I was thinking that an even better activity would be a "[inter]National Crypto Day", when we all send each other lots of encrypted messages. In fact, I just sent the following message to a friend of mine (it's a private message, but I don't see any harm in posting it here):
This would seem to be even more effective than "M-x spook" - style keyword spoofing, since we'd really be showing that "they" can't read our mail if we don't want them to, and exercising our right to send mail that they can't read.
Additionally, I would suggest sending some messages that are just randomly-generated gibberish. This would be a response to any proposal to ban crypto: it would make the point that you can't prove that a given message consisting of seemingly-random data is actually a ciphertext at all, and not really just random. Banning the transmission of meaningless random numbers would be absurd, which means that banning the transmission of ciphertext would be impossible. Suppose I had sent the following to my friend instead:
I would be seriously impressed if anyone (aside from myself or my friend) could prove that the first is a secret message, let alone tell me what it says. However, neither can I prove that the second is not a secret message (in fact, you don't have to believe me when I claim that), so if they were to reverse the burden of proof, I would have a problem -- they could punish me for refusing to surrender a non-existent key. See this story from a couple weeks ago.
Of course, all this assumes that there really is a "they" trying to read everyone's mail -- I'm stating no opinion on that.
I left out a few words: "being twice the number of letters in the alphabet" should have been "being two more than twice the number of letters in the alphabet", since Solitaire uses the standard 52 cards of the deck, with two mapping to each letter of the alphabet, plus the two jokers as fenceposts, or placeholders, to divide the deck into two groups, for a total of 54.
The algorithm consists of several fancy cuts, moving the jokers around, etc., and ultimately selects a single card. The process also puts the deck into a different permutation (and there are about 2.3 * 10^71 permutations), so repeated applications will produce a pseudo-random sequence of cards. The numerical values of the cards are used as a one-time pad to encrypt the message. This solves the problem of key distribution for one-time pads, because two parties can start with identically-shuffled decks and generate the same keystream. I don't know the math of how it works, but supposedly the pseudo-random stream is good enough for the system to be pretty secure.
The number 54, being twice the number of letters in the alphabet, as well as the size of a deck of playing cards, including jokers, sort of brings to mind Bruce Schneier's "Solitaire" cryptosystem, or at least it does when you mention Cryptonomicon at the same time.
If you haven't read it: Schneier devised a cryptosystem called "Solitaire" that is you can use to encrypt and decrypt messages by hand with a pack of cards. In Cryptonomicon, two characters use this method to communicate, and Schneier's paper describing the system is included as an appendix.
Though low-tech, Solitaire is supposedly secure against modern attacks, so if this contest uses it, it would be pretty hard. Then again, the number 54 and the relation to the book are the only evidence for it, so it's probably unrelated.
They always release a new Milestone the day after I download the previous one. Of course, in this case, I'll admit it was kind of dumb of me to download M9 yesterday when the Milestone page clearly stated the target date for M10 was 10/08/99. That whole discussion just got me itching to download something, so I did.
I'm about two hours away from being able to try M10, but I have some comments on the Mac version of M9: the browser functionality is coming along nicely -- for example, it loads Slashdot perfectly, though it still can't log in. The front-end is starting to look relatively polished, but my concern is that it doesn't look like a Mac application. Actually, with previous builds, I assumed that it was simply because the GUI-polishing was being left for last. Now, though, it seems to be starting to look the way it's meant to look, if that makes any sense. I'm seeing the interface that they're designing, and it doesn't seem Mac-like.
As I understand it, they use some sort of cross-platform front-end framework to make all versions use a single codebase. This doesn't seem to use native widgets, so things look funny. This could be okay, if the widgets were similar enough, and if they worked right (even so, the non-native look-and-feel would be a turn-off for many Mac-users), but in fact, many of the text fields, radio buttons, etc., don't draw quite right. Text fields are too tall and narrow for their text, radio buttons are too small, and a few of their pixels get chopped off, the sidebar in the Preferences window sort of jumps when I click something, etc.
This sort of thing is a problem with cross-platform GUIs, I guess, because corresponding things have different relative sizes, and some amount of platform-specific attention is needed to make everything fit. I don't know how this compares to other platforms; maybe the Mac version is just not getting enough attention.
By the way, what's with the second date for M12 on the Milestone Page? "M12 - 12/7/99 - 15/99/99"? Is somebody having Y2K trouble?
"An Architectural Review Board should be formed to maintain the source trees. "
This is a really dumb idea - a govt. committee that decides software architechture...
I didn't say it would be a government committee. I don't think it should be, because of your points, plus I think that really would be too much government interference. I want the government to make this happen, but not be (unnecessarily) involved in the implementation.
It would be an industry committee, like the way OpenGL is managed, or the Object Management Group (OMG) for CORBA. An Open Source project like this would need an organization to serve as a code maintainer, and I just thought some rules for its formation, such as the restrictions on Microsoft and the opportunity for anyone who produces an operating system or a web browser to be represented, should be part of the process. Someone has to approve patches, and the proposal wouldn't be complete without some suggestion as to how it would be managed. Membership in the group would be pretty much open to anyone, though there would have to be some mechanism to make sure it didn't get stuffed with people with any particular agenda; I just listed some of those who would probably be most interested in being members.
I guess the reason I scare-quoted "official" is that there wouldn't be any government mandate making this group's distribution any more valid than anyone else's, assuming someone else decided to fork off a version and set up their own source tree. This would just assure that there would be a version to step into the gap left by Microsoft's closed one. Primarily, they would be there to keep an eye on what Microsoft did with the code. Realistically, Microsoft would, for a while at least, remain the major developer and distributor, and this would just make sure it stayed open.
Anyway, I guess the ARB is not the most important part of the proposal. I just thought it would be a good idea to specify some of how the code would be maintained, and to keep Microsoft from retaining de facto control over it, but the really important part is that the code be donated to the public under the GPL. I don't see how anything less could undo the damage Microsoft has, in my opinion, done.
The source code for the offending programs, namely Windows 95/98, Internet Explorer, and whatever other products were used for anticompetitive purposes, should be seized and given to the public under the GPL to compensate for the damage they have done.
An Architectural Review Board should be formed to maintain the source trees. The ARB should represent high-tech companies such as IBM, Sun, Apple, SGI, HP, Compaq, Red Hat, etc. Microsoft can be a member, but not with any controlling authority. The ARB should also include representatives of the OSS/Free Software community. These would be a bit harder to choose, but clearly the FSF should be represented, as well as the Mozilla team, etc.
Changes to the main source tree would have to be approved by the ARB. The top-of-trunk code, as well as any derivatives, would be publicly available to anyone wanting to look at it or use it, under the terms of the GPL. That means that anyone wanting to fork off their own version and distribute it could do so, but not without keeping their own source open, etc. Microsoft would be allowed to "innovate" by making contributions (sure, why not?) -- even their application-development divisions would be able to work in tandem with the OS development (again, why not?), but there would be no secret APIs, because all such contributions would have to be open. As an extra provision, Microsoft (at least for some period of time) would not be allowed to distribute any version not approved by the ARB, or any binary-only form. That is, their contributions would have to be adopted into the "official" top-of-trunk version before they could be distributed. This would prevent their use of "embrace-and-extend(-and-extinguish)" tactics.
This might seem excessive, or "cruel and unusual", but it's really perfectly reasonable. If software can be "intellectual property" at all, then it can be confiscated, just like any other kind of property, if it is used to commit a crime. Since this property was the instrument used to commit the crimes, confiscating it would be no different from confiscating a drug dealer's car. Placing it under the GPL would then amount to donating it to the injured parties.
This is approximately the same comment I made back in February, in response to a proposed "source-licensing punishment". The archived page does not have comment links, but mine is about three-fifths of the way down, or just search for my name.
