I realize that I'm walking a fine line with this question, but seriously... If the readers of/. decide that they want to fill a message thread with conspiracy theories, or diatribe about the poster, what's the problem here? They've read the article (in some cases), they've came to the site, they're interested in it enough to leave feedback, you've got your readers, etc.
It's a reasonable question.
I think the issue is that the goal of Slashdot is to be a content filter -- to take the torrent of news from the Internet and just hand us the things that we're most likely to be interested in. The moderation system is designed to do the same things for comments.
The problem with conspiracy or Slashdot-related stuff is that there is a small group that is vocal that cares deeply about this, and a large group that is probably less vocal that doesn't care about it.
As a result, people who don't care about this have to dig through comments that they are uninterested in, which makes Slashdot less useful as a filter for interesting news.
That doesn't mean that the conspiracy theorists should be *silenced*, but I think it's reasonable to allow people who really don't want to see discussion about Slashdot mechanisms to skip it.
1. Infomercials. How many articles are posted to the main page every day on slashdot? And how many truly unique products are released into the market? One of the good things about product reviews/advertisements posted to slashdot is that people can comment on competing products, other solutions, or the product itself. If there is nothing new or innovative about a hyped product, isn't it good for that to be public knowledge?
I see a lot of articles about ideas in a research lab...but they usually get misrepresented as something that's about to come out. This is *particularly* true of storage technology -- there are a lot of articles about how we could have, say, 10TB on a keychain drive. The article summary sounds as if this is something about to come out, but then, when you read the article, you realize that the product is at best years away. Usually the average quality of comments on such stories is not very good, since there are few people, outside of maybe some storage researchers, who can contribute very useful comments on it.
2. Minor gaming stories on the main page. Yes, sometimes they appear on the main page when likely they shouldn't. But gaming is a nerd mainstay, and shouldn't be ignored. Also, gaming drives a ton of internet revenue. Whether intentional or not, I don't begrudge gaming stories potentially leading to more revenue for slashdot.
I like gaming stories, because there's often a lot of interesting discussion about game design, and sometimes some good inteviews from people in the game development industry. The gaming stories that I don't like are stories about how console X has just gained 3% market share more than console Y -- it's guaranteed to generate nothing but a very predictable flamewar.
Actually, I think that some of it is simply the fact that the April Fools stories for the past couple years, IMHO, haven't been that great. I remember reading some...sometime...that were pretty funny.
I do wish that the April Fools articles were marked as such, at least at the bottom of the summary or something, for those gullible among us...
Make it 1/day, 5/week, whatever you think is fair, but after that threshold, they still get the stories posted, but the links get a "nofollow". That way you're still sharing interesting, unique stories with the Slashdot readers, but people have no incentive to submit hordes of stories.
Doesn't work.
Even if each user only got *one* story accepted per userid, they could still just keep setting up new accounts. A slashdot UID is not an expensive ID -- anyone can create a number of them -- and thus, restrictions on a per-account basis, especially when money is involved, just don't work all that well.
It's easy to get into the swing of criticizing "the anonymous editors" because they're just some inhuman amorphous blobs, and everyone else is doing it, and things aren't perfect. I've certainly ragged on editing to an unfair degree before...but something like this makes you realize that, hey, there's a person back there with feelings too.
I don't want to read a front page story on the NY Times about them changing their paper stock. I don't want to watch a segment on CNN about the CNN Make Up lady. I just want my news!
Perhaps you can make slashdot.slashdot.org and make accounts and AC have it off by default on the main page.
I'd argue that this is a different case from the NYT and CNN. First, your typical user on Slashdot is a lot more capable of making helpful suggestions than your typical CNN reader is for CNN. Second, Slashdot is still a young medium compared to CNN -- forums like this one are still evolving rapidly. Third, a lot of people are interested, from a research standpoint, in how Slashdot works, in the same way that a lot of people are interested in suggesting fixes for email and search engines and IM and so forth. I see a huge number of interesting technical ideas bubbling around here each day, and it would be a shame (IMHO, of course) to exclude Slashdot from those ideas. Fourth, Slashdot is an interactive medium. This isn't just a story about Slashdot -- it's a forum in which ideas and debate can go back and forth. A TV show about CNN can't offer this.
If you consider the ranking of a post as a simple trust metric, you start out with 1 point (or maybe 0 if you've been naughty or 2 if you've been nice). There are some limits on the rate at which people can post (and can post-but-still-be-modded-down), but ultimately, someone can churn out a lot of useless posts fairly easily. Generally, there are many 0-rated posts, and few +5 rated posts. This means that it's easier to pick the wheat from the chaff than the chaff from the wheat.
The drawbacks? Either you cap a rating (in which case anyone that manages to get a false post to +5 gets to keep it at +5) or you don't cap a rating (and any good posts consume the mod points of everyone as they keep rating up the highest articles out there.
Perhaps making 0 the lowest score possible -- i.e. nobody can spend mod points on lowering trolls and flames below ACs, but they can reverse previous moderations upwards.
But, and maybe this is where I should have done more filtering work, the comments and moderation of comments seems to have dropped in quality. The signal to noise ratio is horrible IMNSHO.
I hear people complaining about this on an awful lot of forums. Two ideas come to mind (both of which may apply to some extent):
(a) People are wrong. This is supported by the number of people who consistently claim things like "kids are worse today than they used to be", when it seems pretty unlikely that society has seen a monotonic downfall in the ethics/virtues/whatever of kids. Nostalgia really does seem to make things look golden, and thus make the current state of affairs look worse.
