or a hundred years ago? I assert it is much easier now. The Internet allows for a degree of social fluidity that isn't possible in the old days.

You also seem to neglect the ability of technology to induce fragmentation in an existing group. From such simple things as the ability to ignore a selected user or subject to sophisticated preferences or ratings, there are ways to cut out what parts or people of the group you don't like. Ie, social fragmentation can be built in.

Stating the obvious since 1969.
[ Parent ]

technology can help (none / 0) (#75)
by speek on Mon Aug 26, 2002 at 04:21:32 PM EST

You are right that there is a maximum size to a working community. But, you're wrong that the problem can't be attacked with technology.

K5 could be divided into groups of users. The simplest way to divide users is by user id number. Any random method will do however. Determine some fixed group sizes. For instance, the smallest size group could be 100 users, or 1000 - whatever is determined to work best. These groups are then grouped. So forth and so on, as much as is deemed necessary. At first, just one layer is sufficient.

How does this help? Submitted stories are seen only by your group-mates. The group votes it's own stories up. Stories posted by the group are still only visible to that group on K5's main pages, but now, it becomes a submitted story that all groups can see and vote on. Another voting round, and the story is either posted universally or left as specific to that group.

Individuals can no longer bother anyone but their group mates. Only an entire group can succeed in getting stories in front of all of K5. No matter how large K5 gets on the whole, group size stays fixed. If the number of groups gets too large, another layer is created.

This is what you do if you're serious about K5 as a community. Group members will bump into each other frequently, and get to know each other well. Ties between groups will be formed in K5's universal section.

If K5 is viewed as a news service, then this is not a good idea. In that case, you could do something similar by randomly weeding stories on an individual basis. For instance, when a user submits a story, only 1000 users can see it, perhaps the next 1000 to view the story. Thereafter it disappears until those 1000 vote it up, at which point it becomes viewable by everyone. Same idea, but the boundaries aren't fixed. Fixed boundaries promote community, non-fixed is just a scaling thing.

But, in any case, fragmentation is achieved, without a lot of the damage that usually goes with it.

--
what would be cool, is if there was like a bat signal for tombuck - [ Parent ]

thoughts (3.50 / 2) (#67)
by Shren on Mon Aug 26, 2002 at 01:34:00 PM EST

Proposal P1: Rework the editing

The basic idea behind editing is that a flawed story, with the input of a hundred monkeys, can suddenly become a quality story. In a sudden moment of enlightenment, I've decided that's probably not true. Wise authors make the changes suggested by editors and thus garner votes, but does it really make the story better? It seems to me like it just allows more people to piss in the soup before it's served.

My new opinion : Edit it yourself. Maybe a mailing list for this purpose would be nice. But the delusion that the edit queue makes quality articles out of bad articles is just that - a delusion. So all the edit queue ever actually achieves is minor bits and pieces, which hardly need the system, number of eyes, and controversy of the edit queue.

I hearby reverse my calls for a longer edit queue and call for an edit queue time of 0 seconds. Thank you.

3

Proposal P2: Add more comment types

It's called a diary.

1

Proposal P3: Refine ratings

This is an arbitrary change which has nothing to do with scaling - you're bundling a feature you want with a bunch of other suggestions. This isn't the US Congress.

0

Proposal P4: Merge hot-list and story ratings

Interesting. I don't use the hotlist features at all, so I really can't comment.

none

Proposal P5: Restructure the discussion

I think you have a flaw here. You postulate a new system and a desired result. It doesn't seem that your new system will result in your desired result. I'd say that it would break down but it doesn't look like it would even achieve your desired result for any period of time.

1

Response to P3 (4.50 / 2) (#69)
by curunir on Mon Aug 26, 2002 at 02:41:58 PM EST

I think you're right to point to the moderation system needing improvement. However I think you go about it in the wrong way. Your proposal is basically advocating the moderation system on The Other Site TM. It's clear from their example that that sort of moderation is just as prone to abuse as scoop's is.

Yet there has to be some way to quantify the fact that a (4.90/10) has received better moderation than a (5.00/1). What I would propose is something akin to how ice skating is judged in the Olympics. In that case, the statistical outliers are dropped from the rating. So imagine the case of the (4.90/10). If you dropped the highest moderation and lowest moderation, you'd be left with the equivalent of a (5.00/8) which would rate higher than the (5.00/1) This would also limit the effect that one angry moderator can have by dropping a 1 on a post. Instead of bringing down the average for the post, it would just negate one of the many higher moderations leaving the rest intact.

