IRC log, #wikipedia channel, 2008-01-04

freenode #wikipedia

2008-01-04 23:45 < White_Cat> you can do that with </noinclude>
2008-01-04 23:57 < TimStarling> I'll break that
2008-01-04 23:57 < gwern> Krimpet: ah, but what will be the host language for the emulator?
2008-01-04 23:57 * Krimpet holds the 6502 and Z80 close to her heart
2008-01-04 23:57 < TimStarling> well, I'll make it so the comment takes precedence
2008-01-04 23:57 < quanticle> gwern: "It is obvious" == problems solvable by undergrads; "It is true" == problems solved by Grad Students; "It is interesting" == problems solved by PhDs; "A deep property" == problems that earn people the Fields Medal
2008-01-04 23:58 < White_Cat> TimStarling can I break things too?
2008-01-04 23:58 * White_Cat grabs quanticle's head
2008-01-04 23:58 < White_Cat> can I break this?
2008-01-04 23:58 < gwern> quanticle: so that's what those phrases mean
2008-01-04 23:58 * quanticle breaks White_Cat preemptively
2008-01-04 23:58 < TimStarling> Random832: I think that's possible already
2008-01-04 23:58 < quanticle> gwern: Its a handy reference guide that was given to me by a professor
2008-01-04 23:58 < theINC> i wonder what happened to the other cat guy
2008-01-04 23:58 * White_Cat cant be broken
2008-01-04 23:58 < theINC> brown_cat
2008-01-04 23:59 < gwern> actually, when it comes to wikimarkup, I've been wondering for a long time - can you link to specific page revisions as internal links and not external links?
2008-01-04 23:59 < Random832> TimStarling: {{subst:REVISIONID}} doesn't work, and was closed WONTFIX
2008-01-04 23:59 < quanticle> gwern: Its surprising how macho academic lingo gets
2008-01-04 23:59 < TimStarling> yeah?
2008-01-04 23:59 < TimStarling> hmmm
2008-01-04 23:59 < White_Cat> {{subst:TimStarling}}
2008-01-04 23:59 < White_Cat> I ought to create a TimStarling template
2008-01-04 23:59 < Krimpet> gwern: well, there's already a 6502 assembler in JavaScript that could be usable
2008-01-04 23:59 < TimStarling> you know, I actually wanted to remove {{REVISIONID}} altogether
2008-01-04 23:59 * Random832 would like to do oldid=[new-revision]&diff=last or oldid=[last-revision]&diff=next
2008-01-04 23:59 < TimStarling> maybe you should tell me what the application is
2008-01-05 00:00 < TimStarling> before I go crazy with an axe some time
2008-01-05 00:00 < Random832> in a template that can be substed
2008-01-05 00:00 < White_Cat> TimStarling why not remove all code from mediawiki at it :)
2008-01-05 00:00 < White_Cat> would be much easier to maintain
2008-01-05 00:00 < Random832> and result in a link to the diff where the edit is made
2008-01-05 00:00 < TimStarling> currently using {{REVISIONID}} will double the parse time
2008-01-05 00:00 < TimStarling> that's not cool
2008-01-05 00:00 < gwern> Random832: I'd like internal links to page revisions - it bothers me to be pasteing in secure.wikimedia.org ext links when everyone else is using en.wikipedia.org links
2008-01-05 00:00 < TimStarling> well, on save at least
2008-01-05 00:01 < Random832> well, if it's being substed, it only needs to do that to save
2008-01-05 00:01 < Random832> why would it double the parse time?
2008-01-05 00:01 < gwern> are save parsing times a big chunk of performance? I didn't think edits were that common compared to display
2008-01-05 00:01 < TimStarling> because it has to parse it once to check for spam blacklisted URLs, and again after the revision ID is known
2008-01-05 00:01 < TimStarling> you can't know the revision ID before the article is inserted
2008-01-05 00:02 < Random832> oh, i've actually been meaning to talk to a dev about the spam blacklist
2008-01-05 00:02 < TimStarling> and you can't insert the article before you've checked for spam
2008-01-05 00:02 < TimStarling> and you can't determine external links without a full parse
2008-01-05 00:02 < Random832> in addition to being irritating by preventing saves for users who are NOT inserting links, it doesn't work.
2008-01-05 00:02 < Random832> I have a foolproof exploit to enable any spam link to be inserted.
2008-01-05 00:03 < Random832> I have not told anyone [except a few admins], but it exists
2008-01-05 00:03 < Random832> why not check for spam links on _render_, anyway?
2008-01-05 00:03 < Random832> and automatically break such links by not outputing an <a href> tag
2008-01-05 00:04 < Krimpet> Random832, that sounds like it'd be more computationally intensive, since the page is rendered more often than it's saved
2008-01-05 00:04 < gwern> "When some people die, it's time to be sad. But when other people die, like really evil people, or the Irish, it's time to celebrate." Jimmy Bender, "Greg the Bunny"
2008-01-05 00:04 < quanticle> Random832: Because then you'd have to check for spam everytime the page is displayed
2008-01-05 00:04 < Random832> Krimpet: but it only has to do the check when it knows it's doing something that would insert an external link <a> tag
2008-01-05 00:04 < quanticle> gwern: Is that from the most depressing blog post ever?
2008-01-05 00:04 < TimStarling> well, item 1, feel free to report this exploit to me by private message
2008-01-05 00:04 < Random832> quanticle: no, only every time it's rendered
2008-01-05 00:04 < TimStarling> item 2, the reason we don't do it on _render_ is becuase _render_ is done before _save_
2008-01-05 00:05 < gwern> quanticle: yes
2008-01-05 00:05 < TimStarling> not after
2008-01-05 00:05 < quanticle> Random832: Isn't the page rerendered after every save?