(b) It is an emergent property of open forums to get more noise over time. This is supported by the number of people on *other* open forums, like Usenet, that complain about how noise is always increasing. Perhaps as a forum gets good and attracts interesting people, it gains a reputation, and new people hop in to meet those people and get a soapbox to address all the other new people. The number of hangers-on in the general population exceeds the exceptional, and thus the quality falls.
Oh, and one other possibility:
(c) The merit of a forum is measured as how much a person learns from it, and how challenged and stimulated they are by the debate. If the content of the content remains the same in that forum, eventually old-timers will have learned most of what the forum has to teach them, and they will start to say "there's not much interesting stuff here".
When the submitter posts a review of the book, and the link to the book is one where they make money selling it. Slashdot is a news site, not your personal bookstore.
Agreed, but I would also suggest that there are many ways around it -- at best this would just dampen the problem. The link could be to a page that contains (or *will* contain, after the editor glances at the page) an affiliate link.
Incorrect Information: Often the summary INCORRECTLY summarizes the article, fabricating facts that only leads up to even more people saying "RTFA". What is the point of the summary if not to summarize the article?
I agree. If someone submits an incorrect summary, they shouldn't get the benefit of getting their name on the story submission. I'd rather see Taco's possibly incorrect take on what the story is actually about than some of the clearly incorrect summaries that link to (potentially interesting) stories.
Grammar and Spelling: How hard is it to spell-check the summary? Some of the articles I've seen are just embarassing.
I'm not familiar with the business side of Slashdot, but I'm going to bet that Taco can find some journalism student who wants an internship to clean up submission grammar and spelling ("edited tech news site with hundreds of thousands of readers for a year" probably looks pretty awfully good on a resume, and lots of college students are burning time here anyway...).
Really, Slashdot and its editors should not throw stones while sitting in a greenhouse, so to speak. For a site with, what, how many paid editors, the summaries are of low qulity, there are many dupes, submissions seem to be only rarely spell-checked and sometimes I am sure whatever editor took over a story never RTFA either.
But in doing this article, they're clearly trying to improve things. I'm not going to kick them for asking for ideas and suggestions.
Not doing that sort of basic validity checking is why Slashdot is just the world's most famous blog instead of a trusted news site.
The comments seem to do a pretty good job of finding problems.
Comments that completely reject the original submission usually make their way to the summary in the form of a followup -- this does *usually* happen, though it isn't perfect.
My big beef is that often article summaries -- and *especially* article titles -- are written to be sensationalist or with an incredible degree of bias. For example, maybe Bill Gates donates ten million dollars to AIDS research, and we see an article title that reads "Gates Uses Ill-Gotten Goods to Promote Personal Image", or Sun decides to keep the source to a particular product closed and we get a summary title of "Sun Attacks Open Source".
That gets the attention of people once or twice, but it's really irritating when it keeps happening day in and day out.
I don't have a perfect fix, but I do have some possible ideas -- one is to have editors, if they're unwilling to edit article titles because of concerns of censorship, at least post a followup at the bottom to try to make POVs a bit more neutral.
Another possibility would be to have some mechanism to allow one comment to be voted onto the article summary as a "correction". Not sure how this would work, but it'd be a starting place.
Another way, to try to deal with bias in comments, would be to have a post attribute "Biased". Slashdot is currently driven by opinion, and maybe this would be too strong (I could see it being -0 by default). However, it would be nice to allow someone to look for more neutral posts if that's what they want. I rarely think "Gee, there isn't enough heated argument on Slashdot" and often think "Boy, I wish I could filter out some of this advocacy stuff".
I *really* wish that Slashdot would dispense with the one-post-every-two-minutes rule. I understand that there are concerns about spam, but there's a pretty straightforward fix. Just require an *average* of one-post-every-two-minutes, but allow higher burst rates. Often, I have short responses to comments "No, that's not true, see the following webpage:", and write them more frequently than one every two minutes, and develop a backlog of comments. It would be really nice if the rule was, say, no more than five posts every ten minutes. This *still* keeps spammers down to the same rate, but allows some bursty behavior on the part of posters. An alternative would be to allow a "queue" of some size, say, five posts to be built up...but backlogged posts only appear every two minutes. That way, I don't have to sit around with a bunch of tabs open with finished-but-not-submitted posts open waiting for the current deadline to expire so that I can submit my post.
If this is unacceptable, removal of the two-minute limit for members would be probably the single most likely thing that would convince me to get a Slashdot membership. I already filter the ads, and the main irritation for me is simply the time limits.
I'd like to see the default post ordering be "show newest comments first". This would help ensure that all postings get a fair shot at being viewed, and help eliminate the first post problem.
I'd kind of like to be able to get the friends database in a text file on CD -- nothing in there is private, and it'd be really great to be able to analyze it with something other than screen scraping.
I think that it would be neat if Slashdot added an API to Slashcode to browse and post comments, so that people could make readers (a la Usenet readers). That would be a neat member perk. I can think of a ton of things that could be done from even a simple client, if only there was a good, reliable way of accessing Slashdot. Screen scraping is hard on the servers and not that reliable.
It would be nice if users could attach a couple of small files (even 1K) to their account that are keyed with a simple string. For example,
Maybe popularity shouldn't be the most important criteria about what is published.
The only time this becomes a problem is either (a) if you want to feed users something that they don't want -- probably not a viable strategy or (b) if the the group for whom consumption of the post is being approved differs somehow from the group producing popularity data. If there were 10,000 typicals, I'd trust them pretty well as to whether or not I should read something that they liked.
I think that a lot of Slashdotters gripe about mainstream news sources (CNN/newspapers/talk radio/etc) because most Slashdotters are not particularly good examples of mainstream America. CNN is aiming for Joe Sixpack, and Joe Sixpack can't solder or write graph traversal algorithms.
