[05:29:46] legoktm: did you get access to pgehres' stuff? [05:29:51] I didn't. [06:50:57] pour one out for the toolserver [07:36:31] * legoktm grumbles [07:36:37] loginwiki is in small.dblist [07:36:51] I guess that's based on # of articles, not # of users? [07:44:58] I'd guess [07:45:07] legoktm: also almost completely devoid of edits [12:02:29] Firefox keeps telling me "The connection was reset/interrupted". [12:02:59] Are the servers down? [12:03:20] No problems here [12:04:56] sha1sum: No known problems at the moment. Do you still have these problems? [12:06:27] They started shortly before I joined this room but they are gone now. (They were not limited to my home computer though.) [12:11:23] "dispenser: MIT or GPL 2+ upon death or after 6 months of inactivity" http://koti.kapsi.fi/~federico/tsusers/licenses.txt [12:11:37] * Nemo_bis looks for legal definitions of "inactivity" [13:01:23] Nemo_bis: That's out of date, current one doesn't allow porting [13:02:32] Plus there's license incompatibility due to me copying & pasting snippets from the web [15:54:43] James_F: Weird tags again [15:54:47] https://meta.wikimedia.org/w/index.php?title=Tech/News/2014/28&diff=9051697&oldid=9033376 [17:36:44] I was looking for graphs giving some idea of the trends in upload.wikimedia.org traffic but there doesn't seem to be anything meaningful. [17:41:03] Nemo_bis: When more people are awake there's more traffic [17:41:05] JOB DONE [19:16:33] Hello, there must have been some recent changes made to the Wikipedia site. The resolution seems worse, blurred, on my Ipad 2. I don't remember it looking like that yesterday. Does anybody know what happened, and if it might be fixed somehow? I tried both Safari and Chrome on the Ipad with the same result... [19:16:51] Pescoluse: known bug, let me find it [19:18:45] Pescoluse: https://bugzilla.wikimedia.org/show_bug.cgi?id=67286 [19:19:13] Pescoluse: the bug is fixed already, but the fix is not yet live on Wikimedia sites [19:20:11] That fix was merged yesterday, it seems? [19:20:18] The mobile team may want to request a cherry-pick [19:20:21] brion: ---^^ [19:20:54] Ah! Thanks a lot. Then the pages will eventually look better again soon :-) [19:22:34] Pescoluse: Yeah, it was fixed on Monday. But in the normal process, that fix won't go live until next Thursday, so I'm suggesting they might want to fast-track it [19:22:47] RoanKattouw: it’s nothing to do with MobileFrontend, it’s core [19:22:54] Oh, heh [19:23:05] it’s specifically the desktop site that is affected [19:23:30] Oh it's desktop on mobile? [19:23:52] depending on whether you consider an ipad mobile :) [19:23:56] brion: Are you gonna be around between 4 and 4:30 today? [19:24:04] RoanKattouw: i can be online yeah [19:24:04] Cause if so I'll just list it for SWAT [19:24:08] and probably do today's SWAT too [19:24:09] ok [19:24:09] Sweet [19:24:37] Pescoluse: I'm putting this fix on today's fast-track list so it'll go out around 4pm Pacific Time [19:24:51] excellent [19:25:02] i’ll have my simulators out and ready to test :D [19:26:02] i just happened to ask ori about this, too [19:26:13] so coordin- oh, RoanKattouw's on it. [19:27:44] Sounds great! [19:28:36] And it is only when reading the desktop site on the Ipad, the mobile site looks crisp. [19:34:37] brion: Also that's a really interesting side-effect of -webkit-backface-visibility: hidden; that I didn't know about. I mean I already knew that property was radioactive and should basically never be used, but I didn't know it could cause blurryness [19:34:53] I have seen it cause transparent images to render solid black, and I have seen it cause selection interaction issues (!) [19:35:01] fun :D [20:35:41] hi guys. i was wondering if this is the place to talk about Cite.php. i am hard core into the formatting of citation metadata, and i use name- and list-based references with {{rp}} for page numbers. [[template:RP]] indicates that this is the currently technically superior and forward-compatible way, but that improvements to Cite.php are needed in order to reconcile the use of these multiple techniques into one, as well as to complete the ... [20:35:48] ... obsolescence of {{sfn}}. have i assessed the situation accurately? [20:35:53] i'm not a programmer, just an editor [20:37:20] dtm: if by 'improvements' you