[04:38] <ddaa> lifeless: okay, knits upgrade started
[04:38] <ddaa> I reduced the max number of job per slave to 3, to avoid running out of memory
[04:39] <ddaa> also hacked a better bzr BatchProgress so long conversions will not time out
[12:22] <sabdfl> good work ddaa
[12:22] <sabdfl> so, what's up with staging? where's the log?
[01:51] <lifeless> sabdfl: I think its a code skew issue of some sort
[01:51] <lifeless> possible the new zope
[01:52] <lifeless> I'm digging
[01:52] <lifeless> db looks completely fine to me
[01:58] <mhb> hello
[02:00] <mhb> I hope this is the right place to ask ... I would like to know when we finally be able to translate Edgy Eft into local languages
[02:00] <mhb> I asked at the ML ubuntu-translators, but sadly, nobody replied
[02:05] <lifeless> sabdfl: making progress, it is db.
[02:05] <lifeless> something has borked the permissions on the karma cache
[02:06] <mhb> or can you recommend me another IRC channel or mail address where I can get a reply, please?
[02:12] <lifeless> mhb: hi
[02:12] <lifeless> the folk who know are usually here
[02:12] <lifeless> but its middle of the weekend
[02:13] <lifeless> I would not expect a reply until moday
[02:13] <mhb> lifeless: oh well
[02:14] <lifeless> you can try #rosetta, #ubuntu-devel, or the launchpad-users mailing list
[02:14] <lifeless> mailing lists are usually the best place
[02:14] <lifeless> and the ubuntu-translators list is definately the right place
[02:14] <mhb> nobody replied for more then a week or so
[02:14] <lifeless> how long has your message been unanswered on that list ?
[02:14] <mhb> than
[02:14] <lifeless> I suggest mailing again
[02:14] <lifeless> sorry I cant offer better than that
[02:15] <lifeless> pitti *may* know, or carlos, but neither are around just this minute
[02:15] <mhb> lifeless: if you meet someone who's responsible for Rosetta, tell him to check the ubuntu-translators list mor often :o)
[02:18] <mhb> lifeless: well I guess mailing the same question to the same ML won't help a lot
[02:18] <lifeless> seriously, it may well help
[02:18] <lifeless> sometimes things get dropped 
[02:18] <mhb> I'll check the archive
[02:19] <mhb> Thu Jul 27 17:58:39 BST 2006 
[02:20] <mhb> a *lot* more than a week :o) the question is here, btw: https://lists.ubuntu.com/archives/ubuntu-translators/2006-July/000743.html
[02:23] <kiko-afk> mhb, I know that edgy translations will be open RSN
[02:23] <kiko-afk> carlos has been working on updating the database to allow for it
[02:27] <mhb> kiko-afk: thanks
[02:28] <mhb> kiko-afk: I can't believe that, though ... the first time I asked this question was about a month ago, at 10th of July
[02:29] <mhb> kiko-afk: 15:39 < carlos> mhb: after that I will test it and try to move it into production... I guess we could say one week or so
[02:29] <kiko-afk> mhb, so I was unaware of this issue myself back then
[02:29] <kiko-afk> mhb, but I became aware of it last week 
[02:29] <kiko-afk> I think that this has been very poorly communicated, but I do think it is on-track (i.e. it will not be delayed forever)
[02:29] <kiko-afk> mhb, I'll have carlos update me and then /I'll/ post to ubuntu-translators.
[02:30] <mhb> ok, thank you
[02:30] <mhb> I'm sorry, but it disappoints me when someone says "a week" and nothing happens for more than a month
[02:31] <kiko> mhb, well, I agree that the communication has been bad, but I think they ran into more problems than expected (as usual with complex software). you should still have been updated though.
[02:34] <mhb> kiko: it's more painful for me because some of the Czech translators asked me about it and I (as one of the Administrators on that team) promised them to ask someone in charge of that
[02:35] <kiko> mhb, IKWYM. I think you can tell them that:
[02:35] <kiko> a) the changes to the DB were more complicated and took longer than expected.