Unless I'm mistaken (it could happen), a quantum computer could break a key of any length in constant time, up to the number of qubits that it is able to maintain. That last part is the important disclaimer, since making qubits behave correctly (i.e., making them collapse correctly and not prematurely) is really hard, and gets harder the more you try to use.
Last I heard (and I don't claim to be at all up-to-date), getting a quantum computer to maintain 4 qubits, so as to factor 15 into 3 and 5, would be considered quite a feat. A 512-qubit computer would work on the same principles, and would factor 512-bit numbers just as quickly as it would factor 15, but I'm as skeptical as everyone else as to whether they've managed to construct such a thing.
Even if they have, it doesn't follow that larger keys are vulnerable: even if they managed the amazing feat of building a 512-qubit quantum computer, it still wouldn't be able to factor numbers larger than 512 bits. Breaking a four-kilobit keys would require yet another three and a half kilo-qubits, which they don't necessarily have.
On the other hand, making the leap from a couple of qubits to 512 would suggest that they have developed a technique for adding qubits fairly easily, perhaps with polynomial or even linear cost, in which case they would be able to make one with four kilo-qubits, a mega-qubit, or whatever, without exponentially greater effort.
I thought I read that it was something like this: First of all, Oracle did it in ~71 seconds, meaning that SQL Server would have to come in just under 2 hours to "win".
However, it also turned out that due to some loophole in the definition of the TPC-D benchmark, it was arguably legal to set up the system to effectively precompute the query while loading the dataset, which is not counted in the time, so that all you have to do on the clock is spit out the results. However many hours the system takes to load the dataset does not count.
Using this technique, Oracle did it in about half a second, and SQL Server did it in less than two -- still slower, but not 100x. If you consider this technique valid, then I guess MS won, but it's pretty clear, to me at least, that in any reasonable interpretation, this would constitute cheating.
I never heard of any times for a SQL Server system doing it the honest way, or even whether or not anyone got it to happen at all.
As I said above, I'm not too impressed with the "Clotho" idea -- it sounds too much like a "portal", and I ranted a bit about those, too. I'm not very worried about any "cyberclysm", either.
What I think Katz misses is that we won't just use machines, as we do now, to exchange and process information, but also to understand it. He mentions Ray Kurzweil, but if he'd read and understood The Age of Spiritual Machines, he would rememebr that Kurzweil expects us to merge with the machines, not just use them as assistants, which is what I've also been imagining for a while now.
Consider a scientific application in which the computer collects and processes a huge amount of data -- far more than the scientist could ever keep in his head. The computer doesn't just spit out the answer in the form of a number, let alone a new theorem or a complete paper. Instead, it generates a visualization that allows the scientist to understand the data set intuitively despite being unable to see all the details. The computer then also spits out some numbers that are the result of a full analysis, and between these, the scientist now understands a greater amount of information than he could have without the machine.
This is not even a very good metaphor for the real idea, which is that computers will become part of our consciousness. A direct neural interface doesn't just mean that you'll see a window manager in your head and move the mouse by thought -- you'll actually be able to be aware of all the information as it is being processed. Imagine your mind working thousands of times faster: then time would subjectively pass more slowly, and you'd be able to ingest far more information in the same amount of time. Now add in improved recall, visualization, and mathematical ability.
Not only will we have access to more information, we'll be able to experience more of it, more fully. I wish I could read every single comment on Slashdot, and post ten times more than I do, but even if I spent 168 hours a week at it, there's just no way, because my mind is too slow. Katz' idea sounds basically like perfecting the moderation system and adding good search capabilities. That would be nice, but it's nowhere near the full experience that I envision.
I can't stress enough that I'm not just talking about having all that information at our virtual fingertips, but really knowing it as intimately as we do our names and the first five natural numbers. There need be no "cyberclysm" if our intelligence increases as quickly as our needs. There may be some hiccups and bottlenecks if the paces don't match perfectly, but there's no inevitable explosion looming.
I don't really think of search engines this way, since they are about searching for pages with the particular information that I am looking for. The only sense in which they "tell" me what to read is that they produce a list of things that match the description I give, which is really not the same thing. Of course, search engines are now acting increasingly like "portals", bombarding us with book, movie, and music recommendations, etc., along with our query results.
The very concept of "portals", however, really bugs me, since they are all about telling you where to go. They present a list of things that they think you'll find interesting, and you're supposed to just follow them. It's fast, easy, and convenient! It saves you the trouble of searching through the whole web by taking you straight to the good stuff! You don't have to understand those big, complicated URLs; just click here! Arghh! I for one see it as an attempt to undermine the user-centric, interactive nature of the web, which is what makes it so different from other media in the first place. This is what people like Katz have gotten so excited about -- the democratization of information flow. But with a "portal", it turns back into a provider/consumer relationship. Of course, the user can still navigate freely, but portals are all about discouraging this. "Ideally" (as far as they're concerned), the user would confine his entire web experience to the links presented by the portal. Why do you think we keep hearing about "eyeballs"? It's because what matters to them is the ads, which are based on the model of them telling the consumer what to see, what to think, and what to buy. Basically, they're trying to turn the web into television.
This is really what "Clotho" sounds like. I know the idea is for it to rise above all the commercialism, but it just does not seem possible for something like this to remain "pure". It seems inevitable, first, that there would be a proliferation of sites claiming to offer this service, but which would really be just a bunch of search engines and "portals", and also that they would end up being big, overhyped (and ridiculously overvalued) commercial entities that would try to control what content, and what ads, reach all those "eyeballs".
I'd like to see a list of links to the mp3 files for all the episodes right here on Slashdot's "index.pl?section=radio" page. I don't really care whether the files are mirrored here, elsewhere, or linked directly to The Sync. I'd just like to be able to get the latest episode, and any that I've missed, right here, instead of having to go to http://thesync.com/geeks for the latest and http://thesync.com/geeks/archives.html for the older ones.
As for putting it on the front page vs. the "radio" section, I think the best thing is to keep it here in the side section, but mention it (at least in a "quickie") on the front page every couple months, or every three-to-five episodes, or something like that, just as a sort of "for those of you just joining us" thing.
Not to disagree, or at least not entirely, with the basic sentiment, but I have to call you on this: your use of the term "strip-mining" stikes me as a cheap rhetorical device. It invokes images of the destruction done by evil corporations as they strip the surface off huge patches of land because it's a cheaper way of getting at the ore than digging a little, etc. However, does this really apply to the moon, which has no ecosphere? I.e., there's nothing to destroy. Really, how could you even tell if a piece of the lunar surface had been "strip-mined"? Moreover, who would care? The emotional appeal regarding the badness of "strip-mining" is meaningless.
By the way, according to Strangers From the Sky, (I know, the books aren't "official") we actually encounter Vulcans some twenty years before the Amity incident -- one of their scoutships crashes on Earth and, though the whole incident is covered up, Kirk and friends, who, through one of their little accidents, just happen to be around, help them get home. Please don't say your entire knowledge of Star Trek history is based on that TNG movie.
Right. The different toolsets available for the respective systems should be considered in rating them. That is, if a program exists for one and not the other, and it is better than the nearest equivalent on the other system, then that should be counted as a "point" in favor of that system.
I put "point" in quotes because it should not actually be counted directly, as in a checklist. Instead, the best tool should be used on each system in performing the test, and whatever advantage it gives will automatically reflect well on that system.
Rather than requiring both systems to use the same software, the testers should spend similar amounts of effort and expertise looking at the tools available for each system and selecting the best they can find.
I've been thinking of somethingh along the same lines as Emacs' "M-x spook" command -- "M-x spook" inserts a few dozen randomly-selected keywords (e.g., nuclear, Iran, terrorist, president, bomb, etc.). The idea is for people to routinely append this to their e-mails on the theory that Echelon is scanning messages for such keywords, and that causing it to come up with huge numbers of false hits would decrease its effectiveness, ideally making it impractical. At least one Slashdot user has done this in his signtaure -- he has a comment further down on this thread.