2008-01-05 00:05 < PatPeter> I just got:
2008-01-05 00:05 < Random832> quanticle: save and purge, and when transcluded templates are updated
2008-01-05 00:05 < PatPeter> {| style="background: none;"
2008-01-05 00:05 < PatPeter> |- | valign="top" | Database locked
2008-01-05 00:05 < PatPeter> | The Wikipedia database is temporarily in read-only mode for the following reason:
2008-01-05 00:05 < PatPeter> The database has been automatically locked while the slave database servers catch up to the master
2008-01-05 00:05 < PatPeter> This is probably due to routine maintenance; if so, you will be able to edit again within a few minutes. We apologize for any inconvenience this might have caused. You can continue to browse Wikipedia articles while the database is locked. For further information, you can visit the #wikipedia channel on the freenode IRC network.
2008-01-05 00:05 < PatPeter> |}
2008-01-05 00:06 < gwern> what if on detecting a spam link on rendering, you save a new revision breaking the link? but then you'd need to whitelist the new revision to reap any performance benefit... oy
2008-01-05 00:06 < Random832> gwern: you can't "save a new revision breaking the link" because breaking the link is an arbitrarily difficult task
2008-01-05 00:07 < Krimpet> ...PatPeter is banned -_-
2008-01-05 00:07 < TimStarling> I'll tell you what: I'll fix it when someone starts spamming with it, ok?
2008-01-05 00:07 < TimStarling> spammers are really really stupid, I don't know if you've noticed this
2008-01-05 00:07 < TimStarling> relying on spammer stupidity will get you a long way
2008-01-05 00:08 < Random832> ok, ok - but, my suggestion would also stop annoying legitimate users
2008-01-05 00:08 < TimStarling> you mean when the spam blacklist is changed and the next user to attempt to edit the article has their edit rejected?
2008-01-05 00:09 < Random832> well, that's one case.
2008-01-05 00:09 < Random832> the other is the exploit i suggested, and less sophisticated variations of it that also work
2008-01-05 00:10 < TimStarling> well, the issue is that we don't really want spam in the article, even if it is disabled spam
2008-01-05 00:10 < TimStarling> that's why it works the way it does
2008-01-05 00:10 < Random832> eh
2008-01-05 00:10 < Random832> it should at least ignore section editing - the user doesn't, in that case, have the ability to easily go through and see where the spam is
2008-01-05 00:11 < Random832> [section editing where the spam is in another section]
2008-01-05 00:11 < TimStarling> ok
2008-01-05 00:11 < Random832> (and, I heard recently that the rejection page doesn't have an edit box on it? that should change, if that's the case)
2008-01-05 00:11 < TimStarling> anyway, I think we were talking about {{REVISIONID}} a while ago, a subject which is rather closer to my present work
2008-01-05 00:12 < Random832> yeah, that
2008-01-05 00:12 < gwern> 'Mr Sterne has also contacted veterans of the US 5th Ranger Battalion, who confirmed that they took Maisy Battery from the Germans. They also revealed that they found $4.2?million worth of French francs, which was shared among the men.' <-- I guess even Rangers have their price
2008-01-05 00:12 < TimStarling> {{subst:REVISIONID}} doesn't work because the revision ID is not known during PST
2008-01-05 00:12 < TimStarling> that should be no big suprise
2008-01-05 00:12 < Random832> some weeks ago, i went to create a template i'd had in mind for a while, which would allow you to [subst the template] provide a link to the diff for the edit you are making
2008-01-05 00:12 < Random832> ok... what about the page's _current_ revision id?
2008-01-05 00:13 < Random832> it wouldn't be ideal, but my idea would still work
2008-01-05 00:13 < Random832> and it would make a certain kind of sense - you could even display the same for {{REVISIONID}} in preview
2008-01-05 00:13 < TimStarling> why would you want a diff link to the edit you're making?
2008-01-05 00:14 < Random832> well, the use case i had in mind would be if i'm removing or refactoring something on the talk page, and want to make a note
2008-01-05 00:14 < quanticle> gwern: Where is that from?
2008-01-05 00:15 < Random832> it's not that important, i was just wondering how hard would it be to do
2008-01-05 00:16 < gwern> quanticle: reddit article on re-discovery of a nazi gun battery in normandy
2008-01-05 00:18 < TimStarling> make a note in the next edit
2008-01-05 00:18 < TimStarling> it really makes life a whole lot simpler
2008-01-05 00:18 < Random832> all right
2008-01-05 00:20 < Random832> we've spent too much time on this i guess
2008-01-05 00:21 < max_> god evening
2008-01-05 00:21 < max_> .. or good, whichever
2008-01-05 00:24 < max_> I am trying to get a copy of the wikimedia dump of enwiki and I'm confused as to which file I'm looking for
2008-01-05 00:25 < FastLizard4> max_: Not sure, in fact, I thought that the dumps were not accessible to the public, but I may be wrong
2008-01-05 00:26 < gwern> max_: it really depends on what you actually want
2008-01-05 00:26 < gwern> FastLizard4: check out WP:DUMP, man. they are very publicly accessible
2008-01-05 00:27 < max_> faslizard, download.wikimedia.org
2008-01-05 00:27 < max_> what kind of copyleft encyclopedia would it be if you couldn't copy it
2008-01-05 00:27 < FastLizard4> huh
2008-01-05 00:28 < FastLizard4> From reading the page, you probably want pages-articles.xml.bz2
2008-01-05 00:28 < max_> allright, I think I've found what i've been looking for acutally
2008-01-05 00:29 < max_> download.wikimedia.org/enwiki/20071018/pages-meta-current.xml.bz2
2008-01-05 00:31 < max_> gwern, my aim is to have an offline accessibe wikpedia to give to someone for christmas :D
2008-01-05 00:32 < gwern> max_: ok... so I guess then current-articles (main, current revisions only) sounds like what you want?
2008-01-05 00:32 < gwern> and, uh, wasn't christmas a little while ago?
2008-01-05 00:32 < gwern> more to the point, I wouldn't be going that solo. plenty of people have worked on doing that before. why not reuse their work?
2008-01-05 00:33 < Tony_Sidaway> max_: as I understand it there have been offline Wikipedias. There used to be a page (and perhaps it still exists) about how to get a dump of Wikipedia onto your PDA.