[02:35] <mhb> kiko: so when nothing happened even though I told them otherwise, the blame goes to Ubuntu as a whole and myself
[02:35] <kiko> b) there will be changes done while Launchpad is offline next tuesday that relate to this, and it shouldn't take long from then
[02:36] <kiko> c) regardless, I'll post to ubuntu-translators to update you guys on how things are going
[02:36] <kiko> d) don't loose faith, it's tough at spots but we are really working hard to make this the best translation tool. ever.
[02:36] <kiko> mhb, just hang in there until monday
[02:38] <mhb> kiko: tell me, will there be a search for a specific word/sentence through all of Rosetta ?
[02:39] <mhb> kiko: that's what practically *every* person I showed Rosetta to asked about
[02:39] <kiko> mhb, it's one of the earliest planned features, so yes, it will happen. we discussed it in cape town in february 2005! I know it's a top requested feature
[02:39] <kiko> the problem with it is that that when we did experiment with that search it brought the servers to a halt ;)
[02:40] <kiko> so we've been considering ways of making it faster -- we're talking matching through millions of strings. lots of data.
[02:40] <kiko> let  me check on one thing.
[02:42] <mhb> kiko: IMVHO the first thing that comes to my mind with all these delays and features is the fact that Rosetta is closed-source application
[02:43] <kiko> hey
[02:43] <kiko> if you have volunteers that are interested in getting their hands dirty we have arranged code access in the past.
[02:43] <kiko> however, it's not easy work -- so don't expect the search feature, for instance, to be change 50 lines and hooray! :)
[02:43] <mhb> kiko: I know volunteers that refuse to work on it UNTIL it gets open-source
[02:43] <kiko> (if it was it would already have been done)
[02:44] <kiko> heh
[02:44] <mhb> kiko: like me, for instance
[02:44] <kiko> well, we've all got our opinions.
[02:44] <kiko> but
[02:44] <kiko> I should make it clear that this product is /really/ being worked on /for you/. there is no agenda other than building the best collaboration tool for open source
[02:45] <kiko> there are a number of reasons it's not free software yet. but it will be when the time comes
[02:46] <mhb> kiko: maybe we could define a "closed approach" - that's an approach when you can't answer the question directly and precisely
[02:47] <kiko> I don't understand what you mean.
[02:48] <mhb> kiko: an "open approach" is when the distribution says that it will 99% come out at 6.10 and I can trust it.
[02:48] <kiko> come out?
[02:48] <kiko> oh, the date.
[02:48] <mhb> get released
[02:48] <kiko> and a closed approach?
[02:50] <mhb> kiko: the "closed approach" uses words like "ASAP" and "when the time comes" ... and furthermore they avoid the direct answer why so
[02:50] <mhb> kiko: take Vista, for instance - nobody except the core devs know *exactly* why it got delayed
[02:50] <kiko> sorry, are you implying that I am avoiding direct questions?
[02:51] <mhb> kiko: I feel some of your/Canonical's answers are more "closed" than I'd like them to be
[02:52] <kiko> look, I can't tell you exactly why this has been delayed because I don't know myself, but I /can/ ask and then post you a summary with a good estimate of when it will happen. don't ask a developer when a feature will be ready because he will always get it wrong. ask me.
[02:53] <kiko> the difference to an open source project is yes, you can't see what the commits contain, and that open source projects don't really make date commitments for features.
[02:53] <kiko> but you do see what every individual commit was about
[02:54] <kiko> and we do make date commitments, which makes us prone to slipping from time to time.
[02:54] <mhb> kiko: IMHO with every slip you lose a bit of the community support
[02:55] <kiko> mhb, yeah, you're right.
[02:58] <kiko> we just need to do better.
[02:59] <mhb> kiko: Rosetta slipped 3 times for me - lack of the most wanted feature for quite some time, closed-source and now the delays
[03:00] <mhb> kiko: a lot of people leave our team because they feel they don't have anything to do
[03:01] <kiko> mhb, that's interesting. is Dapper fully translated into czech?
[03:01] <mhb> kiko: nope
[03:01] <kiko> hmmm. I don't understand what you mean then.
[03:01] <mhb> kiko: you keep adding new features into Dapper? 