Anyway, my idea: on the theory that "they" are using encrypted communication as grounds for suspicion of other crimes (i.e., "He sent an encrypted e-mail, so he must have something to hide..."), what if people just start e-mailing random noise to each other? Since, with the right algorithms, it is not even possible to prove that there is a message hidden in a transmission, "they" would not be able to tell the difference between an e-mail that contains an encrypted message and one that is just random noise. Hence, if people just routinely mail a few dozen kilobytes of "/dev/random" output back and forth, what could "they" do? It would become impractical for "them" even to identify suspects based on use of encryption, since there would be too many false hits.
I like the idea of giving moderators access to only some randomly-selected subset of comments (an appropriate value of "20%"). As I pointed out the other day, moderation tends to be unfairly concentrated on the earlier comments, since these are the ones that more moderators see, and since, even when a moderator reads an old story, he will probably want to use his precious points on a "live" story, rather than "waste" them on something that not as many people will see. The problem is that stories go by, and comments appear, so quickly. I suspect that people actually keep reading stories somewhat longer, after the commenting has died down, so it is actually desirable to have moderation happen symmetrically, aside from the issue of fairness to people who take more time and post thoughtful comments.
I had also thought of assigning moderator access on a story-by-story basis, but I think this is better, since it could also take timing into account, making sure that all parts of a discussion get similar coverage, as well as reducing the possibility of abuse.
Also, "Put 5x... as many moderators into the system" is not quite clear. I assume the increase would be implemented by some combination of (a) giving access to more people, (b) giving each person more points, and (c) giving each person points more frequently. Is that right? I suggest using mostly (c), and not as much (a).
As for "Increase the percentage of moderatable comments based on karma", you could just as well grant more points, or grant points more often, based on high karma. The point is that now karma would affect a moderator's "power" in more than an all-or-nothing way, which is a qualitative difference. I guess having it affect the percentage is the lowest-impact way of doing this, since it only give the moderator more choices. This change could be kind of controversial, but I think it's fine. In fact, I think giving more points would make more sense than giving more choices, since karma is (theoretically, at least) a predictor for making good moderation judgements, and not so much for deciding which comments to moderate. There is still no way for moderators to collaborate, meaning that that decision should be centralized.
Finally, I kind of like M2 the way it is, on a separate page. The fact that it's an added chore seems like a good selector, in that only dedicated people will do it. Also, it has a nice side effect of making me read some comments that I had otherwise missed. I was even thinking that ordinary moderation should be pulled out onto a separate page, to make it a bit more of a chore. Maybe the front page could have an additional link "Moderate" next to "Read More...", which would take you to a differently-formatted version of the page, optimized for moderation: nested view, names, sigs, and current scores stripped, and maybe even restricted to show only the comments for which moderatability has been assigned, with their context.
By the way (yeah, I know I said "finally" last paragraph), here's a suggestion for the problem of context in M2 pages, as well as for my idea above: show a comment's ancestors, but no siblings. I know replies sometimes include sideways references, but most of the flow should be recoverable by just following the chain straight up.
I don't know if it solves all the problems, but I think your idea definitely has potential. Something like this would certainly be an improvement, though a lot of details would have to be worked out. I for one would like to see the idea developed.
I guess there's still the problem that the names are textual, which could lead to contention -- if they are required to be the organization's real name, people would just start clamoring to get the "cool" company names. Already, plenty of internet start-ups have "dot com" in their names. Then again, they wouldn't have to do that if normal names could be used to find web sites.
Of course, the existence of a problem does not imply the existence of a solution -- and how could you possibly avoid contention while using textual names? I think the problem is inherent to the concept, but I suppose it can be worked around and minimized.
Would your idea allow exact duplicates? That would seem to cause trouble, but I think it would actually help: when someone registers a name that is identical to an existing one, he could be required to provide some distinguishing information, and the other guy could be contacted and given the opportunity to provide some additional information as well, so that when you do the query, you would see the same name twice, with some annotations to help tell them apart.
Alternately, it could allow both entries to have exactly the same name, but contact both parties and give them the opportunity to add annotaions. In the meantime, the entries could just look the same, which is presumably undesirable to both parties. Even if one or the other is playing dirty, e.g., trying to steal the other guy's "eyeballs" (ugh!), at least one of them will want to add something.
It's just a directory entry, though, so we can't let them write a book in the description field. Hopefully most people wouldn't get too carried away, but there should be a hard limit at which it resorts to first come, first served and the new guy would have to think of something else, or maybe just append a number. To clarify, I'm talking about someone duplicating not only the name but the description, in which case there would almost have to be someone who is clearly trying to play dirty, so it could be decided fairly enough.
Also, it would be a pain to have to do this query every time, so users need a way to get back to the site they actually wanted once they've found it. This can be handled by a simple bookmarking system like browsers already have, though it should store the actual IP addresses, or some other unique identifier, rather than the query that was used. Also, since users would want to rely on them more, it would be important to have a better system for organizing bookmarks. Maybe some kind of lightweight database application on the user end.
Slashdot has been having trouble scaling to keep up with its growth, and I'm not talking about the servers. With the rate that new stories go up, and the number of comments on each, it has been becoming difficult to keep up. Discussions don't usually stay active long enough to do them justice, because they get crowded out, with some stories falling off the bottom of the page the same day they are posted.
I've been thinking that the problem is that it is organized linearly in time, which doesn't scale too well. Maybe a solution would be to "add an axis", making the layout more two-dimensional by breaking it into sections, each of which would proceed linearly. It's not perfect, but it would allow each story to stay visible for a longer period in its section. It would still be hard to keep up with everything, simply because of the sheer volume, but people who want to skim everything and comment sporadically could do so, while those who want to discuss a single issue in greater depth would also be able to.
Maybe some of the most interesting threads could even be kept alive by using activity as well as age to determine the order of the stories. Also, when the number of comments on a story gets too unwieldy, maybe some of the most active sub-threads could be pulled out into new stories -- these could be kept in a separate sidebar for each section.
I've been thinking about something like this too. Of course, as Todd Knarr pointed out in his reply, what you're describing does not sound very different from a search engine, especially one like Yahoo, which keeps the sites in a hierarchy by category.
Search engines are nice, but the problem is that now they've turned into "portals", becoming some of the very worst offenders in the hunt for "eyeballs". That is, they don't just find you sites that match your query anymore: now they are all clamoring to be the center of the entire web experience, by providing news, links, etc. Of course, they can't force you to use these, but they make it so convenient that a lot of people tend to just go where the portal tells them to, instead of exploring the web on their own. To me at least, all this talk of "eyeballs" is incredibly offensive, because it implicitly assumes that I'll look at whatever they show me and buy whatever they advertise. In short, it is an attempt to turn the web into television.
I think the problem with domain names is that they are logical identifiers with semantic meanings. When hosts were just identified by IP numbers, it was not contentious, but when textual names are mapped to those numbers, suddenly everyone wanted the cool names, and now that businesses and consumers have gotten involved, they also want the catchy names.
I can't think of a way to fix this except something like your directory idea, but I hope it can avoid becoming too commercialized. The thing is, people don't seem to mind the concept of a phone book -- if you want to call someone and don't have the number, you look up the name in a directory and then make the call. This directory system should be more like that.
"Ultimately, the domain name system is about your ability to be seen and heard in the digital age."
This infuriated me, since it seems to me that that is exactly what the domain name system should not be about. All anyone seems to care about is having a URL that is catchy and easy to remember. The actual quality of the content or servce that a site provides hardly seems to matter by contrast.
I think a lot of us are against both domain squatters and domain bullies. It's not about big guys vs. little guys this time. Rather, it's because both activities are based on this horribly commercialized concept that the web is all about companies pushing and shoving to attract the all-valuable "eyeballs" to their sites.
One more thing: you said "Well, I just can't wait for the use of domain names to be deprecated." Do you know something I don't about an alternative system that will take care of these problems, or are you just expressing the same hope that I share, that some such alternative will come along eventually?