Obviously the thresholds for this would have to be played with. I'd try something like dropping a high-low pair for every 6 posts to start with...keeping the core 2/3's of all moderations. But I think that overall, this would make scoop's moderation system less prone to abuses by a single moderator.

simpler method available (none / 0) (#73)
by speek on Mon Aug 26, 2002 at 04:02:31 PM EST

It would be simpler to start comments at a medium rating (say 3) with a certain weight (say 10). Then, begin averaging as normal from there. A 5 rating would bring the average up to (10 * 3 + 5)/11 = 3.18. The next 5 rating increases it to 3.33. This works best if you remove all the excess options (ie 2,3,4).

--
what would be cool, is if there was like a bat signal for tombuck - [ Parent ]

Ratings are stupid anyway (5.00 / 3) (#70)
by majubma on Mon Aug 26, 2002 at 03:01:05 PM EST

 The ratings system we have appears to be deliberately obfuscated.
Rating on kuro5hin is set up as a democratic process. As I've discussed before, participating in a democratic process entails making actions based on the effects they will have.

Ratings have four main effects, as follows.

* Ratings determine whether a comment is hidden from view. Comments are hidden when their average rating is less than one.
* Ratings determine whether a user is "trusted." Users who post highly-rated comments gain the ability to view comments that were rated zero, as well as to give zero ratings to comments themselves.
* Ratings may affect how users view comments, depending on their sort and display preferences.
* Ratings may cause complaints and extended whines in the diaries if others percieve the ratings to be "unfair."

When rating a comment, one should obviously choose a rating that has the desired effects. The problem is that the effects of rating a comment are only marginally connected to the actual quality of the comment in question. Particularly since a large number of users do not actually use the ratings in their display preferences, (well, I don't, at least) and the number of comments that actually wind up being hidden is exceedingly low. The most powerful effects of rating a comment is that of affecting its author's "trusted" status and making him post extended rants in the diary section.

Then, according to basic pragmatism, when I rate a comment I should consider whether I want the user to be trusted or not, and whether my rating will provoke outrage. The problem is that a user's comments are a poor indicator of whether that user deserves trusted user status. A better indicator would be the user's rating history.

So, you can look at the user's rating history, to see if they rate acceptably... The first problem is that if they're trusted and you're not, you don't get to see what they've rated zero! How am I supposed to decide whether a user deserves to keep his trusted status, without this information? Further, it's a real pain in the ass to have to run through a user's ratings history in order to rate their comments properly.

The second problem is that the recipients of ratings expect their ratings to be indicative of
the "quality" of the comment, rather than any reasonable pragmatic expectation. If I find a user who IMO abuses the zero rating (never mind that I have to have trusted myself in order to do this) and attempt to correct the situation by giving him low ratings, I'll incite diary-section whines and retaliatory ratings.

When weighing all the consequences of rating a comment, I find that there is no incentive for my to rate comments. In summary:

* I don't sort based on comment ratings, and I'd actually rather other people didn't either.
* I can't recall coming across across a visible comment that I thought should be hidden.
* I don't want to incite people to post junk to the diary section.
* It's too hard to review other's rating histories in order to decide whether they should be trusted.

THEREFORE:

Ratings, at least for this user, are broken. Therefore, I stopped rating and stopped paying attention to ratings. This has improved my experience of kuro5hin.


--Thaddeus Q. Thaddelonium, the most crookedest octopus lawyer in the West.

ratings (none / 0) (#91)
by akb on Tue Aug 27, 2002 at 12:06:59 AM EST

Great post, I disagree :)

You lay out very well what I take to be a case for why the combination of the rating system and comment  system aren't working.  I'd like to suggest why they aren't working and maybe a way to help.

Basically, I think there needs to be more feedback between ratings and comments.  You make the case for there not being a strong connection between the two.    I think it would be worth k5 consider limiting comment posting based on the ratings system.

I see a lot of people that exclusively rate comments  1 or 5.  They aren't rating the comments based on how they add or detract from a discussion, they are voting to maximum effect what they agree with.  I also see a lot of off topic, low content posts, such as those that plague the other site.

If there were a cost to rating someone extremely one might consider more carefully when giving a rating.  By rating you would be "spending" your ability to make comments.  Conversely, if you make lots of low content posts, rather than a few high content informative engaging posts, you would be limited in how many ratings you could make.  The heuristics to get this halfway decent seem feasible but it would be very difficult to measure.