2008-01-05 00:33 < pengo> why is it i see a lot more links to wikipedia in channels other than this one
2008-01-05 00:33 < max_> I would feel better to acquire the dump from a torrent
2008-01-05 00:34 < Tony_Sidaway> pengo: cos we've been there, done that ;)
2008-01-05 00:34 < max_> and if someone already created a viewer for it, then that's great
2008-01-05 00:34 < Tony_Sidaway> max_: I don't even know what a torrent is. What's wrong with ftp?
2008-01-05 00:35 < max_> nothing is wrong, in a torrent the bandwidth for downloading is drawn mainly from other people that have the file
2008-01-05 00:35 < gwern> Tony_Sidaway: it's a weird ass ancient insecure protocol with a ridiculous number of corner cases
2008-01-05 00:35 < Tony_Sidaway> gwern: it's a commonly used download protocol.
2008-01-05 00:35 < gwern> and telnet used to be a common login protocol
2008-01-05 00:35 < Tony_Sidaway> Who cares if ftp is insecure?
2008-01-05 00:36 < max_> hey, no fite! :D
2008-01-05 00:36 < gwern> anyone who cares about security?
2008-01-05 00:36 < quanticle> max_: Torrents reduce the load on Wikipedia servers
2008-01-05 00:36 < max_> absolutely
2008-01-05 00:37 < Tony_Sidaway> gwern: relying on the transmission medium to ensure security? Not very clever.
2008-01-05 00:37 < max_> I've searched for third part wiki dumps and torrents, the wiki site itself does not use torrents
2008-01-05 00:37 < TimStarling> what's the security problem with anonymous FTP?
2008-01-05 00:37 < Tony_Sidaway> Ah torrent is some kind of peer-to-peer thngie? And this is MORE secure? :0
2008-01-05 00:37 < max_> do you guys have any suggestions as for already published offline wiki dumps?
2008-01-05 00:38 < gwern> Tony_Sidaway: transport layer security has its place.
2008-01-05 00:38 < max_> Tony_Sidaway: it makes extensive use of encryption and hashing to assure that the peers canot alter the data
2008-01-05 00:38 < Tony_Sidaway> max_: indeed, and I bet it could be implemented on top of ftp ;)
2008-01-05 00:38 < gwern> TimStarling: bounce scanning, for starters
2008-01-05 00:38 < LuciferTiger> can i ask one really annoying question?
2008-01-05 00:39 < gwern> breaking the ftp server, for another
2008-01-05 00:39 < gwern> LuciferTiger: that was it. thanks for playing
2008-01-05 00:39 < LuciferTiger> lol okie
2008-01-05 00:39 < max_> naw, shoot, LuciferTiger
2008-01-05 00:39 < max_> but don't ask to ask :P
2008-01-05 00:40 < gwern> ([[FTP bounce attack]] still astounds me)
2008-01-05 00:40 < White_Cat> STD Lucifer?
2008-01-05 00:40 < White_Cat> That cant be right?
2008-01-05 00:40 < LuciferTiger> any chan other than wikipedia on here where i could talk about comp stuff?
2008-01-05 00:40 < White_Cat> I blew that the smethers
2008-01-05 00:40 < max_> gwern, I'm looking that one up, sounds interesting
2008-01-05 00:40 < White_Cat> You shivan scum
2008-01-05 00:40 < LuciferTiger> i hit on wikipedia just randomly
2008-01-05 00:40 * White_Cat slaps LuciferTiger a bit with a large menos grande...
2008-01-05 00:40 < White_Cat> :D
2008-01-05 00:40 < gwern> max_: I've used nmap's implementation on occasion. it's useful sometimes
2008-01-05 00:41 < LuciferTiger> that hurt... :P
2008-01-05 00:41 < White_Cat> Sorry Lucifer was a capital ship in FreeSpace
2008-01-05 00:41 < max_> LuciferTiger, this is freenode, people are assholes, just absorb more knowlege than annoying stuff you spew out and soon, you will find cool peoples with similar interests
2008-01-05 00:41 < LuciferTiger> oh im more like the light bearer kind of lucifer
2008-01-05 00:42 < LuciferTiger> ahh thanks
2008-01-05 00:42 < max_> on efnet there's like the electronics channel for electronics, c++ people are cool, ops in help are cool once you get to know them, compsci is where real smart people talk about computers and math
2008-01-05 00:42 < White_Cat> talking about nodes
2008-01-05 00:42 < White_Cat> Lucifer is destroyed in the Sol node
2008-01-05 00:43 < Tony_Sidaway> gwern: that is an extremely old exploit. And yet ftp is still in very wide use without significant problems. How do you explain this?
2008-01-05 00:43 < max_> ftp is useful
2008-01-05 00:44 < max_> and convenient
2008-01-05 00:44 < pengo> http has less overhead
2008-01-05 00:44 < pengo> and cheese tastes better
2008-01-05 00:45 < pengo> and torrential rains are good for rice farmers
2008-01-05 00:45 < Tony_Sidaway> http is okay too. But it isn't secure either.
2008-01-05 00:45 < pengo> what are we talking about?
2008-01-05 00:45 < gwern> Tony_Sidaway: C is still in very wide use despite buffer overflows being an extremely old class of exploits which keeps showing up. How do you explain this?
2008-01-05 00:46 < Tony_Sidaway> gwern: you just answered the question.
2008-01-05 00:46 < max_> why do buffer overflows keep showing up!
2008-01-05 00:46 < Tony_Sidaway> quite!
2008-01-05 00:46 < gwern> max_: people keep using non-type or memory safe languages, is the short answer
2008-01-05 00:47 < pengo> i'm finding this conversation very boring so i'm oging to mention Britney Spears locked up in psych ward
2008-01-05 00:47 < Tony_Sidaway> C in the right hands is an extremely powerful tool, as another answer.
2008-01-05 00:47 < pengo> despite my better judgement
2008-01-05 00:48 < max_> can't we go all proof style mathematical on it and make algorithms that can't overflow?
2008-01-05 00:48 < FastLizard4> max_: Did you get it downloaded eventually?