Cironian was asking whether the MP3 decoding circuitry would significantly increase the power drain. That is, he realizes that this thing has the same mechanical and optical parts as a CD player, plus the additional chips for the decoding, and he wondered whether this addition would cause the battery life to suffer relative to a CD player.
Hrunting pointed out that, for reference, a Rio lasts longer on one battery than a CD player on two, from which we can conclude that its decoding parts take less than half the power of the CD player's mechanical and optical parts. Even less, really, since both also have the load of the actual output signal. Hence, with the addition of the decoder, the whole thing should take at most half again the power of the ordinary CD player, probably a lot less, reducing the battery life by at most a third, probably a lot less. So it was a perfectly good answer, if not fully explicit.
My addition: common sense says that a microchip should draw far less power than either a motor+laser or the output signal (though I'm just guessing, since I don't have any numbers), so to answer the original question, I would expect the extra load to be dwarfed by the others -- practically lost in the noise. I don't think it would make much difference at all.
ObDrool: I really want one of these, though I'm still pretty interested in one that would work with an IBM MicroDrive (340 MB, 1", same form factor as complact flash cards). The article on the RCA "Lyra" was the only one I saw that mentioned this possibility, but wouldn't it work with anything that accepts a compact flash card? Specifically, does anyone know if a Rio can take a MicroDrive? If not, why, and if so, why don't they advertise it?
It looked like my karma was bumped up when I meta-moderated yesterday, but not today. Maybe it's some fraction of a point (rounded up), maybe he changed it, or maybe I hallucinated.
What I didn't get is, he said the reward for meta-moderating would never put you over the top for (I assume he meant) moderator access or the initial 2. In that case, what's it good for? Unless karma has some additional meaning, I don't get it. If this reward is not counted toward the things that karma is good for, then it doesn't serve any real purpose. Just seeing a bigger number on your profile page would be a psychological reward, but it would also cause confusion if the number stopped corresponding to the actual value used to calculate the other things.
I think having it count is fine. Moderating and meta-moderating are really "community service", since they are activities that take time and thought and aren't as rewarding as posting. This is even more true of meta-moderation, since its results are not as visible and it cannot be done as part of ordinary reading, but requires going to a special page. However, it's right that the reward should be small, since otherwise it would be too easy -- it's hard to get unearned karma by having one's posts upgraded, since someone else's decision is involved, but this would be easier to abuse.
It also wasn't quite clear (or maybe I missed it) what effect meta-moderation has. If a moderator's decisions are found to be unfair, is his access revoked, or does he lose karma (eventually making him ineligible), or what? Such punishments should not be on a hair-trigger, but it makes sense to have some such system. (What else is the point, right?)
How do these sound for karma awards:
Using all granted moderator points: +1 Since it's perfoming a service, and you had to refrain from posting in order to do it.
Meta-moderating for a particular day: +1/3 Also a service, and takes extra time, but easier to do, and hence easier to abuse, so smaller amount.
Having one's moderations judged ("fair"/"unfair"): (+/-)1/10 I assume this is the whole point of meta-moderation. For symmetry, there should be both a reward and a punishment. They should both be small to keep the total effect reasonable.
I'm not so sure about having karma expire -- the problem is that it is unfair to infrequent posters, who might always make good comments, but never enough within the time window. Of course, I recognize that the alternative greatly favors longtime members, Maybe the best would be to use the average score, or base the expiration on "the most recent n comments", instead of "the past m months." Plus, I like the idea of having it be a cumulative history of one's participation. It would be better for the threshold to be incrementally increased for older members, instead.
For one thing, the "controls for everyone" thing would work a lot better if the points don't expire. Then, you wouldn't have to worry about your points being "wasted" if you don't moderate every day. Also, a problem is what if you moderate two comments on a single page when you only have one point? it would only be counted for (I assume) the first one, but maybe you thought the second was more important to moderate, if you had to choose, and the point would be wasted on the one that you didn't want to moderate as much. When you know exactly how many points you have, you can make more intelligent decisions as to how to make them count.
Then again, I guess the point was to encourage moderators to detach their egos from their moderation, and having no idea whether or not it would actually be counted would sort of have to have that effect.
Dunno: it's interesting, but I don't really like the idea that much. I have to agree that it's better to know when you have points. As Jburkholder said, you should think in a different mode, i.e., evaluating the quality of each post rather than just reading them for content. Ryandav alluded to the same point regarding browsing at -1, though I always do that anyway. Another reason it's really important to know when you're moderating is that you have to know to refrain from posting on a discussion so you can moderate it. It would be a shame to refrain from posting just so you could submit some votes, only to have them ignored.
This sounds better, if I underatand correctly -- I assume you mean that when the system grants points to someone, it would also assign a subset of current stories in which those points may be used. Would the user be notified, perhaps with a "MODERATE" mark on the Slashdot front page for the selected stories? That would have the effect of evening moderation out among stories of varying popularity, and increase fairness. It would also help moderators with the "so many posts, so few points" dilemma. The problem with moderators thinking "I have to make these points count by finding some really good posts to upgrade." is that a lot of posts that deserve a 2 or 3 never get it.
That brings up an interesting question: I'll admit that I'm pretty interested in my "karma", and want to gain as much as possible, insofar as it's a measure of "acceptance" within the community, or some such warm fuzzy feeling. But is it good for people to become obsessed with this? If people start thinking "I want to get my posts moderated up, so I can become a moderator and gain some POWER; better yet, wait until I break the 20-barrier, so I can get automatic 2's! Then everyone will have to read what I write! <evil cackle>", it will become an ego thing (or more so), which could interfere with the "purity" of the discussions. Moderation already did some of that, I guess, but this would seem to increase it. On the other hand, maybe the effect is good, since it would encourage people to think more and flame less.
See the difference? Worrying about those posts that "deserve" to be upgraded only matters if we care about karma for its own sake, i.e., we consider it a reward, and think the poster should be rewarded for writing something good, even if it's not a story that many people are interested in, or if the story is already a few hours old and nobody's reading it anymore [1]. But if the purpose of moderation is only to provide an aid to readers who want to find the good posts, then it's not such a problem, since in those cases fewer people need the aid anyway.
[1] Seriously, it's amazing how fast the stories go by these days. I guess it's an inevitable result of Slashdot becoming so big, but it makes it a lot harder to participate in the discussions. If all the stories have over 300 comments by the time you get to them, all you can really do is read. If you take the time to write a really thoughtful comment, chances are hardly anyone will still be reading the story by the time you post it. This is often my problem. I'm not sure how your idea would help with this problem, but I think it's an important one.
Slashdot Mirror
I was thinking that an even better activity would be a "[inter]National Crypto Day", when we all send each other lots of encrypted messages. In fact, I just sent the following message to a friend of mine (it's a private message, but I don't see any harm in posting it here):
btfzuqhqzeclyfvchjxgdrkpetnwbvosxfqswbv
upashzmbrexhinncdshzutjnlizgmzntgwgyqmr
tqklcsncasicehglnsslprhhjlgbflsoarsukuf
ytlupzwwhkjxhtsydpnodnapgcajjeeowbpmpkw
hprrdonbijmyhbpiurzkroajcxbhsfzxivjorho
uokreaftfhuygtleqeejovbjhsspiaxpgtztejx
tyzyprwuyoeaczpjmhbrjptxmjhxhubocfjtuce
ncijzzejaterdyqzzqvyjfcwydiwopoqcjttscq
lbuurunpkpzluripwnclvkubdzjoocvtifknbpm
mwnrwqzpwznrlwbcqavnyjbtuknyusdlwbuuoer
rbhvghjatdqnqjxvmznkbswsdgoraifwxavemnr
rrdlnffgijxtyelixvlyilrrdswlnmblhyvrmvn
ggubrlgwhewnjpoidkstnhnlcbmkuqjvegknufw
zcwlkwdhkwqzrkpbmoirbigdxidyamnvw bxkid
mkacukj
This would seem to be even more effective than "M-x spook" - style keyword spoofing, since we'd really be showing that "they" can't read our mail if we don't want them to, and exercising our right to send mail that they can't read.