Of course its possible that this proposal might just encourage the factionalized block voting that seems to be occuring, making this all a numbers game for people to manipulate.

The really important value in making a discussion community that doesn't suck is listening, and you can't play a numbers game to enforce that.

Collaborative Video Blog demandmedia.net
[ Parent ]

Why can't we hide people? (4.00 / 2) (#71)
by bigbtommy on Mon Aug 26, 2002 at 03:45:48 PM EST

To make the system simpler, why can't we just hide stuff from authors we don't like. That would be especially useful for the diary section...
-- bbCity.co.uk - When I see kids, I speed up
NIMBY (4.50 / 2) (#92)
by codemonkey_uk on Tue Aug 27, 2002 at 05:19:56 AM EST

The argument against "kill-lists" for this kind of site is that it is a "Not In My Back Yard" answer that hides, rather than solves, the problem. What could happen if you implement it is you create a barrier to entry where the more experienced / long term users of K5 have extensive kill-lists, but new users are exposed to a sea if spam.

That and the fact that it just doesn't work. All the noise comes from new users (or should I say, new accounts). The genuine people who hang around for a while tend to out grow the irritating newbie behaviour, and in the end add value to the site. A kill list would condemn them forever.
---
Thad
"The most savage controversies are those about matters as to which there is no good evidence either way." - Bertrand Russell
[ Parent ]

maybe still workable (3.00 / 1) (#99)
by Shpongle Spore on Tue Aug 27, 2002 at 04:11:32 PM EST

Both of the problems you mention look pretty easy to solve. First, give new users (and visitors) a default kill list consisting of everyone who is on N people's kill lists. Adjust N until it seems about right.

The second problem it solvable by making kill list entries expire automatically after, say, a month. That way you have never have to see an annoying person's posts more than once a month, but everyone gets as many second chances as they need to start acting mature.
__
I wish I was in Austin, at the Chili Parlor bar,
drinking 'Mad Dog' margaritas and not caring where you are
[ Parent ]

kill lists don't change behaviour (4.00 / 1) (#101)
by martingale on Tue Aug 27, 2002 at 08:22:24 PM EST

I object to kill listsfor the following reason: they don't produce feedback to the trolls you want to be protected from. Basically, in a world full of individual kill lists, there is no incentive for trolls to reduce their output, since their audience consists of those "lucky" few whose kill lists don't (yet) include them. In fact, with everyone having a kill list, I expect that the most profitable path for trolls would be to use an immense number of throwaway accounts: post a troll, get added to everyone's kill list, create new account, repeat.

A system which encourages people to vote on (read against) trolls is I think much better, since at least there is feedback. If you want the equivalent of a kill list, just browse at +5. (which won't get rid of every troll either, but certainly penalizes unimaginative trolls). Also, don't read the queue, which has a much higher ratio of abusive stories.

[ Parent ]

Read flags (4.50 / 2) (#72)
by gidds on Mon Aug 26, 2002 at 03:46:22 PM EST

One of the main disadvantages of blogs like this, and the strongest factor in stories dying IMO, is that if I come back to a story later I get all the old comments along with any new ones. Okay, I can sort by date, but that doesn't interact well with threading, and still leads to huge web pages.

If you don't mind a brief plug, I'm a member of CIX.  I spend more time there than on all other discussion sites together, and one of the main reasons (apart from the strong community spirit, interesting people, etc.) is that discussion is properly threaded and each message has a `read' flag.  Some discussions last months, but I've no trouble keeping track because the software does that for me.  This gives space for discussions to mature, and emphasises content over speed.

Andy/

Um? (3.00 / 1) (#80)
by rusty on Mon Aug 26, 2002 at 07:06:51 PM EST

We have threading and "read" flags. I don't understand the distinction you're drawing.

____
Not the real rusty
[ Parent ]
It looks to me like (4.00 / 1) (#83)
by aphrael on Mon Aug 26, 2002 at 07:25:11 PM EST

he wants the option to only be shown new messages.

[ Parent ]
read flags, etc. (4.00 / 1) (#100)
by Shpongle Spore on Tue Aug 27, 2002 at 04:18:10 PM EST

Better control over the read flag is kind of a big issue for me as well...I keep thinking I'll post a meta article about it one day except it would be too short.

Basically I wish there were some way to reset the read flag on your messages--maybe just enter a date and make all messages past that date unread. This would be really helpful in those situations where you load a page and don't finish reading it for whatever reason. Once you come back to the page and reload it the read flags are no good since you haven't actually read all the "old" messages.