2008-01-05 00:49 < pengo> max_: sure.. and we can spend six years writing hello world too,
2008-01-05 00:49 < gwern> max_: sure. good luck with that, btw. academics have been trying to crack that one for decades, and haven't managed to make C really tractable
2008-01-05 00:49 < Tony_Sidaway> max_: yes but there aren't that many people prepared to program low-level stuff in high level languages that are susceptible to proof.
2008-01-05 00:50 < max_> fascinating
2008-01-05 00:50 < White_Cat> Tony_Sidaway you recall my PM?
2008-01-05 00:50 < gwern> they exist, though. I like the various kernels written in Haskell as examples of that sort of thing
2008-01-05 00:51 < Tony_Sidaway> White_Cat: I listened for a bit but had to leave with no response after a bit.
2008-01-05 00:51 < Tony_Sidaway> I had asked you to link me
2008-01-05 00:51 < White_Cat> Tony_Sidaway that didnt come through
2008-01-05 00:51 < White_Cat> I will link you now
2008-01-05 00:51 < quanticle> gwern: C is fast. Its "portable assembly", as one professor put it...
2008-01-05 00:51 < FastLizard4> oh dear, assembly programming...
2008-01-05 00:52 < Tony_Sidaway> It is, pretty much, portable assembler. And not that portable when you get to word sizes.
2008-01-05 00:52 < Tony_Sidaway> Not to mention endianism
2008-01-05 00:52 < gwern> quanticle: autoconf begs to differ about the portable aspect... and as other folks never tire of pointing out, you can get C speeds or better in other higher-level languages. its low-level nature breaks a lot of optimizations
2008-01-05 00:53 < gwern> I don't particularly care about speed though, but I do believe them. at worst, you can always use your language's FFI to fall back on C
2008-01-05 00:54 < quanticle> gwern: I'm not talking about portablility in that respect. I'm saying that there's a C compiler for damn near every architecture out there.
2008-01-05 00:54 * gwern remembers reading one fellow's thesis. the trouble he had getting any form of tail recursion into gcc becuase of C!
2008-01-05 00:55 < gwern> quanticle: are you expecting to run your program on nearly every architecture out there? you'd better target netbsd or linux then also
2008-01-05 00:56 < max_> I do not know if I even want to know what it is that allows code to work across architechtures and operating systems
2008-01-05 00:57 < pengo> wow this channel is boring today
2008-01-05 00:57 * LuciferTiger remembers someone bootstrapping C compiler to insert backdoors into c login code without it actually being in the code.
2008-01-05 00:57 < quanticle> gwern: Also, while C may not be especially fast on PC architectures, it is a hell of a lot faster than anything else on mobile devices or embedded systems
2008-01-05 00:57 < LuciferTiger> all this so that his boss woudnt see that he put a backdoor there
2008-01-05 00:57 < quanticle> LuciferTiger: Yeah. That's actually a rather famous example of how you have to have trusted toolchains
2008-01-05 00:58 < gwern> when I first read of thompson's backdoor, it blew my mind
2008-01-05 00:58 < LuciferTiger> gwern: lol same here
2008-01-05 00:59 < quanticle> LuciferTiger: http://www.ece.cmu.edu/~ganger/712.fall02/papers/p761-thompson.pdf
2008-01-05 00:59 < quanticle> Heh. I still have that bookmarked...
2008-01-05 00:59 < LuciferTiger> wow i was about to post a similar link
2008-01-05 01:00 < LuciferTiger> all my "bookmarks" are on stumbleupon which when i last checked was rather unsearchable
2008-01-05 01:00 < gwern> I was impressed by how to detect such an evil compiler - use a differently evil compiler!
2008-01-05 01:00 < FastLizard4> storm worm...
2008-01-05 01:01 < max_> hmm
2008-01-05 01:02 < quanticle> FastLizard4 ...uses a completely different vulnerability: human nature
2008-01-05 01:02 < max_> proving in general that code is not compromised is still not within the scope of automated proof systems
2008-01-05 01:02 < FastLizard4> quanticle: ?
2008-01-05 01:03 < quanticle> FastLizard4: I was finishing your statement
2008-01-05 01:03 < FastLizard4> oh, yes
2008-01-05 01:04 < FastLizard4> Wow, it is pouring rain outside of my house
2008-01-05 01:04 < quanticle> max_: In fact, some vulnerabilities are impossible to disprove, since they're not halting problems. Buffer overflow is one of them...
2008-01-05 01:13 < pengo> only in theory
2008-01-05 01:13 < TimStarling> that's a lot of negatives...
2008-01-05 01:14 < TimStarling> the halting problem does not prevent vulnerability detection
2008-01-05 01:14 < TimStarling> if that's what quanticle was trying to say
2008-01-05 01:15 < TimStarling> because in vulnerability detection, you can discard suspicious code and demand that the author rewrite it
2008-01-05 01:15 < TimStarling> you can just reject any code that is too hard to analyse
2008-01-05 01:15 < TimStarling> that's how you get around the problem of uncomputability
2008-01-05 01:17 < Random832> you can demand that the author rewrite it or what?
2008-01-05 01:17 < LuciferTiger> ofcourse u an
2008-01-05 01:18 * Random832 goes into #mediawiki and demands that you rewrite some part of the wiki engine
2008-01-05 01:18 < LuciferTiger> thats the wole point of it
2008-01-05 01:18 < LuciferTiger> Random832: if you prove somethings wrong and if u have a stake in it... you can.
2008-01-05 01:21 < gwern> ah, one of the premises of static typing. 'it's ok to disallow some correct programs as long as we can be certain our type checking procedure terminates and will not allow an illegal program'
2008-01-05 01:21 < naught101> http://en.wikipedia.org/wiki/Image:Carbon-dioxide-3D-vdW.svg#filehistory
2008-01-05 01:21 < TimStarling> a human reviewer should only accept code that is obviously secure
2008-01-05 01:21 < naught101> that image is vandalised somehow...