[03:02] <kiko> the strings you add to dapper will be inherited by edgy, and by future releases
[03:02] <mhb> kiko: nope, and we do not keep adding translations into the Dapper applications - and if we do, we do it in upstream
[03:03] <kiko> and dapper will be around for a long time, so getting dapper language packs better and better will definitely promote better translation of software in general to czech.
[03:04] <mhb> kiko: you might know why we do *not* recommend to translate any GNOME/KDE/other non-ubuntu-specific tool in the Dapper rosetta page
[03:04] <mhb> because it's a waste of time - when you translate it in upstream, you help GNOME/KDE itself, not just one release of one distribution
[03:05] <kiko> that's not very true. upstream can pull automatically newer translations from Launchpad, and definitely everybody benefits with more strings translated than with less strings translated.
[03:05] <kiko> you seem to be making an odd point
[03:05] <kiko> you say your team has nothing to do
[03:06] <kiko> and yet you are unwilling to suggest to them to translate strings in dapper, the majority of which are going to be the same strings as in edgy.
[03:06] <mhb> kiko: do you know any larger project that uploads translations from Launchpad?
[03:06] <kiko> and which we make freely available to any upstream.
[03:06] <mhb> kiko: maybe I'm wrong, but  both KDE and GNOME doesn't
[03:07] <mhb> kiko: they use their own translations, we can modify them in Rosetta, but they won't get back to them
[03:07] <kiko> I know that individual projects definitely do, and that KDE and GNOME don't do it as policy because such policy would be controversial at best
[03:08] <kiko> so yeah, we should better relate to upstream. many of them don't even know they can download the translations.
[03:08] <kiko> (but even individual translators can, merging them with the current translations and then committing)
[03:10] <mhb> kiko: I talked to both the main KDE and GNOME coordinator for the Czech language. They refuse to upload anything if we don't send the files to them by hand in the way they want it
[03:11] <kiko> that's interesting. and what is "the way they want it"?
[03:12] <mhb> kiko: for example,  send the *exact* .po file that is in CVS/SVN with the new translations to them via email
[03:12] <kiko> mhb, but it's easy to do so. download the rosetta version, and run the merge commands with the current version. 
[03:13] <mhb> kiko: is it possible to download .po files directly, not through the slow librarian app?
[03:13] <kiko> we also do the upstream imports for a reason: our intention is to track CVS/SVN translations and offer pre-merged translations for them.
[03:14] <mhb> kiko: and don't you think translators want actually to translate, don't they expect the tools to do the administrative work for them?
[03:14] <kiko> heh. yeah, users do expect Rosetta to do everything for them. and we're trying
[03:14] <kiko> mhb, the librarian is not particularly slow -- it's just a fileserver. however, what you want is something like /products/ubuntu/+source/evolution/+pots/evo/download-latest
[03:14] <kiko> which automatically redirects you to the latest copy of the translations output by rosetta
[03:15] <kiko> mhb, let me stop chatting now and summarize what your concerns were so that we can give you an acceptable answer by monday. otherwise I will invest all my saturday into chatting but not fixing the things that are annoying you.
[03:15] <mhb> kiko: OK
[03:16] <kiko> man my wrist hurts today
[03:17] <mhb> sorry for that .o)
[03:18] <kiko> nothing to be sorry about, if the tool is annoying then you should indeed complain, and I should indeed do something about it
[03:20] <mhb> so: we would like to know a date, possibly the exact date when we can start translating edgy. I think if you want to be a friend with the translators in the future, notify them in the ubuntu-translators about almost everything - new features in Rosetta etc.
[03:21] <mhb> everything that is related to them, of course
[03:21] <kiko> mhb, okay, I'll add the weekly rosetta update excerpt and send it to them.
[03:21] <kiko> so you'll receive one this tuesday. and please comment on them if they are unclear or poor, I'll make sure to clarify
[03:22] <mhb> okay
[03:22] <kiko> mhb, is there anywhere else I should post these updates?
[03:22] <kiko> they have a highlights section
[03:23] <kiko> and then a detailed list of the individual commits and what they did
[03:24] <mhb> kiko: I don't know if someone else needs that
[03:24] <kiko> yeah, I don't know either. it's hard to decide when you're being annoying
[03:25] <mhb> kiko: please, tell me - who do you get feedback for Rosetta from?