Another nice feature would be a "next new message" link on every post; it would be a lot nicer than having to use the browser's search feature to find new posts.
__
I wish I was in Austin, at the Chili Parlor bar,
drinking 'Mad Dog' margaritas and not caring where you are
[ Parent ]

P1 and mojo and/or $$$ (4.00 / 2) (#76)
by upsilon on Mon Aug 26, 2002 at 04:23:39 PM EST

Here's another thought for the problem that P1 attempts to address... what if the edit/vote queue were simply off-limits to all but those with sufficiently high mojo? For example, only the people with enough mojo to zero-rate comments were allowed into the editing and voting queues. (This needn't be the criterion used; I just suggest this one because it's already in use elsewhere on the site.)

Mind you, I speak as one who does not have enough mojo for this... but I would not mind being barred from voting and suggesting edits if most of the site were likewise barred... Then when stories leave the queue, conversation can really begin.

I suggest this because it seems that the mojo barrier is there to make sure that only the "best and brightest" of K5 can zero comments (for some definition of "best and brightest", mind you). It only makes sense that those people -- the people that the site seems to collectively respect the most -- vote on the stories.

Alternatively, maybe only people with "premium memberships" get access to the queue... that, I think, would spur a lot of money headed Rusty's way. Maybe a combination -- premium members plus the high-mojo "free" accounts...

There are, of course, two significant downsides to this proposal that I see.

  1. Elitism. K5 seems to pride itself on being a site where the users choose what's important. Suddenly, with the enactment of my idea(s), that's no longer true. K5 becomes a site where a significant subset, but a small subset chooses what's important for everyone else. As such, I don't know if my proposal is worth serious consideration, but I'll still present it for thoughts.
  2. Dead conversations: When a story is "revealed to the masses", they come in and see all the comments written by the high-mojo folks, and maybe all the things that deserved to be said have been said. As such, maybe this proposal would do more conversation-killing than -spurring... I don't really know. The alternative that occurs to me is that when a story leaves the queue, all comments on it get dropped (or hidden, at least)... which may, indeed, provide more impetus for the queue to be about editing and not serious conversation on the topic. This may not be a bad thing, after all...
That's really all I have to say. I've already spouted a good deal of crap (and hopefully some good ideas) in this comment... maybe more later, if anything else occurs to me...
--
Once, I was the King of Spain.
Responses (4.50 / 4) (#79)
by rusty on Mon Aug 26, 2002 at 07:04:47 PM EST

Editing: Your editing proposal is sort of good, I think. Yes, we need more editing time. There's functionality in Scoop now to make that much easier. I would implement that instead as a clear division between the voting queue and the editing queue (make them separate things), and as an optional ability for people to edit longer, while still allowing the readers to de-spamify the edit queue. At least part of this is coming soon.

Comment types: Why would we want to encourage off-topic comments by officially sanctioning them? At least now there's a disincentive to posting off-topic. There are also facilities available to every user (diaries) where the very concept of "off topic" doesn't really exist. I don't see how encouraging offtopic comments elsewhere would be constructive.

Ratings: Yes, 0-5 is simple, and also it's not entirely clear what you're supposed to be rating on. This is on purpose. Trying to restrict rating criteria will never actually force anyone to rate according to the criteria suggested. It will merely allow us to pretend that ratings mean more than they really do, or mean a specific thing that we cannot possibly actually prove they mean.

The reason ratings "meaning" is vague is precisely to avoid the impression that ratings mean something they can't mean. The "meaning" of ratings is some agglomeration of many people's judgements on agreement, quality, appropriateness, and sometimes like or dislike of the author. That's all they'll ever mean, no matter what linguistic sugar-coating we wrap them in, and not assigning a specific "meaning" to them at least hopefully helps people avoid thinking they mean something that they don't.

Story Ratings: I'd like to allow rating of stories like rating comments, for informational and perhaps ordering purposes. But I don't think that could or should replace voting. Story voting is a major part of what K5 is. You could design a system that doesn't work that way, but it would not be K5.

Voting produces a very clear result. Some things are posted, some are not. Rating would allow any story to be posted, and anyone to read anything that had ever been posted. I know the dominant ethos of the net is that everything should be individual preference, but that isn't what K5 is about. What gets posted and what doesn't is purposely a firm decision made by majority-rule. Even if hundreds of people want a story posted, hundreds more can overrule them. This is not an accident, or a weakness. This is a feature. It's a basic principle of the site.