2008-01-05 01:21 < naught101> I can't see how to fix it
2008-01-05 01:21 < TimStarling> computerised reviewers could do the same
2008-01-05 01:21 < naught101> it's mildly amusing
2008-01-05 01:21 < Random832> gwern: that's not one of the premises of static typing. a type system isn't required to be turing-complete in the first place
2008-01-05 01:21 < naught101> and also quite offensive
2008-01-05 01:22 < naught101> ah... commons
2008-01-05 01:23 < LuciferTiger> whats amusing is the blatant lack of spelling skills
2008-01-05 01:23 < gwern> Random832: what's the point of a type checker which might not terminate?
2008-01-05 01:24 < Random832> eh?
2008-01-05 01:24 < Random832> I'm not sure what we're arguing about
2008-01-05 01:24 < naught101> true LuciferTiger
2008-01-05 01:24 < Random832> what are the "some correct programs" that are being discarded in order to make it certain that the type checker will terminate
2008-01-05 01:25 < Random832> I don't think there is any way for there to be a construction which creates a risk of it not terminating, if the type system isn't turing complete / isn't able to express recursion/loops/whatever in the first place
2008-01-05 01:26 < gwern> Random832: rice's theorem strikes again. you can't have a type checker which will always terminate but which will always disallow all programs which could go wrong at runtime and only such programs
2008-01-05 01:27 < Random832> gwern: um, how the hell do you figure
2008-01-05 01:27 < gwern> you could lift into the type system any computation you want - such as whether a given program will terminate.
2008-01-05 01:27 < Random832> um... wtf?
2008-01-05 01:27 < Random832> only if you have a type system a hell of a lot more complex than integer, string, array, struct in the first place
2008-01-05 01:27 < Random832> if you don't, you can't lift any computation into the type system anyway
2008-01-05 01:28 < gwern> Random832: how do you figure you can have your magical type checker, then?
2008-01-05 01:28 < Random832> ...huh?
2008-01-05 01:28 < Random832> by NOT HAVING TYPES THAT ARE ABLE TO EMBODY COMPUTATION
2008-01-05 01:28 < TimStarling> well, C++ types are pretty damn complicated
2008-01-05 01:28 < Random832> if it's not turing complete, you don't even have the issue of there being a risk of it not halting at all
2008-01-05 01:29 < TimStarling> they've been used to do quite a lot of compilation-time calculation
2008-01-05 01:29 < Random832> TimStarling: I rephrase myself: "by not using C++"
2008-01-05 01:29 < gwern> oh. you're talking about a really primitive type system then. I forget where those fall in. probably in disallowing valid programs
2008-01-05 01:29 < LuciferTiger> all this automata stuff is making my head spin.
2008-01-05 01:29 < TimStarling> but nobody actually cares about type checking
2008-01-05 01:30 < TimStarling> gwern was just using it as an analogy
2008-01-05 01:30 * FastLizard4 rejects your reality and substitutes his own
2008-01-05 01:30 < gwern> TimStarling: yeah, but I'm not sure how much you can do on the type level. all the compile-time stuff I understood to be template-based; doing type-level calculations in say Haskell does end up doing the work at compile time but I think that's supposed to be an implementation detail
2008-01-05 01:30 < TimStarling> the conversation (some time ago) was about security
2008-01-05 01:30 < Random832> no, "disallowing valid programs" is when the C++ compiler decides to quit after 15 layers of templates
2008-01-05 01:32 < FastLizard4> C++ is evil
2008-01-05 01:32 < gwern> TimStarling: mm. types are in a sense security, in the same way proofs = programs
2008-01-05 01:34 < gwern> iirc, types are inter-convertible with capabilities, which are themselves as powerful as ACLs
2008-01-05 01:37 < TimStarling> so... MZMcBride tells me in #mediawiki that <onlyinclude> is used sometimes
2008-01-05 01:37 < TimStarling> and I think I've found a way to make it kind of work in the new preprocessor
2008-01-05 01:37 < Random832> <onlyinclude>?
2008-01-05 01:38 < Random832> not <includeonly>?
2008-01-05 01:38 < FastLizard4> <onlyinclude>? That's a new one
2008-01-05 01:38 < TimStarling> yeah, was that conversation before your time?
2008-01-05 01:38 < Random832> what's <onlyinclude>?
2008-01-05 01:38 < FastLizard4> It's <includeonly>, Random832 is right
2008-01-05 01:38 < TimStarling> 2 hours ago now
2008-01-05 01:38 < Random832> if it's for something other than <includeonly>, that's confusing and should be renamed
2008-01-05 01:38 < TimStarling> 2 #wikipedia years, right?
2008-01-05 01:39 < TimStarling> well, it was avar's crazy idea
2008-01-05 01:39 < Random832> actually - what new preprocessor is this?
2008-01-05 01:39 < TimStarling> as it happens, I am wearing a T-shirt today with another of avar's crazy ideas written across it
2008-01-05 01:40 < Random832> what's that?
2008-01-05 01:40 < Messedrocker> Ævar you mean?
2008-01-05 01:40 < TimStarling> my "diff to penultimate revision" T-shirt from WM 2007
2008-01-05 01:40 < Messedrocker> oh boy
2008-01-05 01:40 < Messedrocker> what a stupid t-shirt
2008-01-05 01:40 < Messedrocker> stupid stupid stupid
2008-01-05 01:40 < Messedrocker> also your girlfriend's laptop is cute
2008-01-05 01:40 < Random832> "diff to penultimate revision"?
2008-01-05 01:41 < Messedrocker> Random832, at Wikimania 2007, that phrase was printed on a bunch of t-shirts everyone wore
2008-01-05 01:41 < Messedrocker> and i thought it was really stupid
2008-01-05 01:41 < Random832> you're going to have to explain more clearly in simple english what that is and why it's a crazy idea
2008-01-05 01:41 < Random832> oh, the t-shirt _itself_ is a crazy ide
2008-01-05 01:41 < Random832> a
2008-01-05 01:41 < Scott5114> that's a default mediawiki string
2008-01-05 01:41 < TimStarling> nah, we took it out of mediawiki a while ago
2008-01-05 01:42 < TimStarling> I said that to Ward and he said "I know, but we're preserving it for posterity"
2008-01-05 01:42 < Random832> speaking of wikimania, is there going to be a vote on what location for 2009 or what? I heard toronto was one of the possibilities; that's drivable for me
2008-01-05 01:42 < Random832> apparently it's part of the new messages box?