[03:26] <kiko> mhb, carlos, danilo and SteveA are the people directly responsible, though I do work on Rosetta as well when I have time
[03:26] <mhb> kiko: I mean more of a user feedback
[03:26] <kiko> oh
[03:27] <kiko> to be honest, I mainly get feedback through bugs. we're not very good at keeping an eye on the many lists out there. we're trying this month to start keeping better track 
[03:27] <mhb> kiko: where and how many people should shout "we want this feature" so that someone responsible will realize this?
[03:27] <mhb> :o)
[03:28] <kiko> mhb, filing bugs is useful, and you can forward them directly to me (kiko@canonical.com) if you think they didn't get proper traction
[03:28] <mhb> kiko: OK
[03:29] <kiko> and /do/ forward them
[03:29] <kiko> because they will get traction
[03:30] <mhb> so I should file a bug about "Rosetta should be more upstream-developer-friendly" ?
[03:30] <mhb> especially KDE and GNOME
[03:30] <kiko> I think there are generic bugs like that. the best bugs are those that request very specific features that aid in that
[03:30] <kiko> I think there is even a bug to offer a permanent download link
[03:32] <mhb> off-topic: there are more issues than just the hard communication
[03:32] <kiko> heh. indeed. most of the issues are related to writing code! :)
[03:32] <mhb> for example KDE and GNOME have sometimes a different translation of a specific term
[03:33] <kiko> yes, right
[03:33] <kiko> the right way to handle that
[03:33] <kiko> is to have an order of precedence for translations
[03:33] <kiko> and to allow people to easily review the translations being inherited from other projects
[03:34] <kiko> so for instance KDE translations would prefer to inherit from other KDE products/packages first
[03:34] <kiko> we also need to offer translation overheads
[03:34] <kiko> so that people when translating KDE have easy access to a glossary of the words translated the way KDE does
[03:34] <kiko> it would be nice to even have highlights
[03:35] <mhb> sure, but first, the search feature, please :o)
[03:35] <mhb> we can't even track bugs without it
[03:35] <kiko> yes, I know it's the #1 feature requested.
[03:35] <kiko> mhb, just to be sure, it's searching a string in any language, right?
[03:35] <mhb> I have to grep the .po files on my computer to find out in what package is the mistake
[03:35] <kiko> english and czech in your case?
[03:36] <mhb> yes
[03:36] <kiko> would it make sense to search through /other/ languages in your case, or just those two?
[03:37] <mhb> kiko: well, it would be helpful for some people ... for example Slovak is very similar to our language, everyone in the Czech Rep. can understand Slovak, so someone would like to check if a word is translated in Slovak and if so, how
[03:38] <kiko> I see
[03:38] <kiko> mhb, but would it be acceptable for the UI to say
[03:38] <kiko> Search for: [                             ]  in [ Czech   | v] 
[03:38] <kiko> ?
[03:39] <mhb> it would be for me
[03:39] <kiko> all right
[03:39] <mhb> can I ask you one more question, please?
[03:39] <kiko> sure!
[03:40] <mhb> the Czech translation team does have a little dictionary of terms which are *strongly* recommended to translate as the dictionary suggests
[03:41] <mhb> we use wiki to achieve that
[03:41] <kiko> right
[03:41] <kiko> is it a single wikipage?
[03:41] <mhb> yes, just like that - word and translation etc
[03:41] <kiko> so
[03:41] <kiko> very easy strawman proposal:
[03:41] <kiko> wikilink for language guide
[03:41] <kiko> and display that prominently when translating
[03:42] <mhb> for example that
[03:42] <kiko> other less easy strawman proposal: fetch information from that page and display it (but where?)
[03:42] <mhb> I thought more of a large scale
[03:42] <kiko> a real tool, yeah. there's a spec on that called TranslationGlossaries I think
[03:42] <mhb> make a similar dictionary that can communicate with Rosetta
[03:42] <kiko> and offer it as an overhead
[03:42] <mhb> I'll take a look, thanks
[03:43] <kiko> the cool way to do that
[03:43] <kiko> would be to do it using ajax
[03:43] <kiko> with a floating div containing information on the words you typed into the box
[03:43] <kiko> so let's say I typed 
[03:44] <kiko> "o livro est na mesa" in portuguese
[03:44] <kiko> I'd get a little overhead saying things like "mesa: avoid use in translations related to fish".