Additionally, I would suggest sending some messages that are just randomly-generated gibberish. This would be a response to any proposal to ban crypto: it would make the point that you can't prove that a given message consisting of seemingly-random data is actually a ciphertext at all, and not really just random. Banning the transmission of meaningless random numbers would be absurd, which means that banning the transmission of ciphertext would be impossible. Suppose I had sent the following to my friend instead:
lrfkqyuqfjkxyqvnrtysfrzrmzlygfveulqfpdb
hlqdqrrcrwdnxeuoqqeklaitgdphcspijthbsfy
fvladzpbfudkklrwqaozmixrpifeffeclhbvfuk
byeqfqojwtwosileeztxwjlkngbqqmbxqcqptkh
hqrqdwfcayssyoqcjomwufbdfxudzhiftakczvh
sybloetswcrfhpxprbsshsjxdfilebxwbctoaya
xzfbjbkrxirimqpzwmshlpjhtazhbuxhwadlpto
yeziwkmgsovqzgdixrpddzplcrwnqwqecyjyibf
jykmjfqwltvzkqtpvolphckcyufdqmlglimklfz
ktgygdttnhcvpfdfbrpzlkvshwywshtdgmbqbkk
xcvgumonmwvytbytnuqhmfjaqtgngcwkuzyamne
rphfmwevhwlezohyeehbrcewjxvceziftiqtntf
srptugtiznorvonzjfeacgamayapwlmbzitzszh
zkosvnknberbltlkggdgpljfisyltmmfvhybljv
kypcflt
I would be seriously impressed if anyone (aside from myself or my friend) could prove that the first is a secret message, let alone tell me what it says. However, neither can I prove that the second is not a secret message (in fact, you don't have to believe me when I claim that), so if they were to reverse the burden of proof, I would have a problem -- they could punish me for refusing to surrender a non-existent key. See this story from a couple weeks ago.
Of course, all this assumes that there really is a "they" trying to read everyone's mail -- I'm stating no opinion on that.
David Gould
I left out a few words: "being twice the number of letters in the alphabet" should have been "being two more than twice the number of letters in the alphabet", since Solitaire uses the standard 52 cards of the deck, with two mapping to each letter of the alphabet, plus the two jokers as fenceposts, or placeholders, to divide the deck into two groups, for a total of 54.
The algorithm consists of several fancy cuts, moving the jokers around, etc., and ultimately selects a single card. The process also puts the deck into a different permutation (and there are about 2.3 * 10^71 permutations), so repeated applications will produce a pseudo-random sequence of cards. The numerical values of the cards are used as a one-time pad to encrypt the message. This solves the problem of key distribution for one-time pads, because two parties can start with identically-shuffled decks and generate the same keystream. I don't know the math of how it works, but supposedly the pseudo-random stream is good enough for the system to be pretty secure.
David Gould
The number 54, being twice the number of letters in the alphabet, as well as the size of a deck of playing cards, including jokers, sort of brings to mind Bruce Schneier's "Solitaire" cryptosystem, or at least it does when you mention Cryptonomicon at the same time.
If you haven't read it: Schneier devised a cryptosystem called "Solitaire" that is you can use to encrypt and decrypt messages by hand with a pack of cards. In Cryptonomicon, two characters use this method to communicate, and Schneier's paper describing the system is included as an appendix.
Though low-tech, Solitaire is supposedly secure against modern attacks, so if this contest uses it, it would be pretty hard. Then again, the number 54 and the relation to the book are the only evidence for it, so it's probably unrelated.
David Gould
They always release a new Milestone the day after I download the previous one. Of course, in this case, I'll admit it was kind of dumb of me to download M9 yesterday when the Milestone page clearly stated the target date for M10 was 10/08/99. That whole discussion just got me itching to download something, so I did.
I'm about two hours away from being able to try M10, but I have some comments on the Mac version of M9: the browser functionality is coming along nicely -- for example, it loads Slashdot perfectly, though it still can't log in. The front-end is starting to look relatively polished, but my concern is that it doesn't look like a Mac application. Actually, with previous builds, I assumed that it was simply because the GUI-polishing was being left for last. Now, though, it seems to be starting to look the way it's meant to look, if that makes any sense. I'm seeing the interface that they're designing, and it doesn't seem Mac-like.
As I understand it, they use some sort of cross-platform front-end framework to make all versions use a single codebase. This doesn't seem to use native widgets, so things look funny. This could be okay, if the widgets were similar enough, and if they worked right (even so, the non-native look-and-feel would be a turn-off for many Mac-users), but in fact, many of the text fields, radio buttons, etc., don't draw quite right. Text fields are too tall and narrow for their text, radio buttons are too small, and a few of their pixels get chopped off, the sidebar in the Preferences window sort of jumps when I click something, etc.
This sort of thing is a problem with cross-platform GUIs, I guess, because corresponding things have different relative sizes, and some amount of platform-specific attention is needed to make everything fit. I don't know how this compares to other platforms; maybe the Mac version is just not getting enough attention.
By the way, what's with the second date for M12 on the Milestone Page? "M12 - 12/7/99 - 15/99/99"? Is somebody having Y2K trouble?
David Gould
"An Architectural Review Board should be formed to maintain the source trees. "
This is a really dumb idea - a govt. committee that decides software architechture...
I didn't say it would be a government committee. I don't think it should be, because of your points, plus I think that really would be too much government interference. I want the government to make this happen, but not be (unnecessarily) involved in the implementation.
It would be an industry committee, like the way OpenGL is managed, or the Object Management Group (OMG) for CORBA. An Open Source project like this would need an organization to serve as a code maintainer, and I just thought some rules for its formation, such as the restrictions on Microsoft and the opportunity for anyone who produces an operating system or a web browser to be represented, should be part of the process. Someone has to approve patches, and the proposal wouldn't be complete without some suggestion as to how it would be managed. Membership in the group would be pretty much open to anyone, though there would have to be some mechanism to make sure it didn't get stuffed with people with any particular agenda; I just listed some of those who would probably be most interested in being members.
I guess the reason I scare-quoted "official" is that there wouldn't be any government mandate making this group's distribution any more valid than anyone else's, assuming someone else decided to fork off a version and set up their own source tree. This would just assure that there would be a version to step into the gap left by Microsoft's closed one. Primarily, they would be there to keep an eye on what Microsoft did with the code. Realistically, Microsoft would, for a while at least, remain the major developer and distributor, and this would just make sure it stayed open.
Anyway, I guess the ARB is not the most important part of the proposal. I just thought it would be a good idea to specify some of how the code would be maintained, and to keep Microsoft from retaining de facto control over it, but the really important part is that the code be donated to the public under the GPL. I don't see how anything less could undo the damage Microsoft has, in my opinion, done.
David Gould
The source code for the offending programs, namely Windows 95/98, Internet Explorer, and whatever other products were used for anticompetitive purposes, should be seized and given to the public under the GPL to compensate for the damage they have done.
An Architectural Review Board should be formed to maintain the source trees. The ARB should represent high-tech companies such as IBM, Sun, Apple, SGI, HP, Compaq, Red Hat, etc. Microsoft can be a member, but not with any controlling authority. The ARB should also include representatives of the OSS/Free Software community. These would be a bit harder to choose, but clearly the FSF should be represented, as well as the Mozilla team, etc.