2008-01-05 01:42 < TimStarling> there will be a jury, like every year
2008-01-05 01:42 < Scott5114> I think on Wikipedia it was changed to "last diff" or something
2008-01-05 01:42 < Random832> Scott5114: "last change"
2008-01-05 01:43 < gwern> why penultimate?
2008-01-05 01:43 < Messedrocker> gwern because it best describes the edit
2008-01-05 01:43 < gwern> isn't the revision before the ultimate (final) revision?
2008-01-05 01:43 < gwern> *isn't that
2008-01-05 01:43 < Messedrocker> "diff to penultimate revision" best describes it
2008-01-05 01:43 < Messedrocker> except it's waaaay too wordy
2008-01-05 01:43 < Random832> because it's a diff of between your talk page and the revision before the latest version of it (i.e. the penultimate one)
2008-01-05 01:44 < Random832> gwern: right. and what revision is on the left side when you click the link?
2008-01-05 01:44 < The359> penultimate version would make more sense to me
2008-01-05 01:44 < The359> Penultimate revision would be the second to last change
2008-01-05 01:44 * gwern doesn't understand why 'last diff' isn't a diff between the penultimate and ultimate revisions
2008-01-05 01:44 < Scott5114> I think that's what it is
2008-01-05 01:44 < TimStarling> you know what #wikipedia is?
2008-01-05 01:45 < Random832> gwern: it _is_. it's a diff between the "penultimate revision" on the one hand, and the current text of the page on the other hand
2008-01-05 01:45 < TimStarling> a crazy tangent generation engine
2008-01-05 01:45 < Scott5114> TimStarling: A mental institution? :P
2008-01-05 01:45 < Random832> lol
2008-01-05 01:45 < Random832> anyway
2008-01-05 01:45 < Random832> what were you saying, about a new preprocessor?
2008-01-05 01:45 < gwern> oy. we have some confusing terminology here
2008-01-05 01:45 < TimStarling> yeah, well that's been announced a couple of times
2008-01-05 01:46 < TimStarling> not much difference for the regular user except that it's faster
2008-01-05 01:46 < TimStarling> we're trying very hard to make all the articles render in exactly the same way
2008-01-05 01:46 * Random832 wouldn't worry about the articles
2008-01-05 01:46 * Random832 would be more worried about the templates
2008-01-05 01:46 < TimStarling> articles have templates on them
2008-01-05 01:46 < pengo> TimStarling: will <ref>'s be able to be used as template parameters now?
2008-01-05 01:46 < TimStarling> that's what makes rendering articles difficult
2008-01-05 01:47 < pengo> TimStarling: umm.. i mean within templates
2008-01-05 01:47 < Random832> will <noi<includeonly>ncl</includeonly>ude> work?
2008-01-05 01:47 * Random832 is kidding; that doesn't work now
2008-01-05 01:47 < TimStarling> pengo: maybe I will have time to add a few features after I'm done refactoring, maybe not
2008-01-05 01:47 < Random832> TimStarling: well, it's not really a feature so much as it is a bug.
2008-01-05 01:48 < Random832> the ref/template thing
2008-01-05 01:48 < Random832> (which incidentally seems to work fine some places)
2008-01-05 01:48 < pengo> TimStarling: it's not really so much a feature as an issue of rending order
2008-01-05 01:49 < Random832> rending order?
2008-01-05 01:49 < Random832> the order in which pages are torn apart?
2008-01-05 01:49 < Random832> that's actually - not a bad word to describe it
2008-01-05 01:49 < TimStarling> feature, bug, whatever
2008-01-05 01:50 < TimStarling> I'm not making any promises besides "it'll be as close to what we have already as possible"
2008-01-05 01:50 < Random832> TimStarling: just don't try to hard to make it continue to fail on refs in templates
2008-01-05 01:50 < Messedrocker> Single-user login exists
2008-01-05 01:50 < Messedrocker> I saw it at Wikimania
2008-01-05 01:51 < Messedrocker> TimStarling, I also noticed you use Windows Vista
2008-01-05 01:51 < TimStarling> came with the laptop
2008-01-05 01:51 < TimStarling> no endorsement implied
2008-01-05 01:51 < pengo> TimStarling: so it's purely a speed thing?
2008-01-05 01:51 < Messedrocker> Ah, but if you were a true martyr for the Civil Rights for Software Movement, you'd cast off the shackles of the evil Micro$hit 666 Corporation and bless it with the love of open-source software
2008-01-05 01:52 < Messedrocker> </troll>
2008-01-05 01:52 < TimStarling> pengo: it will hopefully make it easier to fix bugs and add more features
2008-01-05 01:53 < pengo> TimStarling: ah k cool
2008-01-05 01:53 < TimStarling> I'm fixing bug 4899, that will hopefully go away altogether
2008-01-05 01:55 < pengo> i'm finding it surprisingly difficult to find the bug tracker
2008-01-05 01:56 < TimStarling> http://bugzilla.wikimedia.org/4899
2008-01-05 01:57 < TimStarling> also bug 5678 will go away
2008-01-05 01:58 < pengo> ta
2008-01-05 01:58 < TimStarling> I noticed that when someone re-reported it
2008-01-05 01:58 * FastLizard4 never understood bugzilla
2008-01-05 01:59 < pengo> TimStarling: cool. so fixing some real bugs.