Slashdot (well, Slashcode) addresses this problem in a couple of simple ways. First, it is a niche (a large niche, but a niche) site. So it only has to serve the group of "nerds". A typical Slashdotter is probably closer to the typical nerd than the typical Joe Sixpack. Second, it allows you to filter by topic -- if you don't care about BSD articles, you tell Slashdot that you don't want to see any more BSD articles. So you can, in a rudimentary way, take advantage of popular selection, but only that of those people that are fairly similar to you.
This is the primary problem that Slashdot's competition, like reddit and digg and a host of other social-network-type sites, are trying to solve with technical fixes (I personally don't think that they're quite there yet, but I'm guessing that the future is in this direction). The approach is usually more-or-less to try to automatically identify people that have similar preferences, and then select articles for each person based on those of people with similar preferences. That is, at least in theory, better than having one person choose what is interesting, and better than just having a simple scalar metric (this article is a 5.0 out of 10 for a typical Joe Sixpack, maybe, but what is it out of 10 for a guy that likes stories about materials science?)
I think that Slashdot could actually benefit a lot from this approach. Slashdot used to try to filter out the flood of available news down to that-which-is-interesting-to-you. Slashdot also has a comments mechanism. This mechanism works if there are 30 comments. It even works well if there are 100 or maybe 200 comments. It doesn't work so well if there are 1000 comments. I just don't have time to read through all the comments, so I only read through the first page's worth. I'm pretty sure that there's a better mechanism for finding and isolating worthwhile comments in the torrent that comes blasting out with each story, because there usually is interesting stuff on Slashdot.
One other point on reputation-based systems. This isn't very concrete, but I'd like to get the meme out there. The largest problem with reputation-based systems is that they are really hard to bootstrap. They have a network effect -- it only helps you to involve yourself if other people are already involved. A number of really obvious ways to bootstrap such a system (such as giving each entity a certain about of trust) doesn't work that well, due to spammers who sign up many accounts. This is a problem for spam and all sorts of things.
Now, a lot of people realized that you *can* bootstrap such a system by giving everyone a certain amount of trust, but making identities expensive -- that is, nobody can write a perl script that keeps churning out new accounts. You can charge money, or make someone waste some human time (like identifying a CAPTCHA) or whatever.
The problem with *this* is that introducing money into the equation has obvious problems ("Buy your votes here!") and that some people aren't willing to waste human time, and others *are* willing to waste human time (sockpuppet GNAA accounts on Slashdot, for example).
Slashdot does have a valuable asset, though. It has a large database of users who have been posting
I'd have had better luck translating the files into 300 baud audio files and transcribing them onto vinyl LPs.
Actually, that's not a bad idea. Consumer audio media lasts a *long* time relative to computer media, and copying audio from media to media is, well, pretty easy -- we've always had good tools for that. You just need to document your encoding format and then start recording.
Here on slashdot, every time some patent is mentioned at all, there's some cranky old technology guy who remembers doing the same thing back in '78 on some project at Fubartronics Inc.
Those guys are great.
IIRC, prior art requires publication, patenting, or public availability of what is done, though. I'm not sure that some hidden internal of a software system, even if ten years earlier, counts as prior art.
Even if that happened, they'd have to grandfather in old patents. The USPTO is *never* going to say "Oh, you know all those millions of dollars you companies blew on non-applied research? Well, we're taking away your patents." Larry Lessig tried pushing for copyrights to be retroactively reverted to Revolutionary War-era lengths, and SCOTUS basically laughed him down.
Companies would scream bloody murder. Hell, IBM would never, ever be involved in pushing for something like that -- a lot of their assets are involved in their patent portfolio.
I'm willing to envision a world in which there are acceptable software patents being produced -- but I think that we probably will need more change than this.
The sense in which a patent is "obvious" is not "obvious" as defined in the dictionary, the common word. "Obvious" patents are those that are trivial modifications of existing ideas. An "obvious" patent might be a patent on a blue taxi, when someone has patented a yellow taxi already.
This becomes significant when the problem is new (maybe due to a changing environment), and nobody has done anything like it before, but that any expert in the field would, in short order, provide a similar answer to the problem.
For example, let's say that suddenly, for some reason, bioengineers produce flying monkeys that can chew through metal but are, as it happens, afraid of triangles. A traditional monkey pit won't work to hold the monkeys, because they'll fly out. You can't put chain link fence over the top, because the monkeys will chew through the fence. However, you can put a big triangle at the top of the cage, and the monkeys will be scared to fly out. That's pretty obvious to you or me -- given the above problem, anyone's going to say "stick some triangles where you don't want the monkeys to go". It doesn't take years of research or funding or anything. However, because it's not a trivial modification of an *existing* idea -- we've never had monkeys that are afraid of, say, squares before -- the system is not obvious in the eyes of the USPTO, and can be patented.
This is the problem that software developers run into. This particular weakness of the patent system -- it does not address changing environments well -- is *exactly* the kind of thing that the software world has been constantly seeing. The whole thing is constantly moving. Faster processors come along that can handle audio in real time, without needing a custom ASIC? Patent systems having general-purpose processors do various audio tasks. The Internet arrives, and cheap data transfer becomes available to everyone? Patent doing various operations over the Internet that are now feasible. Mobile MP3 players get popular? Patent various methods of moving audio to mobile MP3 players.
But, hell, I'll take this as a worthwhile first step.
Agreed. Maybe this new system won't work. Maybe it will. But I can list a heck of a lot of good things about it:
(a) IBM and RH are involved (and maybe some of those ideas churned up on Slashdot, FWIW). Both have incentive to keep patents workable for open source developers.