[03:44] <kiko> and "livro: KDE prefers the term revista"
[03:44] <kiko> etc
[03:44] <mhb> yeah
[03:45] <mhb> I wanted to code that in AJAX myself
[03:45] <kiko> you know that the largest problem with rosetta
[03:45] <kiko> is actually the @#!!@@ large dataset
[03:45] <mhb> kiko: I can undestand
[03:45] <kiko> which causes us to scramble for performance even for simple suggestion lookup!
[03:46] <kiko> I spent like god knows how many hours looking at the SQL generated
[03:46] <kiko> ugh
[03:46] <kiko> danil[o] ut is currently working on improving that
[03:46] <kiko> made some progress last week
[03:47] <mhb> kiko: as I said, I wanted to do this tool, but I have other projects, so I couldn't lead the development ... and without the Launchpad account information it is also not so useful
[03:47] <kiko> well
[03:47] <mhb> kiko: and I don't want to do any community work that is non-free
[03:47] <kiko> you could develop a prototype 
[03:47] <kiko> that could be useful generally
[03:47] <kiko> client-side only
[03:48] <kiko> and license it to us separately
[03:48] <kiko> while still releasing it under your free software license of choice
[03:48] <kiko> LGPL it
[03:48] <kiko> we'd be happy to improve and give back improvements.
[03:49] <kiko> I did so for the javascript sorting code we use a few months ago
[03:49] <mhb> kiko: yeah, but nobody can promise me that, which saddens me
[03:49] <kiko> well
[03:49] <kiko> if it's LGPLed, being client-side, we do have to redistribute modifications
[03:50] <kiko> and we would if we said we would anyway.
[03:50] <mhb> when the time comes, I know .o)
[03:50] <mhb> no, sorry bout that
[03:50] <kiko> no, I meant immediately
[03:51] <kiko> if your code is LGPLed and we use it
[03:51] <kiko> because it is client-side
[03:51] <kiko> we need to release modifications to it
[03:51] <kiko> immediately
[03:51] <kiko> not "when the time comes"
[03:51] <mhb> ok
[03:51] <kiko> that only holds for client-side code
[03:51] <kiko> but.. it's still binding
[03:52] <mhb> by client-side you mean the code without the database access, or what?
[03:52] <kiko> correct
[03:52] <kiko> using an API to do the query to the remote server
[03:52] <kiko> (and if we implement the server-side you can use that API everywhere, not only in rosetta JS)
[03:52] <mhb> kiko: I think I could do it, but I have a lot of other things going, so it would be ready for edgy+1 translations, I guess
[03:53] <kiko> wow that would be most excellent
[03:53] <sabdfl> hey kiko
[03:53] <kiko> hello sabdfl 
[03:53] <kiko> what's cooking up north?
[03:53] <sabdfl> in case you're wondering, the crashes come from a db issue, not the code, though it's too soon to tell if the code itself has issues
[03:53] <sabdfl> the staging db update lost all db premissions
[03:53] <sabdfl> so lifeless rest the db to the production one, i think
[03:53] <kiko> me, wonder? ;)
[03:54] <sabdfl> which does not include the additional columns needed with the new code
[03:54] <sabdfl> kiko, i will take that pie if there's a single new oops, if i give the go-ahead after testing tomorrow, hmmmk?
[03:55] <kiko> well, I still don't see why we are in a hurry to put this into production -- the points I made still stand
[03:55] <sabdfl> the code keeps track of who started specs and when
[03:55] <sabdfl> the distro sprint is in a week
[03:56] <sabdfl> if we land it this week, we get much better data for edgy
[03:56] <kiko> I see
[03:56] <kiko> you hadn't said that
[03:56] <kiko> when does it start?
[03:56] <sabdfl> i just thought of it this morning
[03:56] <sabdfl> i think in a week
[03:56] <kiko> let me check
[03:56] <kiko> if it does indeed start on the same week as the soyuz sprint then I think it's a good idea
[03:57] <kiko> otherwise it should wait
[03:58] <kiko> sabdfl, it starts on the 21st, so I agree. 