2008-01-05 02:00 < TimStarling> <ref> really needs a rewrite
2008-01-05 02:00 < TimStarling> maybe it can be brought into the core
2008-01-05 02:01 < FastLizard4> Good night, people
2008-01-05 02:02 < TimStarling> everyone knows <ref> sucks in a lot of ways, and it's such a critical component in wikipedia that it deserves a bit of developer time
2008-01-05 02:05 < mazzanet> ha
2008-01-05 02:06 < mazzanet> i {{fact}}ed a page a couple days ago
2008-01-05 02:07 < mazzanet> someone then citated it with a link to the flash intro of a particular website
2008-01-05 02:08 < mazzanet> and i just found a credible source opposing their citation
2008-01-05 02:08 < mazzanet> *chuckles*
2008-01-05 02:10 < TimStarling> adding a citation database, that'd be pretty cool, eh?
2008-01-05 02:10 < TimStarling> bibtex style?
2008-01-05 02:11 < pengo> TimStarling: yup. <ref> really needs a version that lets you define a ref without putting in that particular spot too.
2008-01-05 02:11 < TimStarling> obviously {{cite web|...}} would be better than <ref>{{cite web|...}}</ref>
2008-01-05 02:12 < pengo> <ref showRightHere="false" name="cheese"> i like cheese.<ref name="cheese"/>
2008-01-05 02:12 < TimStarling> :P
2008-01-05 02:13 < pengo> i think refs scare people from editing too much when they're long and embedded in the text
2008-01-05 02:13 < TimStarling> it's half a solution at best
2008-01-05 02:14 < TimStarling> I made a proposal for harvard-style referencing on mediawiki-l
2008-01-05 02:14 < pengo> aren't there templates for that already?
2008-01-05 02:14 < TimStarling> shall I find the archive?
2008-01-05 02:14 < pengo> ok
2008-01-05 02:14 < pengo> or i could just read [[Harvard referencing]] and [[Wikipedia:Harvard referencing]]
2008-01-05 02:15 < TimStarling> well, that won't tell you much
2008-01-05 02:16 < TimStarling> http://lists.wikimedia.org/pipermail/mediawiki-l/2007-December/025107.html
2008-01-05 02:16 < pengo> ahh nice
2008-01-05 02:17 < pengo> tho it creates even more wiki syntax
2008-01-05 02:17 < TimStarling> using punctuation instead of XML makes it easier on the eye
2008-01-05 02:17 < pengo> nods
2008-01-05 02:17 < TimStarling> I think we can handle a bit more wiki syntax
2008-01-05 02:18 < pengo> TimStarling: well you're the one writing the parser, so I won't complain :)
2008-01-05 02:19 < pengo> will there ((citation needed)) too, or is that just islly?
2008-01-05 02:20 < TimStarling> what would it do?
2008-01-05 02:20 < pengo> TimStarling: the same as {{fact}}
2008-01-05 02:20 < pengo> so pointless really
2008-01-05 02:21 < TimStarling> link to [[Wikipedia:Citation needed]]?
2008-01-05 02:21 < pengo> yup
2008-01-05 02:21 < TimStarling> well, that's probably best done by a template
2008-01-05 02:22 < TimStarling> now if it popped up a box and allowed you to enter a reference, that would justify a mediawiki special case
2008-01-05 02:22 < pengo> TimStarling: have you thought about splitting up the category lists at the bottom of articles so that "real" categories (e.g "Birds of Australia") are listed separately from metadata categories (like "References needed since January 2008")
2008-01-05 02:22 < TimStarling> I have thought about hiding the metadata categories altogether
2008-01-05 02:22 < pengo> that would work too
2008-01-05 02:23 < White_Cat> TimStarling
2008-01-05 02:23 < White_Cat> is it possible to better sort categories
2008-01-05 02:23 < pengo> TimStarling: have you thought about hiding all stub notices too? they annoy me :)
2008-01-05 02:23 < White_Cat> they are rather unintelligab;e
2008-01-05 02:24 < White_Cat> pengo you can do that with your monobook
2008-01-05 02:24 < pengo> White_Cat: really?
2008-01-05 02:24 < White_Cat> stub types are clearly marked as stub on the template itself
2008-01-05 02:24 < White_Cat> you can give it a "hidden" style
2008-01-05 02:24 < TimStarling> I've thought about many things
2008-01-05 02:24 < White_Cat> TimStarling thinks
2008-01-05 02:24 < White_Cat> too much
2008-01-05 02:24 < TimStarling> unfortunately I don't have time to do very many of them
2008-01-05 02:25 < TimStarling> I need some underlings
2008-01-05 02:25 < White_Cat> SQLDb
2008-01-05 02:25 < White_Cat> talk to me
2008-01-05 02:26 < pengo> ;i can help you think.
2008-01-05 02:26 < Lubaf> White_Cat: What's new in the world of POV pushing?
2008-01-05 02:27 < Lubaf> I ask not because I think you're a resident, but rather, because I think you observe said world closely.
2008-01-05 02:27 < White_Cat> Lubaf be careful with the next word you say. It may be your last as the king of sparta!
2008-01-05 02:27 < White_Cat> This is MADNESS!
2008-01-05 02:27 < White_Cat> This is SPARTAAAAA!
2008-01-05 02:27 < White_Cat> </300 parody>
2008-01-05 02:28 < White_Cat> Lubaf I am more up to date on the world of Pokemon than POV
2008-01-05 02:28 < White_Cat> both start with "po"
2008-01-05 02:29 < White_Cat> Lubaf what do you want to know?
2008-01-05 02:29 < Lubaf> White_Cat: Nothing new about those idiots in Kurdistan articles?
2008-01-05 02:29 < Lubaf> And episode arguments are dead at the moment?
2008-01-05 02:29 < White_Cat> I dont pay attention to Kurdistan articles
2008-01-05 02:29 < White_Cat> and there hasnt been many edits on those
2008-01-05 02:29 < White_Cat> people just found it boring
2008-01-05 02:29 < White_Cat> pov pushing w/o anyone getting in the way
2008-01-05 02:29 < White_Cat> they are busy in de.wiki
2008-01-05 02:30 < White_Cat> episode thing isnt dead
2008-01-05 02:30 < White_Cat> its just building up slowly
2008-01-05 02:30 < White_Cat> it will explode any moment
2008-01-05 02:30 < Lubaf> Oh, goodie. My power may go out soon.