Discussion restructuring: I don't really understand this. You seem to be trying to correlate quality with length, which strikes me as faulty. But I may be misinterpreting, since I don't really get the idea.

____
Not the real rusty

Improving the way comments are shown would help (4.00 / 1) (#93)
by netman on Tue Aug 27, 2002 at 07:24:14 AM EST

But "too many people" shouldn't make sense in the context of a blog. More people is good because more people write more good text. It's just that finding it in the megabytes gets somewhat tedious. After a time the proportion of interesting people and text becomes too low to bother. Take a look at usenet. Take a look at The Other SiteTM.

Well, I think K5's story comments are easier to read then Slashdot because it tells us which comments are new and which aren't, I really love that feature.

However, I find Usenet much much easier to read.

I think that if there was some sort of optional Usenet-client-like Java, JavaScript or XUL app embedded in the comments view it'd be much easier to browse through them.

Just my 1 cent

Order is for idiots. Geniuses can handle chaos.

Noise vs. signal (3.00 / 1) (#98)
by BLU ICE on Tue Aug 27, 2002 at 03:06:07 PM EST

Maybe untrusted users should be banned from posting diaries. I have noticed a lot of diary spam lately. (Of course, my diaries aren't spam. ;-) )

That along with the existing hidden comments system should cut down on the trolling quite a bit.

"Is the quality of this cocaine satisfactory, Mr. Delorean?"
"As good as gold."

-- I am become Troll, destroyer of threads.
It's like an encyclopedia...sorta: Everything2

maybe you don't want a diary then (none / 0) (#104)
by tealeaf on Thu Aug 29, 2002 at 01:30:46 AM EST

See, a diary is to put whatever you want there.  That means if you wake up one day and the only comment you have to make is "my toiled is clogged!", then you can put this kind of crap in your diary.  That's the point of a diary.  A diary is not to be censored or editorialized.  It's an outlet for the inner light/crap that we have and that's how it should be.

If you make posting of the diary a "priviledged" thing, then it's not a diary anymore.  At any rate, who is to decide what is crap and what isn't?  You think an "untrusted" used cannot post a really insightful diary entry?

I think diary should be 100% open, uncensored, unmoderated.

I scan all the subject lines and only read those diary entries that have interesting subject lines.  So if you see an entry with subject "Blah", just skip it right over.  That's what I do and I get great results.  I hardly read any spam in the diaries, but then again, I don't read many diaries because most people have nothing interesting (to me) to say.  Like I care about someone's bad hair day or someone's girlfriend problems.  Boring.  But sometimes someone's problems are interesting (to me).

So, again, I have no problem with diaries as they are now.  The subject line pretty much says it all.  If someone doesn't put enough energy into the subject line, I don't care what they have to say in the rest of their diary entry anyway.

I think what would be really cool is if we could subscribe to certain diaries.  BTW, if this is already a feature for "paid subscribers", sorry.

[ Parent ]

Limiting posting in the voting queue (4.00 / 1) (#102)
by Arkaein on Tue Aug 27, 2002 at 10:40:07 PM EST

I think there needs to be the ability to comment on stories in the queue, but I also agree that a way is needed to keep the stories fresh by the time they reach the main pages, so that it is not necessary to wade through the garbage in the queue to get in on discussions before they grow stale. The 36 hour maximum time is necessary to allow stories posted at any time of day to be treated more or less equally, but in many cases a decision is easily made before that threshold is reached.

I would be interested in ways to speed up story approval or rejection without significantly affecting quality. Polls regularly estimate approval or sentiment with only a small proportional sample relative to the population. As more users join the site it seems that it should be possible to post stories sooner based on the principle that votes will add up faster, yet will likely be just as representative of the concensus attitude.

To take this idea further, it may be worthwhile to do some kind of statistical analysis to see how many votes are really necessary to predict with only a small margin of error whether a story will be posted under the current system. If the number of votes needed is significantly smaller than what is used now, stories could be posted sooner with no other discernable effect. I've only had two stories posted myself, both to section, both taking around 24 hours to make it out of the queue. In both cases I was able to predict that they would make it pretty easily in about half that time.

----
The ultimate plays for Madden 2003-2006

I like P3 idea (2.00 / 1) (#103)
by tealeaf on Wed Aug 28, 2002 at 03:57:38 PM EST

I think that, in general, giving everyone an option to rate from 1 to 5 (I pretty much never use 0, someone ahead of me usually picks up the spam) gives way too much power to an individual.