[03:59] <mhb> kiko: could you send me some more info/documentation about the line between client-side and server-side in a mail, please? I want to know the rough structure first.
[03:59] <kiko> mhb, sure. are you aware of how ajax works?
[04:00] <kiko> it basically needs a way to contact the server
[04:00] <kiko> it uses an XMLHttpRequest object to pull the data
[04:00] <mhb> kiko: don't hesitate with the details, if there is something I don't know, I'll look it up
[04:00] <kiko> okay, cool.
[04:01] <kiko> mhb, what's your email address?
[04:01] <mhb> kiko: martin.bohm@einstein.cz
[04:01] <kiko> my pleasure
[04:04] <mhb> kiko: I would also be happy if you keep me informed when something related happens, like some other developer starts implementing that or so.
[04:04] <kiko> mhb, will do. the upcoming features section of the report usually tells about the new stuff being worked on
[04:05] <mhb> thank you
[04:06] <mhb> for the lengthy talk. I won't tire your hands anymore :o)
[04:06] <kiko> heh. thanks to you
[04:08] <mhb> I hope you don't forget about the translators next week :o)
[04:08] <mhb> bye all
[04:08] <kiko> I never forget. laters!
[04:24] <Mez> any LP admin: ping
[04:24] <kiko> Mez, yo
[04:24] <kiko> what's cookin
[04:30] <Mez> kiko: problems with bazaar... it's fscked up an upload- and i cant delete them
[04:30] <kiko> Mez, isn't it the case that you can ssh into the supermirror and delete it manually?
[04:30] <Mez> kiko: no permissions apparently
[04:36] <Mez> kiko : it gives me permission denied- so I need an admin to do it for me
[04:40] <kiko> mmmm
[04:40] <kiko> I can't do anything on the SM. I think it needs to be lifeless 
[04:41] <Mez> :'(
[04:41] <Mez> lifeless: ping
[04:41] <Mez> hmm - I believe I remember lifeless from ubz :P
[04:47] <LarstiQ> kiko: also, rm -rfing a directory seems to be problematic over sftp
[04:48] <Mez> LarstiQ, indeed ... but thats an sftp bug
[05:06] <Shart> hi
[05:07] <Shart> i want to request ubuntu cd, but i have a problem
[05:07] <kiko> hello Shart 
[05:07] <kiko> what's the problem?
[05:08] <Shart> there is phone number required
[05:08] <kiko> right
[05:08] <Shart> are you really want to call me?
[05:08] <kiko> that's required because you may be contacted if your package is held up in customs.
[05:08] <kiko> it's required by the shipping company.
[05:09] <kiko> canonical will not call you, no
[05:09] <Shart> i don't think it's cheap to call to Belarus
[05:09] <kiko> but a customs agent or the shipping agency on your side of the world may. it only happens in rare occasions.
[05:09] <kiko> it will be a call from somebody local
[05:09] <kiko> not from us
[05:09] <kiko> and it very rarely happens
[05:10] <Shart> ok
[05:10] <Shart> tanks
[05:10] <Shart> thanks* :)
[05:11] <Shart> is it really free of charge?
[05:11] <kiko> you're most welcome
[05:11] <kiko> yes, it is.
[05:12] <Shart> ok
[05:12] <Shart> we don't have so money as you do
[05:13] <Shart> )
[05:13] <kiko> I don't have any money
[05:13] <kiko> and I don't know what you are talking about! 
[05:13] <Shart> ok, sorry
[05:13] <Shart> bye
[05:15] <kiko-fud> aw
[10:18] <mdke> does anyone know whether, if I assign a bug to someone, and comment, they get that bug email, even if they weren't subscribed to the bug before that?
[10:18] <mdke> I mean, if I assign the bug and comment at the same time
[10:30] <LarstiQ> good question
[10:39] <BjornT> mdke: yes, the one you assign to the bug will get the comment.
[10:59] <mdke> BjornT: good, thanks. The absence of the good ol' bugzilla "Mail sent to X, Y, Z" made me worry
[11:13] <teolemon> hi