Changes to the main source tree would have to be approved by the ARB. The top-of-trunk code, as well as any derivatives, would be publicly available to anyone wanting to look at it or use it, under the terms of the GPL. That means that anyone wanting to fork off their own version and distribute it could do so, but not without keeping their own source open, etc. Microsoft would be allowed to "innovate" by making contributions (sure, why not?) -- even their application-development divisions would be able to work in tandem with the OS development (again, why not?), but there would be no secret APIs, because all such contributions would have to be open. As an extra provision, Microsoft (at least for some period of time) would not be allowed to distribute any version not approved by the ARB, or any binary-only form. That is, their contributions would have to be adopted into the "official" top-of-trunk version before they could be distributed. This would prevent their use of "embrace-and-extend(-and-extinguish)" tactics.
This might seem excessive, or "cruel and unusual", but it's really perfectly reasonable. If software can be "intellectual property" at all, then it can be confiscated, just like any other kind of property, if it is used to commit a crime. Since this property was the instrument used to commit the crimes, confiscating it would be no different from confiscating a drug dealer's car. Placing it under the GPL would then amount to donating it to the injured parties.
This is approximately the same comment I made back in February, in response to a proposed "source-licensing punishment". The archived page does not have comment links, but mine is about three-fifths of the way down, or just search for my name.
David Gould
Unless I'm mistaken (it could happen), a quantum computer could break a key of any length in constant time, up to the number of qubits that it is able to maintain. That last part is the important disclaimer, since making qubits behave correctly (i.e., making them collapse correctly and not prematurely) is really hard, and gets harder the more you try to use.
Last I heard (and I don't claim to be at all up-to-date), getting a quantum computer to maintain 4 qubits, so as to factor 15 into 3 and 5, would be considered quite a feat. A 512-qubit computer would work on the same principles, and would factor 512-bit numbers just as quickly as it would factor 15, but I'm as skeptical as everyone else as to whether they've managed to construct such a thing.
Even if they have, it doesn't follow that larger keys are vulnerable: even if they managed the amazing feat of building a 512-qubit quantum computer, it still wouldn't be able to factor numbers larger than 512 bits. Breaking a four-kilobit keys would require yet another three and a half kilo-qubits, which they don't necessarily have.
On the other hand, making the leap from a couple of qubits to 512 would suggest that they have developed a technique for adding qubits fairly easily, perhaps with polynomial or even linear cost, in which case they would be able to make one with four kilo-qubits, a mega-qubit, or whatever, without exponentially greater effort.
David Gould
I thought I read that it was something like this: First of all, Oracle did it in ~71 seconds, meaning that SQL Server would have to come in just under 2 hours to "win".
However, it also turned out that due to some loophole in the definition of the TPC-D benchmark, it was arguably legal to set up the system to effectively precompute the query while loading the dataset, which is not counted in the time, so that all you have to do on the clock is spit out the results. However many hours the system takes to load the dataset does not count.
Using this technique, Oracle did it in about half a second, and SQL Server did it in less than two -- still slower, but not 100x. If you consider this technique valid, then I guess MS won, but it's pretty clear, to me at least, that in any reasonable interpretation, this would constitute cheating.
I never heard of any times for a SQL Server system doing it the honest way, or even whether or not anyone got it to happen at all.
David Gould
As I said above, I'm not too impressed with the "Clotho" idea -- it sounds too much like a "portal", and I ranted a bit about those, too. I'm not very worried about any "cyberclysm", either.
What I think Katz misses is that we won't just use machines, as we do now, to exchange and process information, but also to understand it. He mentions Ray Kurzweil, but if he'd read and understood The Age of Spiritual Machines, he would rememebr that Kurzweil expects us to merge with the machines, not just use them as assistants, which is what I've also been imagining for a while now.
Consider a scientific application in which the computer collects and processes a huge amount of data -- far more than the scientist could ever keep in his head. The computer doesn't just spit out the answer in the form of a number, let alone a new theorem or a complete paper. Instead, it generates a visualization that allows the scientist to understand the data set intuitively despite being unable to see all the details. The computer then also spits out some numbers that are the result of a full analysis, and between these, the scientist now understands a greater amount of information than he could have without the machine.
This is not even a very good metaphor for the real idea, which is that computers will become part of our consciousness. A direct neural interface doesn't just mean that you'll see a window manager in your head and move the mouse by thought -- you'll actually be able to be aware of all the information as it is being processed. Imagine your mind working thousands of times faster: then time would subjectively pass more slowly, and you'd be able to ingest far more information in the same amount of time. Now add in improved recall, visualization, and mathematical ability.
Not only will we have access to more information, we'll be able to experience more of it, more fully. I wish I could read every single comment on Slashdot, and post ten times more than I do, but even if I spent 168 hours a week at it, there's just no way, because my mind is too slow. Katz' idea sounds basically like perfecting the moderation system and adding good search capabilities. That would be nice, but it's nowhere near the full experience that I envision.
I can't stress enough that I'm not just talking about having all that information at our virtual fingertips, but really knowing it as intimately as we do our names and the first five natural numbers. There need be no "cyberclysm" if our intelligence increases as quickly as our needs. There may be some hiccups and bottlenecks if the paces don't match perfectly, but there's no inevitable explosion looming.
David Gould
I don't really think of search engines this way, since they are about searching for pages with the particular information that I am looking for. The only sense in which they "tell" me what to read is that they produce a list of things that match the description I give, which is really not the same thing. Of course, search engines are now acting increasingly like "portals", bombarding us with book, movie, and music recommendations, etc., along with our query results.
The very concept of "portals", however, really bugs me, since they are all about telling you where to go. They present a list of things that they think you'll find interesting, and you're supposed to just follow them. It's fast, easy, and convenient! It saves you the trouble of searching through the whole web by taking you straight to the good stuff! You don't have to understand those big, complicated URLs; just click here! Arghh! I for one see it as an attempt to undermine the user-centric, interactive nature of the web, which is what makes it so different from other media in the first place. This is what people like Katz have gotten so excited about -- the democratization of information flow. But with a "portal", it turns back into a provider/consumer relationship. Of course, the user can still navigate freely, but portals are all about discouraging this. "Ideally" (as far as they're concerned), the user would confine his entire web experience to the links presented by the portal. Why do you think we keep hearing about "eyeballs"? It's because what matters to them is the ads, which are based on the model of them telling the consumer what to see, what to think, and what to buy. Basically, they're trying to turn the web into television.
This is really what "Clotho" sounds like. I know the idea is for it to rise above all the commercialism, but it just does not seem possible for something like this to remain "pure". It seems inevitable, first, that there would be a proliferation of sites claiming to offer this service, but which would really be just a bunch of search engines and "portals", and also that they would end up being big, overhyped (and ridiculously overvalued) commercial entities that would try to control what content, and what ads, reach all those "eyeballs".
David Gould
I'd like to see a list of links to the mp3 files for all the episodes right here on Slashdot's "index.pl?section=radio" page. I don't really care whether the files are mirrored here, elsewhere, or linked directly to The Sync. I'd just like to be able to get the latest episode, and any that I've missed, right here, instead of having to go to http://thesync.com/geeks for the latest and http://thesync.com/geeks/archives.html for the older ones.
As for putting it on the front page vs. the "radio" section, I think the best thing is to keep it here in the side section, but mention it (at least in a "quickie") on the front page every couple months, or every three-to-five episodes, or something like that, just as a sort of "for those of you just joining us" thing.
David Gould
Not to disagree, or at least not entirely, with the basic sentiment, but I have to call you on this: your use of the term "strip-mining" stikes me as a cheap rhetorical device. It invokes images of the destruction done by evil corporations as they strip the surface off huge patches of land because it's a cheaper way of getting at the ore than digging a little, etc. However, does this really apply to the moon, which has no ecosphere? I.e., there's nothing to destroy. Really, how could you even tell if a piece of the lunar surface had been "strip-mined"? Moreover, who would care? The emotional appeal regarding the badness of "strip-mining" is meaningless.
By the way, according to Strangers From the Sky, (I know, the books aren't "official") we actually encounter Vulcans some twenty years before the Amity incident -- one of their scoutships crashes on Earth and, though the whole incident is covered up, Kirk and friends, who, through one of their little accidents, just happen to be around, help them get home. Please don't say your entire knowledge of Star Trek history is based on that TNG movie.