(b) The USPTO is aware of the problem, has acknowledged that it can improve, and is trying ideas to do what it can. This is a *huge* leap.
(c) It's an idea to try out. It might make things better. If it turns out to be a disaster, it's easy to go back to the way things were. It might make things worse. There are almost certainly going to have to be changes made to the original idea, but I suspect that it's better than what's going on now.
(d) Effort is being made to keep this process inexpensive. That's good for individuals, like open source authors.
(e) It also solves another major problem with research -- trying to evaluate how well a researcher is doing. Right now, it's *damned* difficult for a researcher to prove that they are doing good work. A company hires a researcher, and their only metric is how many papers or how many patents they churn out. The incentive is thus to churn out low quality papers. Now they have a new metric -- if someone churns out a lot of low-grade, heavily-challenged patents, their patent is "less good" than the guy down the hall who came up with an idea that nobody's found any prior art for.
(f) It makes the open source community an *extremely* powerful player in this world. In the past, the ability of the blog/open source world to, once irritated, perform research as to the validity of claims and shoot things down has been absolutely stunning. Look at the forged Bush papers, the SCO case, Microsoft marketing FUD, and so forth. Smart people with spare time on the Internet are can now actually *do* something other than flail about, angry about some nonsense patent.
Am I dubious about some aspects of the patent process (especially software patents?), even with this fix? Sure. I worry that they last too long to be appropriate for the software world, that there is incentive to make patents difficult to read, that the cost of a third party getting a patent rejected exceeds the cost of getting a patent approved, that patents are largely not incredibly amazing ideas, that a company designing a product around a patent is constrained to use the process described in that patent, even if they come up with a better process along the way, that the USPTO doesn't recognize the number of new ideas that *any* software developer produces simply in the course of doing his work, and so forth. However, this isn't going to be worse than things were before, and might be a lot better.
Think about it. Slashdot posts "this patent is stupid" and links to a patent. Patent gets a deluge of prior art and research attached to it. Patent becomes impossible to enforce (anyone who might try to enforce it is going to be challenged). Just the threat of that happening is not minor. Slashdot literally gains the ability to nullify bogus patents. Sure, maybe it can only do so for a few patents a year, but it's the really nasty important ones, the ones that threaten open source projects.
Please, unless you go to an ivy league school, Cal Tech, MIT, etc. your degree doesn't mean jack shit.
Your degree still doesn't need to mean jack shit.
What you learn depends mostly on what you want to learn. You can coast along and get decent grades by just cramming before exams and not reading any of your textbooks or reading until you really, fully understand everything. You aren't gonna learn much, though.
I've read a bunch of Declan McCullagh's articles, mostly from Slashdot links. I have to say that he is one of the very few journalists out there who comes off as both technically knowledgeable and pretty accurate. Most tech writers are, well, pretty awful.
Also, he's notable in that he seems to stay ahead of the curve. Usually, when I read about a tech article in the paper or other media, it's months after I'm aware of the problem and the technical rammifications. In contrast, when I read a DM article, it's consistently the first place I've found out about the issue (to be fair, I often first learn about something through Slashdot, which links me to a DM article, so this may not be entirely fair to the other guys).
Granted, there's serious question as to whether you could ever push something like this through the House or Senate. However, Minnesota has a fix.
Minnesota's state constitution prohibits the passage of laws that cover more than one subject. If a judge determines that a law contains unrelated bundled elements, he can throw them out
In 2004, this hit the news when Judge John Finley threw out, as unconstitutional, a rider regarding gun carry rights. The rider was attached to a natural resources bill.
This is a very good design. Basically, it provides strong incentive for the crafters of a bill to make it as focused as possible. If they don't, anyone out there can challenge it and shoot it down. I've yet to hear of any other remotely workable solution to riders yet.
The problem is that, while the highly technically-literate people on Slashdot don't fear much in the way of online harassment (because they know what avenues they can take, how information can leak, and actually what a harasser can do), most people do not enjoy this situation. It's *alarming* to most people to get unknown IMs or emails -- they don't understand this Internet thing and what risks might be involved, and a law can at least make them feel better.
I kinda think that this is a bad idea too, but it's not grounded in a government conspiracy to take your rights away. I dislike it because it can't work. Trying to filter out "bad" content through legal means is essentially impossible.
Furthermore, I only want to see legal remedies if technical ones are impossible. Legal fixes are expensive, slow-to-react, potentially abusable, often screw the little guy over, and contribute to the masses of unnecessary litigation in the United States. If it's possible to resolve problems withour resorting to litigation, it should be done.
However, online trust is an area that is improving in leaps and bounds and is an incredibly exciting area to research. It's like Google all over again -- there is valuable information with which one can provide an essential service -- all one has to do is write the correct code to harvest and analyze that data. A technical fix to avoid stuff that you don't want to see isn't far-fetched at all.
I'd *much* rather see the federal government just back off and let the techies have motivation to produce technical fixes...
A simple dual-layer 9gb DVD is plenty for hi-res textures. How many cutting edge PC games take more than 9gb? None.
Of course, you might argue that this is only true because, well, a simple dual-layer DVD is only 9gb.
I do kind of agree, though -- aside from making life slightly easier by not making game developers have to compress anything, I can't think of all that much that larger storage formats have done for gaming. The fastest way to consume disk space is with video, and 9GB can store an awful lot of video.
Slashdot Mirror
I realize that I'm walking a fine line with this question, but seriously... If the readers of /. decide that they want to fill a message thread with conspiracy theories, or diatribe about the poster, what's the problem here? They've read the article (in some cases), they've came to the site, they're interested in it enough to leave feedback, you've got your readers, etc.