2008-01-05 02:30 < White_Cat> and it will be a big mess
2008-01-05 02:30 < White_Cat> Lubaf oh no
2008-01-05 02:30 < White_Cat> episode thing wont blow THAT bad
2008-01-05 02:30 < Lubaf> Wind picked up, and that means that lamers who let their trees grow into powerlines are about to wreak havok.
2008-01-05 02:31 < Lubaf> No, it died down.
2008-01-05 02:31 < White_Cat> it did not
2008-01-05 02:31 < Lubaf> White_Cat: The wind, not the episode flame wars.
2008-01-05 02:31 < White_Cat> aw
2008-01-05 02:31 < White_Cat> :D
2008-01-05 02:32 < White_Cat> Lubaf I am currently being lynched for including sources on an article
2008-01-05 02:32 < White_Cat> turkish governmental sources mind you
2008-01-05 02:32 < Lubaf> Ah. Those POV pushers.
2008-01-05 02:32 < White_Cat> You know... The Nazi Government wanaby country...
2008-01-05 02:32 < White_Cat> Lubaf oh it isnt just pov pushers
2008-01-05 02:32 < Lubaf> Have they tried to get you blocked yet?
2008-01-05 02:32 < White_Cat> a decent number of admins are also lycnching me
2008-01-05 02:32 < White_Cat> I will be blocked soon enough
2008-01-05 02:33 < TimStarling> I hope not literally lynching
2008-01-05 02:33 < TimStarling> that would be something of an emergency
2008-01-05 02:33 < TimStarling> and very impressive that you would be able to type on IRC during it
2008-01-05 02:34 < Lubaf> TIMMAH!
2008-01-05 02:34 < Lubaf> TIMMAH! LIVIN A LIE!
2008-01-05 02:34 * Lubaf starts playing the guitar solo here.
2008-01-05 02:37 < White_Cat> TimStarling a literal lynch would not be possible
2008-01-05 02:37 < White_Cat> getting to me requires more than a large mob
2008-01-05 02:41 < Lubaf> You can't just improvise a lynching; you need organization.
2008-01-05 02:41 < Lubaf> Who's going to bring the rope?
2008-01-05 02:41 < Lubaf> Who's going to bring spare torches?
2008-01-05 02:42 < Lubaf> Which tree are we going to use?
2008-01-05 02:46 < Lubaf> Seriously.
2008-01-05 02:46 < Lubaf> Is anybody here?
2008-01-05 02:47 < TimStarling> we're laughing on the inside
2008-01-05 02:50 < pengo> my insides just ordered pizza
2008-01-05 02:53 < jtizzle1854> does anyone know where I can find information about the various movie rating systems of movies of the different countries of the world? I found it on wikipedia but it doesn't contain the country I want info for I want info for ratings of russia and ukraine
2008-01-05 02:54 < The359> I'd take a stab at us not having an article on it
2008-01-05 02:54 < jtizzle1854> huh?
2008-01-05 02:55 < The359> we most likely don't have that information on Wikipedia
2008-01-05 02:55 < pengo> i think we do
2008-01-05 02:56 < pengo> try [[R-rating]] and links from there
2008-01-05 02:56 < White_Cat> Lubaf
2008-01-05 02:56 < jtizzle1854> i know
2008-01-05 02:56 < White_Cat> HAH
2008-01-05 02:56 < jtizzle1854> you don't that's why i asked y'all
2008-01-05 02:56 * SQLDb wakes up
2008-01-05 02:56 < The359> why would we know?
2008-01-05 02:56 < jtizzle1854> but that doesn't mean the info's not out there
2008-01-05 02:56 < The359> did you try Google?
2008-01-05 02:56 < White_Cat> SQLDb good
2008-01-05 02:56 < White_Cat> now
2008-01-05 02:56 < SQLDb> Heya White_Cat
2008-01-05 02:56 < White_Cat> shall we play a game?
2008-01-05 02:56 < jtizzle1854> i was just wondering
2008-01-05 02:56 < Rinn> Hard R rating?
2008-01-05 02:57 < SQLDb> White_Cat: Starcraft? :)
2008-01-05 02:57 < White_Cat> Which side do you want?
2008-01-05 02:57 < White_Cat> Wargames
2008-01-05 02:57 < SQLDb> heh
2008-01-05 02:57 < White_Cat> thermonuclear warfare
2008-01-05 02:57 < SQLDb> whoot!
2008-01-05 02:57 < SQLDb> China! :)
2008-01-05 02:57 < White_Cat> with live ammo
2008-01-05 02:57 < White_Cat> no china
2008-01-05 02:57 < SQLDb> hah
2008-01-05 02:57 < White_Cat> USSR and USA
2008-01-05 02:57 < The359> Defcon?
2008-01-05 02:57 < SQLDb> meh. I'm finishing up a game myself... "Which lead is hot?" :)
2008-01-05 02:58 < White_Cat> SQLDb I need your help
2008-01-05 02:58 * SQLDb chuckles and puts away the multimeter
2008-01-05 02:58 < White_Cat> I'll pm
2008-01-05 02:59 < pengo> jtizzle1854: Actually R-rating should link to [[Motion picture rating system]]
2008-01-05 03:03 < jtizzle1854> ok
2008-01-05 03:03 < jtizzle1854> to other countries?
2008-01-05 03:03 < jtizzle1854> not the US
2008-01-05 03:03 < pengo> yes that has other countries
2008-01-05 03:04 < pengo> including US
2008-01-05 03:04 < jtizzle1854> ok
2008-01-05 03:04 < ryanssss> BLESS YOU!
2008-01-05 03:04 < pengo> I came across that page when it was a redirect from [[R rating]] when i was looking for "R ratings" of insulation
2008-01-05 03:04 < jtizzle1854> what r-rating link in wikipedia?
2008-01-05 03:05 < jtizzle1854> what's the address pengo?

Wikipedia-Watch home page | Wikipedia-Watch hive mind

These logs from a freenode IRC channel were emailed to PIR by anonymous
third parties. They are made available by PIR under Section 230 of the CDA.