mean 'total rewrite', then yes [20:37:22] That's en.wiki specific information [20:38:03] dtm: Cite doesn't support any of the citation templates, and doesn't "know" about them – it only generates markers like "[1]" that link to wikitext entries in references list [20:38:30] what is inside these entries is currently not Cite's business at all [20:38:51] MatmaRex: okay but how about page numbers? are there plans to unify those in? [20:39:03] or generally to expand Cite's business [20:39:12] dtm: there are no plans to unify anything in, as far as i know [20:39:17] okay [20:39:28] someone might have plans, but i haven't heard of anything :) [20:39:28] is Cite being presently rewritten? [20:39:34] nope [20:39:46] well, unless someone is doing that covertly somewhere [20:41:05] Well, there was some integration of *extremely common* usages [20:41:11] For VisualEditor's sake [20:41:40] dtm: Wait. [20:42:01] dtm: Anything – /anything/ – that says don't write but instead write {{…}} is "better" is a lie. [20:42:22] dtm: Templates make it worse, not better, trying to impose local hacks (that break) instead of fixing the actual software problem. [20:42:57] dtm: {{rp}} and {{sfn}} are parts of the problem, like {{reflist}}, and should be deprecated. [20:43:11] dtm: More positively, what are you actually trying to do and what features would you like? [20:44:25] James_F: hmmmmmmmmmm. it seems to me that {{sfn}} is pretty obviously ridiculous, and that {{rp}} is actually functional. and, [[template:rp]] says that it is future proofed in the sense that it can be automatically replaced in the future by a bot, pending infrastructural upgrades to reconcile page numbers directly into the reference in a properly flexible way. [20:44:59] dtm: Whoever wrote it didn't talk to me, who would have pointed out that they're wrong. :-) [20:45:23] James_F: what i do all the time is cite the heck out of things, down to the page numbers. Shirley, you're aware of the problem of citing lots of pages from the same work. that's the main challenge that i'm finding to be inelegant. [20:45:26] dtm: Adding page= to s to allow multiple-explained-re-use is something we're looking at. [20:45:45] yeah. which is presently what {{rp}} does, and that's the only way i know of to do so. [20:45:56] dtm: But sadly we don't have much time to do it because we have to spend so much time fixing up after {{sfn}} and {{rp}} and the like. [20:46:01] though it leaves the overall citation in two pieces. [20:46:07] Which is the main problem. [20:46:08] how do you fix up after them? [20:46:12] if you dont mind me asking [20:46:30] i really really care about this and i do it all day, and i'd convert everything to something better if it existed [20:46:35] Anguished screams in IRC when our incremental improvements break their stupidly-coded templates. :-) [20:47:34] in fact, both times i've upgraded articles from {{sfn}} to named/list/{{rp}}, it got instantly reverted, saying that the decision to have four thousand redundant references at the bottom of every article as a victim of its own success, is clearly superior by fiat [20:47:48] and i'm sure that if i further inquired, i'd find that this is definitely not WP:OWNER either [20:48:02] but rather that they do what is right, and it is right because they do it. [20:48:29] We could modify sfn with a self-completing subst:, which would be pretty evil but would work. [20:48:41] K4-713: +v? ;-) [20:48:48] what? a self-deleting subst, ya say? [20:48:59] j/k [20:49:04] :-I [20:49:05] James_F: I was about to ask specifically you about that. [20:49:17] Yeah, it's a horrible hack where a template self-subst:s. But it might just work™. [20:49:21] K4-713: What about it? [20:49:27] ...it's new. [20:49:32] * James_F checks [20:50:02] James_F: And, you know... I might go mad with power or something. [20:50:07] K4-713: It… shouldn't be happening, as you're not in the access list. [20:50:15] RoanKattouw: You are – can you see what's happening and fix? [20:50:35] dtm: Anyway, I hope we can make some progress on page= which will be useful. [20:50:41] dtm: Will write up some thoughts. [20:50:53] < James_F> dtm: Adding page= to s to allow multiple-explained-re-use is something we're looking at. <-- so when you say this, do you mean that you're integrating the concept of {{rp}} directly into ? so that i can have multiple pages in a named and listed ref, like [20:50:57] James_F: As far as I know, the only place I should get +v is -fundraising. [20:50:58] ? [20:52:27] James_F: I am what now? [20:52:31] Near K4? [20:52:31] Fact number one fact number two definitely reliable and not at all made up crap. [20:52:37] like that? [20:52:40] dtm: Exactly. You'd have …, and thereafter would re-use it, but would re-use it with ", page 33–44" added to the end of the re-use, probably as a sub-list. [20:52:46] Yes. [20:52:50] s w e e t [20:53:10] RoanKattouw: You're in this channel's access list, I'm not, so K4-713's access list oddity is something you're more likely to be able to fix than me, [20:53:26] Oh OK lemme check [20:54:12] James_F: so if you do that, plus a bot which automatically converts {{rp}} and possibly the train wreck that is {{sfn}} into that, and possibly implement the functionality of this plus reflinks plus checklinks into visualeditor, then we're winning like Charlie Sheen? [20:54:17] I think it's because she has an @wikimedia cloak [20:54:24] dtm: Possibly. [20:54:25] That gets her +V [20:54:30] RoanKattouw: I have an @wikimedia cloak… [20:54:31] James_F: i want to encourage this most heartily, sir [20:54:41] James_F: i want to encourage the human people who do this, sir [20:54:48] dtm: :-) [20:55:14] Maybe she was voiced once and it remembers? I don't know [20:55:51] James_F: so what would happen if other efforts (as you mentioned about break/fix firefighting) were combined into that? or is it a super huge undertaking that might require a kickstarter? [20:55:51] K4-713: You could try to un-voice yourself I guess? [20:55:53] dtm: We aim to please. :-) [20:56:07] dtm: No Kickstarter necessary. :-) [20:56:11] James_F: are we talking about a few months [20:56:20] RoanKattouw: I don't necessarily mind. It's just happened automatically a couple times now, and I'm confused. [20:56:22] dtm: Around then, possibly sooner. [20:56:39] James_F: i want it! i mean. uhhhh. are there any roadblocks? [20:56:58] dtm: Community happiness / wariness to change is the main one. :-) [20:57:02] James_F: would it be possible to auto upgrade {{rp}} and {{sfn}} [20:57:12] because yes we need a migration path [20:57:25] dtm: We could probably get a bot to do them, but it might be hard. [20:57:29] in order to avoid malaise and dumbness [20:57:47] James_F: well see that's why i suggested GUI integration. for manual assistance. [20:57:57] i mean. i'm sure you know that, but i just wanted to indicate that i knew that. [20:58:09] VisualEditor will have the GUI integration side of it, believe me. [20:58:19] We're responsible for that too. [20:58:32] James_F: i used reflinks many times a day, and i'd massage its output. sometimes i'd even delete stupidly half-formed refs down to the bare url for a moment so that i could run reflinks on it. [20:58:33] Building a dedicated GUI just for this, though, seems a bit wasteful. [20:58:38] yes. [20:58:46] that's why i suggested visualeditor [20:59:09] * James_F nods. [20:59:11] i am all hard core and stuff, and yeah i'm even a unix hacker like the Jurassic Park kid, but i want to migrate to visualeditor someday. [20:59:19] i'm one of those pragmatic types [20:59:21] Good. We want you to, too. :-) [20:59:38] dtm: You'll probably like what I have planned for next-gen Reflinks, if WMF ever approves the space for it :-P [20:59:48] i tried the first widespread release of it and got the newsletter until i realized it was a quagmire of controversy, so i let that pass [20:59:57] DispenserAFK: oh please! please do! [21:00:13] RoanKattouw: hah... I also can't -v myself. [21:00:54] If only the web had archiving tools already... [21:00:59] K4-713: Because you're not in the list. :-) [21:01:03] Yep. [21:01:10] DispenserAFK: a930913 just told me the reflinks story today. i had no idea it was proprietary software, based on a 24TB database (FOR WHAT???) with ultimatums on WMF hosting. [21:01:18] so that's my wiki gossip for the day [21:01:32] dtm: Have you seen our plans to auto-create cites based on URL/ISBN/DOI/etc.? [21:01:57] Technically open sourced [21:02:06] James_F: negative [21:02:15] James_F: do this; i beg of you [21:02:40] i type my citations manually most of the time. in the url bar of the given web site. [21:02:43] >:-I [21:02:54] i've done so many hundreds of citations that i have the parameters memorized [21:02:58] and i use them all [21:03:01] so, what is the process nowadays of enabling a non-controversial MediaWiki extension on a Wikimedia project? [21:03:22] dtm: https://www.mediawiki.org/wiki/Citoid is the initial code. [21:03:30] i would do beta testing, people [21:03:36] Get local community consensus, file a bug on Bugzilla, pray to God someone will submit a config patch. [21:03:49] dungodung: Is it already enabled on a wiki? [21:03:53] i was ranked #10 on beta testing Vada, last i knew [21:03:56] Review! [21:04:04] James_F: than [21:04:04] x [21:04:13] James_F: it is on some wikis, but not on the one I want :) [21:04:29] Which extension? [21:04:29] DPL [21:04:34] dungodung: Ah, in that case it's relatively easy. Make sure you've got community consensus… Oh. [21:04:43] On a production wiki? No. [21:04:45] non-controversial, heh? [21:04:49] * James_F laughs. [21:04:54] heh [21:05:05] DPL is the only extension we've got deployed that can trivially take down bits of the cluster. [21:05:15] what if the community (which I'm kinda part of) told me they wanted it, but they need it soon [21:05:27] it's a small project [21:05:32] Which? [21:05:34] It's seriously unlikely that we'd be able to let it be enabled on any more wikis. [21:05:34] srwikt [21:05:44] Small is good in this case [21:07:02] ok, so if I get community consensus, could I hope to have it? :) [21:07:35] Might need to check in with our DBA... [21:07:43] I mean, I don't really care for myself, it's just that they'd like to be able to display category content in wiki pages [21:07:58] so, unless there's a better way... [21:08:02] DispenserAFK: can i help test your citation stuff? [21:08:10] James_F: and yours [21:09:01] dungodung: Doesn't {{Special:ListPages/Category:Foo}} or similar work? [21:09:29] oh, hadn't tried that :) [21:09:29] Unfortunately not, still waiting for a response from WMF [21:09:40] 5-6 months now [21:10:08] Response to what exactly? [21:10:15] Where? [21:10:59] The 24 TB to make it work [21:11:13] Where is the request? [21:11:40] DispenserAFK: why would citation formatting require 24TB? [21:11:49] Archivinh [21:12:02] So you can do it an entire site at a time [21:12:42] DispenserAFK: "An entire site"? [21:12:48] i'm boggling. [21:13:08] dungodung: That was invented syntax, no idea if that's a real thing. [21:13:12] Enwiki [21:13:14] an entire site of what? and... why 24TB? [21:13:34] and, why not just not do that? [21:13:52] James_F: well, no, there's Special:AllPages, but it doesn't do that [21:14:25] James_F: and Special:CategoryTree doesn't seem to be transcludable [21:15:22] Not the special page! [21:15:25] dtm: 24,000,000 external links * 1 MB average web page size now. Lets say you know the author is in element #articleauthor, you could submit that and it would autofill from that for all other uses in the tool [21:15:31] It has syntax [21:15:58] DispenserAFK: 24,000,000 external links of *what*? ^_^ [21:16:19] HTML, CSS, JS, Images and Video [21:16:33] DispenserAFK: Wait, you want to cache the Internet? [21:16:37] DispenserAFK: That's… insane. [21:16:44] i'm trying to figure out what's not insane about this [21:16:53] how about cut out the insane part? [21:17:08] Why are you caching content? Aren't you only interested in meta-data? [21:17:22] and that's just as much because it's insane, as for expediency [21:17:43] cache is for other ideas down the line [21:17:51] also, compression exists. it really really really exists. [21:18:04] for other, non-insane ideas? [21:18:11] hello dungodung [21:18:18] So… you want to spend a million bucks on an ultra-high-end storage system to cache the Internet (in violation of the TOU about copyrighted materials) for unspecified ideas? [21:18:19] hey Nemo_bis :) [21:18:23] Good luck with that. [21:18:26] s/insane/cutting edge [21:19:14] DispenserAFK: so why didn't you just migrate the non-insane part? [21:19:37] i mean. why not just not be insane? [21:19:43] A million bucks? archive.org buys hard disks at 100 $ / 3 TB [21:19:44] insanity optional [21:20:03] dungodung: what sort of extension would this be? [21:20:16] There may be other players interested in sponsoring my adventures... [21:20:37] DispenserAFK: are you holding reflinks hostage? [21:20:40] Nemo_bis: well, DynamicPageList can list all articles from a specified category. I basically just need that [21:21:04] 24TB storage isn't just 24/disk size [21:21:14] No, it just takes effort for porting [21:21:34] dungodung: the standard DPL or the non-Wikimedia one? [21:21:50] DispenserAFK: where's the source code? [21:21:56] TS [21:22:07] DispenserAFK: it's publicly published under a free license? [21:22:41] Its on archive.org too [21:22:46] If I die [21:22:47] Nemo_bis: didn't know there were versions. all I know is that srwiki has one (prolly the standard one), and I need it on srwikt as well [21:23:05] DispenserAFK: .............. what? [21:23:07] dungodung: well it was enabled even on wikipedias nowadays, shouldn't be hard [21:23:16] anyway back to AFKing [21:23:40] Nemo_bis: ok, so if there's community consensus, I guess we should get it soon [21:23:49] James_F: so.... do... we have non-insane people ready to replace reflinks and checklinks, into visualeditor? [21:24:29] James_F: that's basically what citoid is? [21:25:08] dungodung: I hope so, just standard bug in Site requests with link to consensus and pointing out the extension is already enabled on way bigger wikis [21:25:35] Nemo_bis: right. great, thanks [21:30:40] How long should it take for an extension change to get pushed to the deployment-prep apaches? [21:33:20] dtm: Citoid will take a URL and convert it to a cite; we're planning to use it in VE to let people add new properly-formatted cites, but it could also be used to fix up existing bare links. [21:33:39] James_F: h u z z a h [21:33:49] James_F: so is there a publicly deployed test? [21:34:06] dtm: No, very early days right now. :-) [21:34:28] James_F: okay [21:34:31] dtm: We've not looked at a reference-staleness-checker. [21:35:00] dtm: One of the longer-term things we'd like to achieve is proper structured data storage of cites, which could be regularly checked for updates/staleness/deadness etc. [21:35:17] And of course feeding the links into the Internet Archive and working with them to ensure that we retain caches of those links. [21:35:29] James_F: well, just grabbing the citation metadata would be good. maybe some site-specific templates so you can see how certain sites deliver their author names and stuff [21:35:37] and that'd be fine to start with huh? [21:35:47] Hi [21:35:59] Is it OK if I interject with a technical request? [21:35:59] dtm: Look at https://en.wikipedia.org/wiki/Foobar?veaction=edit in the "Cite" menu. :-) [21:36:03] Qcoder00: Sure. [21:36:03] James_F: is there a way i can be notified of testing of this and of upgrades to Cite? [21:36:29] Is there a way of re-working Extension:Proofreadpage, so it can work with external djvu? [21:36:41] James_F: like when you obsolete {{rp}} and cement the obvious obsolescence of {{sfn}} [21:36:45] (I.e djvu hosted on sites other than commons) [21:37:18] dtm: changes of this magnitude should be announced in tech news [21:37:19] dtm: There's a roughly-monthly posting about Editing team news to enwiki's technology Village Pump (WP:VPT). [21:37:21] !technews [21:37:28] James_F: kthx [21:37:34] dtm: But we'll be talking loudly about it for a while, don't worry. :-) [21:37:45] uh, i keep forgetting this doesn't work here. i'm talking about the same thing as James_F :) [21:37:51] James_F> And of course feeding the links into the Internet Archive and working with them to ensure that we retain caches of those links. [21:37:55] They're already doing that [21:38:08] dtm: https://meta.wikimedia.org/wiki/Tech/News [21:38:13] Qcoder00: That's a question best asked of Tpt, who maintains it. [21:38:16] So labs is takign over from Toolserver.. [21:38:29] I did already leave tham a note but they didn't respond [21:38:45] Nemo_bis: The IA have a feed of structured data of external links used in references in all 900 of our wikis? If so, how, and can we use it? [21:39:03] Another thing that would be useful is the ability to do in browser page rotation of scans when doing edits in Page: namespace [21:39:13] Qcoder00: Have you filed this on Bugzilla? [21:39:18] Not yet [21:39:25] Becaust it's not a bug as such [21:39:33] It's a 'wishlist enhanchment' [21:40:26] James_F: my quote didn't contain any "structured data" bit [21:40:30] Qcoder00: That's what Bugzilla is for. [21:40:56] Nemo_bis: So how do they know which links are just links and which need checking regularly for status? [21:43:09] Sorry, I don't understand this supposed difference [21:43:34] They archive all new links and once they're in wayback machine they're recrawled semi-regularly like any URL [21:43:37] Oh OK [21:43:50] I need to structure my thoughts on this a bit though [21:44:21] (I've not yet determined the current exact coverage of this feature, see https://www.mediawiki.org/wiki/Talk:Archived_Pages ) [21:44:49] MatmaRex: thank you [21:45:42] Also who works on Mediwiki special pages? [21:46:10] It would be 'nice' to have Proofreadpage status integrated into Special:contributons if possible [21:46:16] (Something else for Bugzilla) [21:46:22] Qcoder00: That's a ProofreadPage bug, yeah. [21:46:32] s/bug/enhancement request/ [21:46:34] Whatever.:-) [21:46:43] Tpt is quite responsive about PRP [21:46:50] He is? [21:47:04] Where should I ask him about the page rotation scan thing? [21:47:13] I tried what I thought was there talk page... [21:47:20] He's on the IRC during the daytime [21:47:39] Ah [21:47:52] European daytime that is [21:47:53] Log a bug [21:49:11] I assume you've also heard about VADA ? [21:57:29] James_F|Away: also, I suppose you know already, but: https://archive.org/help/wayback_api.php [22:28:01] James_F|Away: (or anyone else) would i be correct in thinking that the aforementioned problems like of reference pages, and additionally the issue of so-called overlinking, are due to a lack of abstraction between data and presentation? each article, or each user's account, should be configurable as to whether they want to see wikilinks or references, and how many, at all [22:28:38] there shouldn't be any policy at all about it, and there's just a lot of needless social unrest and chaos due only to a technology issue. so it seems to me. [22:28:51] every single word in wikipedia should be wikilinked afaic [22:29:02] give or take. [22:29:03] *yawn* [22:29:18] Nemo_bis: yes, wayback integration would be supreme. man it's hard to find valid content there. [22:29:27] it often archives error pages [22:29:56] and i bemoan the lack of cached images. :( [22:30:08] but wayback is a glorious thing. [22:34:56] it would help so much if there was a realtime slider bar, for a wayback interface. if it could fetch the last several archived pages, and i could slide between them, or have a thumbnail view of them all on one screen, from which to click, and have my wikipedia citation done (archiveurl, deadurl=yes, archivedate), that'd be awesome. [22:35:03] <3 <3 <3 <3 [22:35:41] hurray hurray gotta love the new 80 Gb link https://monitor.archive.org/weathermap/weathermap.html [22:36:30] dtm: something similar to https://toolserver.org/~dispenser/view/Reflinks ? :p [22:37:31] There is a sort of slideshow tool like that, though it's single-WARC-file-based, IIRC. It's doable [22:38:14] Nemo_bis: where's "the internet" on that graph? [22:38:17] It's mainly a problem of not hammering the index and the data transfer too much [22:38:21] greg-g: on the sky [22:38:24] HE and ISC [22:38:31] well, and I2 [22:39:01] I2 is Internet2, the equivalent of GÉANT [22:39:11] yeah [22:39:19] so 16ish gigs up [22:39:27] (I suppose someone would say the opposite but I claim my right to be eurocentric) [22:39:55] Nemo_bis: I worked at the U of Michigan, one of the big I2 partners, so I'm a bit I2 centric ;) [22:40:06] :P [22:40:35] And consider that for instance the Aaron video has some 300 seeders right now for the archive.org torrent, that also helps :) [22:41:18] nice [22:41:48] I'm used to being the only one downloading anything (usually live music shows) that the torrent is slower than http [22:42:02] (the spin up on their end, I presume) [22:53:43] dtm, re: "every single word in wikipedia should be wikilinked" - I mentioned a few of the problems with that, at https://www.mediawiki.org/wiki/Talk:Beta_Features/New_Features#Wikify_Everything [22:55:41] TL;DR: in [[Flag of North Korea]], the phrase "The white stripes symbolize purity" might (would?) get mis-linked to [[The White Stripes]]. (It might be interesting as a user-selectable gadget, but I don't think it could ever be a default site setting) [22:56:27] mm, bad example, i guess a capitalization check would prevent that one. but youknowwhatimean.