I think that, most definitely, we should rate on whether we agree or disagree and not just based on editorial quality.  I see well written boring crap all the time and I don't stop to give a 5.  But if someone points out an interesting idea that I agree and, more rarely, an interesting idea that I disagree with, I will give out a 5, even if there are typos, etc., as long as the essence is there.

Now, +1/-1 is too much like voting and it hides the strength of your agreement or disagreement.

So, I prefer a middle of the road approach:

+1, agree/disagree
+2, agree/disagree
-1, agree/disagree
-2, agree/disagree

Then it becomes clear what the rating means and it enables people to sort the articles more intelligently.  For example, sort high DISAGREE ratings first to get a bunch of posts that even those who disagree found good.  It would be interesting to compare this with sort high AGREE.

Is this complicated?  I don't think so.

What really annoys me is that there are some people who think that 5 should be given to any well written post, regardless of your agreement.  That's bullshit.  Who here reads k5 in an editorial capacity, totally disengaged, and just sifting for technically well written posts?  Is that why you come to k5?  That isn't why I come here.  I come here to discuss things and to read other people's opinion and to state my own.

So, if I disagree strongly, but I am too lazy to write a reply, I'll vote 1.  If I am nice, I will vote 1 and write why I did that, but usually I just get a 1 right back.  This doesn't matter to me, but I can see why people can be discouraged from qualifying their "1" ratings.

Anyway, I think pretty much all rating systems can be abused.  Most important for me is this: if I sort by rating, do I really get high quality posts at the top?  As long as the answer is yes, and as long as not too many good quality posts are buried or unrated, the system works well.

I think k5 system works OK, but it can be better.

another way of thinking about it (3.00 / 1) (#105)
by sarunas on Thu Aug 29, 2002 at 03:13:11 AM EST

I agree with the general ideas you've presented. The different points are good touch up ideas to increase the capacity of the current system. They're at least a good starting point for increasing the complexity and capabilities even if they do end up changing after scrutiny. What 'we do' as a group is act as a big 'idea parsing' engine. We've got input\output functions. Input might be an event or another idea. The internal function between I and O is a combination of the random permutations of culture\individual, influence functions, and a communications reduction (taking complex internal states and reducing them to a shared vocabulary). That's just a complex way of explaining the grapevine effect. What we hope is to control the function of the grapevine. Rather than have the output be a garbled original message, it should be a nice, orderly parsing of the idea, resulting in an intellectual artifact - a point of action, or a mark along the path to a higher goal. In order to break through the current communications bandwidth issue with ideas is to create new tools that we can use. With the new tools we need a new environment with different rules. For the solution to really work well, it has to be intuitive for the user. That's why discussion forums like this work - there's not much to it. But when you make something more complex, there's a lot of people who won't make the jump. The Next Evolution of the community discussion forum needs to have a different set of idea granularities. Instead of having a fixed graph of comments, one needs to be able to both branch out in more dimensions and view the overall structure. An example would be to create a graph of comments instead of just a tree of nestings. A possible change in the granularity would allow a comment to have quotes\references\pointers to other comments\articles\etc and automatically include the proper bibliographical information and citations within the hidden data of the text. This would allow for easier tracking of the logic of a conversation. If someone's early comment is revealed to be innacurate, you can revise the chain of logic, etc. Another evolution would be to open up the points of constraint on various sources of data\interface. We can treat the ongoing discussions as a data stream, or complex object. What needs to be done is adopt a set of base data structures that get populated whenever people add to the graph. These basic structures keep all the book keeping information. It can be kept in a non-descript database as a general repository. Now, we provide access to this data store as a web service, or p2p app, something other than a human\web interface. Now people can write their own front\back ends to the store. They could create an app that requires two monitors, because it displays a large graph of the logic. Someone could make a flow chart of it. Someone could create add their own high level functions and data structures that add logic management functions - examples of which could be buddy\clubs of people who get to comment on their set of the comments-in-the-discussion. Other people could create sets of people who have a high probability of creating meaningful contributions to the discussion and distribute the set in a list that people subscribe to and use to view the commentary. As time passes, you could coalesce objects into volumes and archives. This approach is akin to the idea of freeing up your algorithms. You give some license for free use with small conditions attached, and let people run with it. Conditions might be the retention of the annotations with your data, etc. As a broad name for the idea, you could call the system a net of exegesis.

The future of blog: The scaling barrier | 106 comments (100 topical, 6 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!