It's a reasonable question.
I think the issue is that the goal of Slashdot is to be a content filter -- to take the torrent of news from the Internet and just hand us the things that we're most likely to be interested in. The moderation system is designed to do the same things for comments.
The problem with conspiracy or Slashdot-related stuff is that there is a small group that is vocal that cares deeply about this, and a large group that is probably less vocal that doesn't care about it.
As a result, people who don't care about this have to dig through comments that they are uninterested in, which makes Slashdot less useful as a filter for interesting news.
That doesn't mean that the conspiracy theorists should be *silenced*, but I think it's reasonable to allow people who really don't want to see discussion about Slashdot mechanisms to skip it.
1. Infomercials. How many articles are posted to the main page every day on slashdot? And how many truly unique products are released into the market? One of the good things about product reviews/advertisements posted to slashdot is that people can comment on competing products, other solutions, or the product itself. If there is nothing new or innovative about a hyped product, isn't it good for that to be public knowledge?
I see a lot of articles about ideas in a research lab...but they usually get misrepresented as something that's about to come out. This is *particularly* true of storage technology -- there are a lot of articles about how we could have, say, 10TB on a keychain drive. The article summary sounds as if this is something about to come out, but then, when you read the article, you realize that the product is at best years away. Usually the average quality of comments on such stories is not very good, since there are few people, outside of maybe some storage researchers, who can contribute very useful comments on it.
2. Minor gaming stories on the main page. Yes, sometimes they appear on the main page when likely they shouldn't. But gaming is a nerd mainstay, and shouldn't be ignored. Also, gaming drives a ton of internet revenue. Whether intentional or not, I don't begrudge gaming stories potentially leading to more revenue for slashdot.
I like gaming stories, because there's often a lot of interesting discussion about game design, and sometimes some good inteviews from people in the game development industry. The gaming stories that I don't like are stories about how console X has just gained 3% market share more than console Y -- it's guaranteed to generate nothing but a very predictable flamewar.
Actually, I think that some of it is simply the fact that the April Fools stories for the past couple years, IMHO, haven't been that great. I remember reading some...sometime...that were pretty funny.
I do wish that the April Fools articles were marked as such, at least at the bottom of the summary or something, for those gullible among us...
Make it 1/day, 5/week, whatever you think is fair, but after that threshold, they still get the stories posted, but the links get a "nofollow". That way you're still sharing interesting, unique stories with the Slashdot readers, but people have no incentive to submit hordes of stories.
Doesn't work.
Even if each user only got *one* story accepted per userid, they could still just keep setting up new accounts. A slashdot UID is not an expensive ID -- anyone can create a number of them -- and thus, restrictions on a per-account basis, especially when money is involved, just don't work all that well.
The games stories that involve minimal technical discussion of computer issues could be given their own web site: Gamesdot.
What's wrong with games.slashdot.org?
I'm kinda glad you posted this, actually.
It's easy to get into the swing of criticizing "the anonymous editors" because they're just some inhuman amorphous blobs, and everyone else is doing it, and things aren't perfect. I've certainly ragged on editing to an unfair degree before...but something like this makes you realize that, hey, there's a person back there with feelings too.
I don't want to read a front page story on the NY Times about them changing their paper stock. I don't want to watch a segment on CNN about the CNN Make Up lady. I just want my news!
Perhaps you can make slashdot.slashdot.org and make accounts and AC have it off by default on the main page.
I'd argue that this is a different case from the NYT and CNN. First, your typical user on Slashdot is a lot more capable of making helpful suggestions than your typical CNN reader is for CNN. Second, Slashdot is still a young medium compared to CNN -- forums like this one are still evolving rapidly. Third, a lot of people are interested, from a research standpoint, in how Slashdot works, in the same way that a lot of people are interested in suggesting fixes for email and search engines and IM and so forth. I see a huge number of interesting technical ideas bubbling around here each day, and it would be a shame (IMHO, of course) to exclude Slashdot from those ideas. Fourth, Slashdot is an interactive medium. This isn't just a story about Slashdot -- it's a forum in which ideas and debate can go back and forth. A TV show about CNN can't offer this.
This makes sense, based on how trust works.
If you consider the ranking of a post as a simple trust metric, you start out with 1 point (or maybe 0 if you've been naughty or 2 if you've been nice). There are some limits on the rate at which people can post (and can post-but-still-be-modded-down), but ultimately, someone can churn out a lot of useless posts fairly easily. Generally, there are many 0-rated posts, and few +5 rated posts. This means that it's easier to pick the wheat from the chaff than the chaff from the wheat.
The drawbacks? Either you cap a rating (in which case anyone that manages to get a false post to +5 gets to keep it at +5) or you don't cap a rating (and any good posts consume the mod points of everyone as they keep rating up the highest articles out there.
Perhaps making 0 the lowest score possible -- i.e. nobody can spend mod points on lowering trolls and flames below ACs, but they can reverse previous moderations upwards.
But, and maybe this is where I should have done more filtering work, the comments and moderation of comments seems to have dropped in quality. The signal to noise ratio is horrible IMNSHO.
I hear people complaining about this on an awful lot of forums. Two ideas come to mind (both of which may apply to some extent):
(a) People are wrong. This is supported by the number of people who consistently claim things like "kids are worse today than they used to be", when it seems pretty unlikely that society has seen a monotonic downfall in the ethics/virtues/whatever of kids. Nostalgia really does seem to make things look golden, and thus make the current state of affairs look worse.