David Gould
Right. The different toolsets available for the respective systems should be considered in rating them. That is, if a program exists for one and not the other, and it is better than the nearest equivalent on the other system, then that should be counted as a "point" in favor of that system.
I put "point" in quotes because it should not actually be counted directly, as in a checklist. Instead, the best tool should be used on each system in performing the test, and whatever advantage it gives will automatically reflect well on that system.
Rather than requiring both systems to use the same software, the testers should spend similar amounts of effort and expertise looking at the tools available for each system and selecting the best they can find.
David Gould
I've been thinking of somethingh along the same lines as Emacs' "M-x spook" command -- "M-x spook" inserts a few dozen randomly-selected keywords (e.g., nuclear, Iran, terrorist, president, bomb, etc.). The idea is for people to routinely append this to their e-mails on the theory that Echelon is scanning messages for such keywords, and that causing it to come up with huge numbers of false hits would decrease its effectiveness, ideally making it impractical. At least one Slashdot user has done this in his signtaure -- he has a comment further down on this thread.
Anyway, my idea: on the theory that "they" are using encrypted communication as grounds for suspicion of other crimes (i.e., "He sent an encrypted e-mail, so he must have something to hide..."), what if people just start e-mailing random noise to each other? Since, with the right algorithms, it is not even possible to prove that there is a message hidden in a transmission, "they" would not be able to tell the difference between an e-mail that contains an encrypted message and one that is just random noise. Hence, if people just routinely mail a few dozen kilobytes of "/dev/random" output back and forth, what could "they" do? It would become impractical for "them" even to identify suspects based on use of encryption, since there would be too many false hits.
David Gould
I like the idea of giving moderators access to only some randomly-selected subset of comments (an appropriate value of "20%"). As I pointed out the other day, moderation tends to be unfairly concentrated on the earlier comments, since these are the ones that more moderators see, and since, even when a moderator reads an old story, he will probably want to use his precious points on a "live" story, rather than "waste" them on something that not as many people will see. The problem is that stories go by, and comments appear, so quickly. I suspect that people actually keep reading stories somewhat longer, after the commenting has died down, so it is actually desirable to have moderation happen symmetrically, aside from the issue of fairness to people who take more time and post thoughtful comments.
I had also thought of assigning moderator access on a story-by-story basis, but I think this is better, since it could also take timing into account, making sure that all parts of a discussion get similar coverage, as well as reducing the possibility of abuse.
Also, "Put 5x
As for "Increase the percentage of moderatable comments based on karma", you could just as well grant more points, or grant points more often, based on high karma. The point is that now karma would affect a moderator's "power" in more than an all-or-nothing way, which is a qualitative difference. I guess having it affect the percentage is the lowest-impact way of doing this, since it only give the moderator more choices. This change could be kind of controversial, but I think it's fine. In fact, I think giving more points would make more sense than giving more choices, since karma is (theoretically, at least) a predictor for making good moderation judgements, and not so much for deciding which comments to moderate. There is still no way for moderators to collaborate, meaning that that decision should be centralized.
Finally, I kind of like M2 the way it is, on a separate page. The fact that it's an added chore seems like a good selector, in that only dedicated people will do it. Also, it has a nice side effect of making me read some comments that I had otherwise missed. I was even thinking that ordinary moderation should be pulled out onto a separate page, to make it a bit more of a chore. Maybe the front page could have an additional link "Moderate" next to "Read More...", which would take you to a differently-formatted version of the page, optimized for moderation: nested view, names, sigs, and current scores stripped, and maybe even restricted to show only the comments for which moderatability has been assigned, with their context.
By the way (yeah, I know I said "finally" last paragraph), here's a suggestion for the problem of context in M2 pages, as well as for my idea above: show a comment's ancestors, but no siblings. I know replies sometimes include sideways references, but most of the flow should be recoverable by just following the chain straight up.
David Gould
I don't know if it solves all the problems, but I think your idea definitely has potential. Something like this would certainly be an improvement, though a lot of details would have to be worked out. I for one would like to see the idea developed.
I guess there's still the problem that the names are textual, which could lead to contention -- if they are required to be the organization's real name, people would just start clamoring to get the "cool" company names. Already, plenty of internet start-ups have "dot com" in their names. Then again, they wouldn't have to do that if normal names could be used to find web sites.
Of course, the existence of a problem does not imply the existence of a solution -- and how could you possibly avoid contention while using textual names? I think the problem is inherent to the concept, but I suppose it can be worked around and minimized.
Would your idea allow exact duplicates? That would seem to cause trouble, but I think it would actually help: when someone registers a name that is identical to an existing one, he could be required to provide some distinguishing information, and the other guy could be contacted and given the opportunity to provide some additional information as well, so that when you do the query, you would see the same name twice, with some annotations to help tell them apart.
Alternately, it could allow both entries to have exactly the same name, but contact both parties and give them the opportunity to add annotaions. In the meantime, the entries could just look the same, which is presumably undesirable to both parties. Even if one or the other is playing dirty, e.g., trying to steal the other guy's "eyeballs" (ugh!), at least one of them will want to add something.
It's just a directory entry, though, so we can't let them write a book in the description field. Hopefully most people wouldn't get too carried away, but there should be a hard limit at which it resorts to first come, first served and the new guy would have to think of something else, or maybe just append a number. To clarify, I'm talking about someone duplicating not only the name but the description, in which case there would almost have to be someone who is clearly trying to play dirty, so it could be decided fairly enough.
Also, it would be a pain to have to do this query every time, so users need a way to get back to the site they actually wanted once they've found it. This can be handled by a simple bookmarking system like browsers already have, though it should store the actual IP addresses, or some other unique identifier, rather than the query that was used. Also, since users would want to rely on them more, it would be important to have a better system for organizing bookmarks. Maybe some kind of lightweight database application on the user end.
David Gould
Slashdot has been having trouble scaling to keep up with its growth, and I'm not talking about the servers. With the rate that new stories go up, and the number of comments on each, it has been becoming difficult to keep up. Discussions don't usually stay active long enough to do them justice, because they get crowded out, with some stories falling off the bottom of the page the same day they are posted.
I've been thinking that the problem is that it is organized linearly in time, which doesn't scale too well. Maybe a solution would be to "add an axis", making the layout more two-dimensional by breaking it into sections, each of which would proceed linearly. It's not perfect, but it would allow each story to stay visible for a longer period in its section. It would still be hard to keep up with everything, simply because of the sheer volume, but people who want to skim everything and comment sporadically could do so, while those who want to discuss a single issue in greater depth would also be able to.
Maybe some of the most interesting threads could even be kept alive by using activity as well as age to determine the order of the stories. Also, when the number of comments on a story gets too unwieldy, maybe some of the most active sub-threads could be pulled out into new stories -- these could be kept in a separate sidebar for each section.
David Gould
I've been thinking about something like this too. Of course, as Todd Knarr pointed out in his reply, what you're describing does not sound very different from a search engine, especially one like Yahoo, which keeps the sites in a hierarchy by category.
Search engines are nice, but the problem is that now they've turned into "portals", becoming some of the very worst offenders in the hunt for "eyeballs". That is, they don't just find you sites that match your query anymore: now they are all clamoring to be the center of the entire web experience, by providing news, links, etc. Of course, they can't force you to use these, but they make it so convenient that a lot of people tend to just go where the portal tells them to, instead of exploring the web on their own. To me at least, all this talk of "eyeballs" is incredibly offensive, because it implicitly assumes that I'll look at whatever they show me and buy whatever they advertise. In short, it is an attempt to turn the web into television.
I think the problem with domain names is that they are logical identifiers with semantic meanings. When hosts were just identified by IP numbers, it was not contentious, but when textual names are mapped to those numbers, suddenly everyone wanted the cool names, and now that businesses and consumers have gotten involved, they also want the catchy names.