(b) It is an emergent property of open forums to get more noise over time. This is supported by the number of people on *other* open forums, like Usenet, that complain about how noise is always increasing. Perhaps as a forum gets good and attracts interesting people, it gains a reputation, and new people hop in to meet those people and get a soapbox to address all the other new people. The number of hangers-on in the general population exceeds the exceptional, and thus the quality falls.
Oh, and one other possibility:
(c) The merit of a forum is measured as how much a person learns from it, and how challenged and stimulated they are by the debate. If the content of the content remains the same in that forum, eventually old-timers will have learned most of what the forum has to teach them, and they will start to say "there's not much interesting stuff here".
When the submitter posts a review of the book, and the link to the book is one where they make money selling it. Slashdot is a news site, not your personal bookstore.
Agreed, but I would also suggest that there are many ways around it -- at best this would just dampen the problem. The link could be to a page that contains (or *will* contain, after the editor glances at the page) an affiliate link.
Incorrect Information: Often the summary INCORRECTLY summarizes the article, fabricating facts that only leads up to even more people saying "RTFA". What is the point of the summary if not to summarize the article?
I agree. If someone submits an incorrect summary, they shouldn't get the benefit of getting their name on the story submission. I'd rather see Taco's possibly incorrect take on what the story is actually about than some of the clearly incorrect summaries that link to (potentially interesting) stories.
Grammar and Spelling: How hard is it to spell-check the summary? Some of the articles I've seen are just embarassing.
I'm not familiar with the business side of Slashdot, but I'm going to bet that Taco can find some journalism student who wants an internship to clean up submission grammar and spelling ("edited tech news site with hundreds of thousands of readers for a year" probably looks pretty awfully good on a resume, and lots of college students are burning time here anyway...).
Really, Slashdot and its editors should not throw stones while sitting in a greenhouse, so to speak. For a site with, what, how many paid editors, the summaries are of low qulity, there are many dupes, submissions seem to be only rarely spell-checked and sometimes I am sure whatever editor took over a story never RTFA either.
But in doing this article, they're clearly trying to improve things. I'm not going to kick them for asking for ideas and suggestions.
Not doing that sort of basic validity checking is why Slashdot is just the world's most famous blog instead of a trusted news site.
The comments seem to do a pretty good job of finding problems.
Comments that completely reject the original submission usually make their way to the summary in the form of a followup -- this does *usually* happen, though it isn't perfect.
My big beef is that often article summaries -- and *especially* article titles -- are written to be sensationalist or with an incredible degree of bias. For example, maybe Bill Gates donates ten million dollars to AIDS research, and we see an article title that reads "Gates Uses Ill-Gotten Goods to Promote Personal Image", or Sun decides to keep the source to a particular product closed and we get a summary title of "Sun Attacks Open Source".
That gets the attention of people once or twice, but it's really irritating when it keeps happening day in and day out.
I don't have a perfect fix, but I do have some possible ideas -- one is to have editors, if they're unwilling to edit article titles because of concerns of censorship, at least post a followup at the bottom to try to make POVs a bit more neutral.
Another possibility would be to have some mechanism to allow one comment to be voted onto the article summary as a "correction". Not sure how this would work, but it'd be a starting place.
Another way, to try to deal with bias in comments, would be to have a post attribute "Biased". Slashdot is currently driven by opinion, and maybe this would be too strong (I could see it being -0 by default). However, it would be nice to allow someone to look for more neutral posts if that's what they want. I rarely think "Gee, there isn't enough heated argument on Slashdot" and often think "Boy, I wish I could filter out some of this advocacy stuff".
I *really* wish that Slashdot would dispense with the one-post-every-two-minutes rule. I understand that there are concerns about spam, but there's a pretty straightforward fix. Just require an *average* of one-post-every-two-minutes, but allow higher burst rates. Often, I have short responses to comments "No, that's not true, see the following webpage:", and write them more frequently than one every two minutes, and develop a backlog of comments. It would be really nice if the rule was, say, no more than five posts every ten minutes. This *still* keeps spammers down to the same rate, but allows some bursty behavior on the part of posters. An alternative would be to allow a "queue" of some size, say, five posts to be built up...but backlogged posts only appear every two minutes. That way, I don't have to sit around with a bunch of tabs open with finished-but-not-submitted posts open waiting for the current deadline to expire so that I can submit my post.
If this is unacceptable, removal of the two-minute limit for members would be probably the single most likely thing that would convince me to get a Slashdot membership. I already filter the ads, and the main irritation for me is simply the time limits.
I'd like to see the default post ordering be "show newest comments first". This would help ensure that all postings get a fair shot at being viewed, and help eliminate the first post problem.
I'd kind of like to be able to get the friends database in a text file on CD -- nothing in there is private, and it'd be really great to be able to analyze it with something other than screen scraping.
I think that it would be neat if Slashdot added an API to Slashcode to browse and post comments, so that people could make readers (a la Usenet readers). That would be a neat member perk. I can think of a ton of things that could be done from even a simple client, if only there was a good, reliable way of accessing Slashdot. Screen scraping is hard on the servers and not that reliable.
It would be nice if users could attach a couple of small files (even 1K) to their account that are keyed with a simple string. For example,
Maybe popularity shouldn't be the most important criteria about what is published.
The only time this becomes a problem is either (a) if you want to feed users something that they don't want -- probably not a viable strategy or (b) if the the group for whom consumption of the post is being approved differs somehow from the group producing popularity data. If there were 10,000 typicals, I'd trust them pretty well as to whether or not I should read something that they liked.
I think that a lot of Slashdotters gripe about mainstream news sources (CNN/newspapers/talk radio/etc) because most Slashdotters are not particularly good examples of mainstream America. CNN is aiming for Joe Sixpack, and Joe Sixpack can't solder or write graph traversal algorithms.
Slashdot (well, Slashcode) addresses this problem in a couple of simple ways. First, it is a niche (a large niche, but a niche) site. So it only has to serve the group of "nerds". A typical Slashdotter is probably closer to the typical nerd than the typical Joe Sixpack. Second, it allows you to filter by topic -- if you don't care about BSD articles, you tell Slashdot that you don't want to see any more BSD articles. So you can, in a rudimentary way, take advantage of popular selection, but only that of those people that are fairly similar to you.
This is the primary problem that Slashdot's competition, like reddit and digg and a host of other social-network-type sites, are trying to solve with technical fixes (I personally don't think that they're quite there yet, but I'm guessing that the future is in this direction). The approach is usually more-or-less to try to automatically identify people that have similar preferences, and then select articles for each person based on those of people with similar preferences. That is, at least in theory, better than having one person choose what is interesting, and better than just having a simple scalar metric (this article is a 5.0 out of 10 for a typical Joe Sixpack, maybe, but what is it out of 10 for a guy that likes stories about materials science?)
I think that Slashdot could actually benefit a lot from this approach. Slashdot used to try to filter out the flood of available news down to that-which-is-interesting-to-you. Slashdot also has a comments mechanism. This mechanism works if there are 30 comments. It even works well if there are 100 or maybe 200 comments. It doesn't work so well if there are 1000 comments. I just don't have time to read through all the comments, so I only read through the first page's worth. I'm pretty sure that there's a better mechanism for finding and isolating worthwhile comments in the torrent that comes blasting out with each story, because there usually is interesting stuff on Slashdot.
One other point on reputation-based systems. This isn't very concrete, but I'd like to get the meme out there. The largest problem with reputation-based systems is that they are really hard to bootstrap. They have a network effect -- it only helps you to involve yourself if other people are already involved. A number of really obvious ways to bootstrap such a system (such as giving each entity a certain about of trust) doesn't work that well, due to spammers who sign up many accounts. This is a problem for spam and all sorts of things.
Now, a lot of people realized that you *can* bootstrap such a system by giving everyone a certain amount of trust, but making identities expensive -- that is, nobody can write a perl script that keeps churning out new accounts. You can charge money, or make someone waste some human time (like identifying a CAPTCHA) or whatever.
The problem with *this* is that introducing money into the equation has obvious problems ("Buy your votes here!") and that some people aren't willing to waste human time, and others *are* willing to waste human time (sockpuppet GNAA accounts on Slashdot, for example).
Slashdot does have a valuable asset, though. It has a large database of users who have been posting
I'd have had better luck translating the files into 300 baud audio files and transcribing them onto vinyl LPs.
Actually, that's not a bad idea. Consumer audio media lasts a *long* time relative to computer media, and copying audio from media to media is, well, pretty easy -- we've always had good tools for that. You just need to document your encoding format and then start recording.
Here on slashdot, every time some patent is mentioned at all, there's some cranky old technology guy who remembers doing the same thing back in '78 on some project at Fubartronics Inc.
Those guys are great.
IIRC, prior art requires publication, patenting, or public availability of what is done, though. I'm not sure that some hidden internal of a software system, even if ten years earlier, counts as prior art.
Even if that happened, they'd have to grandfather in old patents. The USPTO is *never* going to say "Oh, you know all those millions of dollars you companies blew on non-applied research? Well, we're taking away your patents." Larry Lessig tried pushing for copyrights to be retroactively reverted to Revolutionary War-era lengths, and SCOTUS basically laughed him down.
Companies would scream bloody murder. Hell, IBM would never, ever be involved in pushing for something like that -- a lot of their assets are involved in their patent portfolio.
I'm willing to envision a world in which there are acceptable software patents being produced -- but I think that we probably will need more change than this.
The sense in which a patent is "obvious" is not "obvious" as defined in the dictionary, the common word. "Obvious" patents are those that are trivial modifications of existing ideas. An "obvious" patent might be a patent on a blue taxi, when someone has patented a yellow taxi already.
This becomes significant when the problem is new (maybe due to a changing environment), and nobody has done anything like it before, but that any expert in the field would, in short order, provide a similar answer to the problem.
For example, let's say that suddenly, for some reason, bioengineers produce flying monkeys that can chew through metal but are, as it happens, afraid of triangles. A traditional monkey pit won't work to hold the monkeys, because they'll fly out. You can't put chain link fence over the top, because the monkeys will chew through the fence. However, you can put a big triangle at the top of the cage, and the monkeys will be scared to fly out. That's pretty obvious to you or me -- given the above problem, anyone's going to say "stick some triangles where you don't want the monkeys to go". It doesn't take years of research or funding or anything. However, because it's not a trivial modification of an *existing* idea -- we've never had monkeys that are afraid of, say, squares before -- the system is not obvious in the eyes of the USPTO, and can be patented.
This is the problem that software developers run into. This particular weakness of the patent system -- it does not address changing environments well -- is *exactly* the kind of thing that the software world has been constantly seeing. The whole thing is constantly moving. Faster processors come along that can handle audio in real time, without needing a custom ASIC? Patent systems having general-purpose processors do various audio tasks. The Internet arrives, and cheap data transfer becomes available to everyone? Patent doing various operations over the Internet that are now feasible. Mobile MP3 players get popular? Patent various methods of moving audio to mobile MP3 players.
But, hell, I'll take this as a worthwhile first step.
Agreed. Maybe this new system won't work. Maybe it will. But I can list a heck of a lot of good things about it:
(a) IBM and RH are involved (and maybe some of those ideas churned up on Slashdot, FWIW). Both have incentive to keep patents workable for open source developers.