I can't think of a way to fix this except something like your directory idea, but I hope it can avoid becoming too commercialized. The thing is, people don't seem to mind the concept of a phone book -- if you want to call someone and don't have the number, you look up the name in a directory and then make the call. This directory system should be more like that.
David Gould
Last week, I saw the following quote on Ditherati:
This infuriated me, since it seems to me that that is exactly what the domain name system should not be about. All anyone seems to care about is having a URL that is catchy and easy to remember. The actual quality of the content or servce that a site provides hardly seems to matter by contrast.
I think a lot of us are against both domain squatters and domain bullies. It's not about big guys vs. little guys this time. Rather, it's because both activities are based on this horribly commercialized concept that the web is all about companies pushing and shoving to attract the all-valuable "eyeballs" to their sites.
One more thing: you said "Well, I just can't wait for the use of domain names to be deprecated." Do you know something I don't about an alternative system that will take care of these problems, or are you just expressing the same hope that I share, that some such alternative will come along eventually?
David Gould
Cironian was asking whether the MP3 decoding circuitry would significantly increase the power drain. That is, he realizes that this thing has the same mechanical and optical parts as a CD player, plus the additional chips for the decoding, and he wondered whether this addition would cause the battery life to suffer relative to a CD player.
Hrunting pointed out that, for reference, a Rio lasts longer on one battery than a CD player on two, from which we can conclude that its decoding parts take less than half the power of the CD player's mechanical and optical parts. Even less, really, since both also have the load of the actual output signal. Hence, with the addition of the decoder, the whole thing should take at most half again the power of the ordinary CD player, probably a lot less, reducing the battery life by at most a third, probably a lot less. So it was a perfectly good answer, if not fully explicit.
My addition: common sense says that a microchip should draw far less power than either a motor+laser or the output signal (though I'm just guessing, since I don't have any numbers), so to answer the original question, I would expect the extra load to be dwarfed by the others -- practically lost in the noise. I don't think it would make much difference at all.
ObDrool: I really want one of these, though I'm still pretty interested in one that would work with an IBM MicroDrive (340 MB, 1", same form factor as complact flash cards). The article on the RCA "Lyra" was the only one I saw that mentioned this possibility, but wouldn't it work with anything that accepts a compact flash card? Specifically, does anyone know if a Rio can take a MicroDrive? If not, why, and if so, why don't they advertise it?
David Gould
It looked like my karma was bumped up when I meta-moderated yesterday, but not today. Maybe it's some fraction of a point (rounded up), maybe he changed it, or maybe I hallucinated.
What I didn't get is, he said the reward for meta-moderating would never put you over the top for (I assume he meant) moderator access or the initial 2. In that case, what's it good for? Unless karma has some additional meaning, I don't get it. If this reward is not counted toward the things that karma is good for, then it doesn't serve any real purpose. Just seeing a bigger number on your profile page would be a psychological reward, but it would also cause confusion if the number stopped corresponding to the actual value used to calculate the other things.
I think having it count is fine. Moderating and meta-moderating are really "community service", since they are activities that take time and thought and aren't as rewarding as posting. This is even more true of meta-moderation, since its results are not as visible and it cannot be done as part of ordinary reading, but requires going to a special page. However, it's right that the reward should be small, since otherwise it would be too easy -- it's hard to get unearned karma by having one's posts upgraded, since someone else's decision is involved, but this would be easier to abuse.
It also wasn't quite clear (or maybe I missed it) what effect meta-moderation has. If a moderator's decisions are found to be unfair, is his access revoked, or does he lose karma (eventually making him ineligible), or what? Such punishments should not be on a hair-trigger, but it makes sense to have some such system. (What else is the point, right?)
How do these sound for karma awards:
Using all granted moderator points: +1
Since it's perfoming a service, and you had to refrain from posting in order to do it.
Meta-moderating for a particular day: +1/3
Also a service, and takes extra time, but easier to do, and hence easier to abuse, so smaller amount.
Having one's moderations judged ("fair"/"unfair"): (+/-)1/10
I assume this is the whole point of meta-moderation. For symmetry, there should be both a reward and a punishment. They should both be small to keep the total effect reasonable.
I'm not so sure about having karma expire -- the problem is that it is unfair to infrequent posters, who might always make good comments, but never enough within the time window. Of course, I recognize that the alternative greatly favors longtime members, Maybe the best would be to use the average score, or base the expiration on "the most recent n comments", instead of "the past m months." Plus, I like the idea of having it be a cumulative history of one's participation. It would be better for the threshold to be incrementally increased for older members, instead.
David Gould
For one thing, the "controls for everyone" thing would work a lot better if the points don't expire. Then, you wouldn't have to worry about your points being "wasted" if you don't moderate every day. Also, a problem is what if you moderate two comments on a single page when you only have one point? it would only be counted for (I assume) the first one, but maybe you thought the second was more important to moderate, if you had to choose, and the point would be wasted on the one that you didn't want to moderate as much. When you know exactly how many points you have, you can make more intelligent decisions as to how to make them count.
Then again, I guess the point was to encourage moderators to detach their egos from their moderation, and having no idea whether or not it would actually be counted would sort of have to have that effect.
Dunno: it's interesting, but I don't really like the idea that much. I have to agree that it's better to know when you have points. As Jburkholder said, you should think in a different mode, i.e., evaluating the quality of each post rather than just reading them for content. Ryandav alluded to the same point regarding browsing at -1, though I always do that anyway. Another reason it's really important to know when you're moderating is that you have to know to refrain from posting on a discussion so you can moderate it. It would be a shame to refrain from posting just so you could submit some votes, only to have them ignored.
David Gould
This sounds better, if I underatand correctly -- I assume you mean that when the system grants points to someone, it would also assign a subset of current stories in which those points may be used. Would the user be notified, perhaps with a "MODERATE" mark on the Slashdot front page for the selected stories? That would have the effect of evening moderation out among stories of varying popularity, and increase fairness. It would also help moderators with the "so many posts, so few points" dilemma. The problem with moderators thinking "I have to make these points count by finding some really good posts to upgrade." is that a lot of posts that deserve a 2 or 3 never get it.
That brings up an interesting question: I'll admit that I'm pretty interested in my "karma", and want to gain as much as possible, insofar as it's a measure of "acceptance" within the community, or some such warm fuzzy feeling. But is it good for people to become obsessed with this? If people start thinking "I want to get my posts moderated up, so I can become a moderator and gain some POWER; better yet, wait until I break the 20-barrier, so I can get automatic 2's! Then everyone will have to read what I write! <evil cackle>", it will become an ego thing (or more so), which could interfere with the "purity" of the discussions. Moderation already did some of that, I guess, but this would seem to increase it. On the other hand, maybe the effect is good, since it would encourage people to think more and flame less.
See the difference? Worrying about those posts that "deserve" to be upgraded only matters if we care about karma for its own sake, i.e., we consider it a reward, and think the poster should be rewarded for writing something good, even if it's not a story that many people are interested in, or if the story is already a few hours old and nobody's reading it anymore [1]. But if the purpose of moderation is only to provide an aid to readers who want to find the good posts, then it's not such a problem, since in those cases fewer people need the aid anyway.
[1] Seriously, it's amazing how fast the stories go by these days. I guess it's an inevitable result of Slashdot becoming so big, but it makes it a lot harder to participate in the discussions. If all the stories have over 300 comments by the time you get to them, all you can really do is read. If you take the time to write a really thoughtful comment, chances are hardly anyone will still be reading the story by the time you post it. This is often my problem. I'm not sure how your idea would help with this problem, but I think it's an important one.
David Gould
"I thought you said he was dead!"
"No, I said he was sleeping with the fishes..."
David Gould
Whenever anyone says, "theoretically", they actually mean, "not really".
That's true, theoretically.
David Gould