(b) The USPTO is aware of the problem, has acknowledged that it can improve, and is trying ideas to do what it can. This is a *huge* leap.
(c) It's an idea to try out. It might make things better. If it turns out to be a disaster, it's easy to go back to the way things were. It might make things worse. There are almost certainly going to have to be changes made to the original idea, but I suspect that it's better than what's going on now.
(d) Effort is being made to keep this process inexpensive. That's good for individuals, like open source authors.
(e) It also solves another major problem with research -- trying to evaluate how well a researcher is doing. Right now, it's *damned* difficult for a researcher to prove that they are doing good work. A company hires a researcher, and their only metric is how many papers or how many patents they churn out. The incentive is thus to churn out low quality papers. Now they have a new metric -- if someone churns out a lot of low-grade, heavily-challenged patents, their patent is "less good" than the guy down the hall who came up with an idea that nobody's found any prior art for.
(f) It makes the open source community an *extremely* powerful player in this world. In the past, the ability of the blog/open source world to, once irritated, perform research as to the validity of claims and shoot things down has been absolutely stunning. Look at the forged Bush papers, the SCO case, Microsoft marketing FUD, and so forth. Smart people with spare time on the Internet are can now actually *do* something other than flail about, angry about some nonsense patent.
Am I dubious about some aspects of the patent process (especially software patents?), even with this fix? Sure. I worry that they last too long to be appropriate for the software world, that there is incentive to make patents difficult to read, that the cost of a third party getting a patent rejected exceeds the cost of getting a patent approved, that patents are largely not incredibly amazing ideas, that a company designing a product around a patent is constrained to use the process described in that patent, even if they come up with a better process along the way, that the USPTO doesn't recognize the number of new ideas that *any* software developer produces simply in the course of doing his work, and so forth. However, this isn't going to be worse than things were before, and might be a lot better.
Think about it. Slashdot posts "this patent is stupid" and links to a patent. Patent gets a deluge of prior art and research attached to it. Patent becomes impossible to enforce (anyone who might try to enforce it is going to be challenged). Just the threat of that happening is not minor. Slashdot literally gains the ability to nullify bogus patents. Sure, maybe it can only do so for a few patents a year, but it's the really nasty important ones, the ones that threaten open source projects.
Thanks to all those folks involved!
Please, unless you go to an ivy league school, Cal Tech, MIT, etc. your degree doesn't mean jack shit.
Your degree still doesn't need to mean jack shit.
What you learn depends mostly on what you want to learn. You can coast along and get decent grades by just cramming before exams and not reading any of your textbooks or reading until you really, fully understand everything. You aren't gonna learn much, though.
I've read a bunch of Declan McCullagh's articles, mostly from Slashdot links. I have to say that he is one of the very few journalists out there who comes off as both technically knowledgeable and pretty accurate. Most tech writers are, well, pretty awful.
Also, he's notable in that he seems to stay ahead of the curve. Usually, when I read about a tech article in the paper or other media, it's months after I'm aware of the problem and the technical rammifications. In contrast, when I read a DM article, it's consistently the first place I've found out about the issue (to be fair, I often first learn about something through Slashdot, which links me to a DM article, so this may not be entirely fair to the other guys).
Granted, there's serious question as to whether you could ever push something like this through the House or Senate. However, Minnesota has a fix.
Minnesota's state constitution prohibits the passage of laws that cover more than one subject. If a judge determines that a law contains unrelated bundled elements, he can throw them out
In 2004, this hit the news when Judge John Finley threw out, as unconstitutional, a rider regarding gun carry rights. The rider was attached to a natural resources bill.
This is a very good design. Basically, it provides strong incentive for the crafters of a bill to make it as focused as possible. If they don't, anyone out there can challenge it and shoot it down. I've yet to hear of any other remotely workable solution to riders yet.
The problem is that, while the highly technically-literate people on Slashdot don't fear much in the way of online harassment (because they know what avenues they can take, how information can leak, and actually what a harasser can do), most people do not enjoy this situation. It's *alarming* to most people to get unknown IMs or emails -- they don't understand this Internet thing and what risks might be involved, and a law can at least make them feel better.
I kinda think that this is a bad idea too, but it's not grounded in a government conspiracy to take your rights away. I dislike it because it can't work. Trying to filter out "bad" content through legal means is essentially impossible.
Furthermore, I only want to see legal remedies if technical ones are impossible. Legal fixes are expensive, slow-to-react, potentially abusable, often screw the little guy over, and contribute to the masses of unnecessary litigation in the United States. If it's possible to resolve problems withour resorting to litigation, it should be done.
However, online trust is an area that is improving in leaps and bounds and is an incredibly exciting area to research. It's like Google all over again -- there is valuable information with which one can provide an essential service -- all one has to do is write the correct code to harvest and analyze that data. A technical fix to avoid stuff that you don't want to see isn't far-fetched at all.
I'd *much* rather see the federal government just back off and let the techies have motivation to produce technical fixes...
The Apache web server (STILL the most popular web server in the world today and consistantly more secure than Microsoft IIS)
"Still"? It's been steadily increasing its lead over IIS for years.
A simple dual-layer 9gb DVD is plenty for hi-res textures. How many cutting edge PC games take more than 9gb? None.
Of course, you might argue that this is only true because, well, a simple dual-layer DVD is only 9gb.
I do kind of agree, though -- aside from making life slightly easier by not making game developers have to compress anything, I can't think of all that much that larger storage formats have done for gaming. The fastest way to consume disk space is with video, and 9GB can store an awful lot of video.
From the standpoint of a company that wants to promote their product and company name? The Windows one. ;-)