IRC Log Meeting 7 February 2006

[Stuart Ellis]

(21:10:08) quaid: <meeting>
(21:10:15) quaid: hey m'friends, how's the week going?
(21:10:36) stickster: Scurrying, hurrying and worrying
(21:10:47) quaid: aye
(21:12:30) quaid:
http://fedoraproject.org/wiki/DocsProject/FedoraDocsSchedule
(21:12:35) quaid: ok, so we had the slip last week of the schedule
(21:12:46) quaid: I still haven't gotten the release-notes to build
properly
(21:13:00) stickster: Let me at it
(21:13:03) quaid: sure
(21:13:16) quaid: I think I need to maybe make the legalnotice* stuff
real XML files
(21:13:22) quaid: and xinclude them, too
(21:13:31) quaid: the problem is "in" the fdp-info stuff
(21:13:39) quaid: in that it has entities and the like
(21:13:58) quaid: stickster: I'll check in what I have right now, as we
move forward w/meeting
(21:14:21) stickster: disco
(21:16:23) megacoder: I guess everyone here is too cheap to have a paid
lwn.net subscription?
(21:16:24) stickster: Do we want to follow the list or just let it all
hang out?
(21:16:44) G2: megacoder: I am ;-)
(21:16:45) stickster: Hey, I resemble that remark
(21:17:12) ***quaid is supposed to be able to use an RH corporate sub,
but can't get it to work for him
(21:17:22) quaid: stickster: follow the list?
(21:17:27) quaid: oh, the schedule
(21:17:41) stickster: yup
(21:17:59) quaid: we'll go down the list, why not, just move quickly
(21:18:07) quaid: ok, OPL ... I haven't sent a draft of anything yet
(21:18:15) quaid: but I did raise the question to Gregdek about the
Wiki ...
(21:18:22) stickster: relnotes == busy
(21:19:20) ***quaid looks for that msg
(21:20:48) quaid: here's his idea:
(21:20:56) quaid: 1. Agree to OPL w/o restrictions.
(21:21:02) quaid: Steps 2a-2c should happen pretty much simultaneously:
(21:21:02) quaid: 2a. Snapshot a list of all wiki users who have
accounts *right now*. Make 
(21:21:02) quaid: that list into a wiki page. Call it the "legacy wiki
user OPL agreement 
(21:21:02) quaid: page" or some such. Put a watch on the page. People
who agree should put 
(21:21:02) quaid: a little "check" or something on the page. With watch,
we should be able 
(21:21:03) quaid: to tell if someone is doing something wacky, like
agreeing on behalf of 
(21:21:05) quaid: other people.
(21:21:07) quaid: 2b. Add language that says "any content here is OPL
content" to the
(21:21:09) quaid: account creation / login page, BIG AND BOLD so people
can't miss it.
(21:21:11) quaid: (http://fedoraproject.org/wiki/UserPreferences). 
(21:21:15) quaid: 2c. Announce OPL policy -- and point legacy users to
the "lwuOap" page, 
(21:21:17) quaid: explaining how to agree to the new policy, and what to
do if you don't 
(21:21:19) quaid: (i.e. go and remove your edits.)
(21:21:21) quaid: 2d. Slap "OPL" on every single page in the footer.
(21:21:23) quaid: 3. Monitor the watch list for compliance.
(21:21:25) ***quaid tickles nman64 here
(21:21:39) quaid: we don't think we need to get any legal approval here,
it's just actions that follow the legal directions we've been given
already.
(21:22:12) G2: looks good to me.
(21:22:20) quaid: 2c worries me, though
(21:22:29) quaid: it's impossible for anyone to remove their
changes/content, I think
(21:22:33) G2: Links to a summary of the OPL in case people haven't been
following the list etc?
(21:22:48) quaid: so, I think we cross the bridge when we get there ...
if someone insists on not agreeing, we have to figure out what can be
done.
(21:22:59) quaid: G2: yeah, within 2c
(21:23:02) stickster: You know *someone* will just to be cantankerous
(21:23:03) G2: yeah. There will be one.
(21:23:13) megacoder: Don't look a me ;-)
(21:23:28) stickster: :-D You're too busy revolving!!!
(21:23:31) quaid: if we get a lot of work because of this
(21:23:46) quaid: I'm going to shout loudly at all the "let's just do
the Wiki without thinking it through first" supporters
(21:24:05) MicroBob [n=bob at pdpc/supporter/sustaining/BobJensen] entered
the room.
(21:24:36) elliss: We kicked around a OPL licensing FAQ when we
discussed this perviously
(21:24:42) elliss: previously
(21:25:02) elliss: That should tackle most of the questions/compliants
(21:25:15) elliss: complaints
(21:25:29) elliss: I need sleep...
(21:25:32) quaid: elliss: did we get any draft going?
(21:25:34) quaid: I forget
(21:25:39) elliss: Yes
(21:25:48) stickster: elliss: You sent something to the list, didn't
you?
(21:25:54) stickster: Or was it wiki?
(21:26:01) elliss: It was a PM attachment
(21:26:08) elliss: In MoinMoin format
(21:26:46) elliss: It'll be somewhere in my mail folders...
(21:26:48) quaid: http://fedoraproject.org/wiki/DocsProject/OPL/FAQ
(21:26:53) G2: cool
(21:26:55) quaid: is that a sane namespace?
(21:27:03) quaid: i mean, can you put it there, if that's a good name :)
(21:27:10) elliss: Licensing/FAQ ?
(21:27:24) quaid: yeah
(21:27:35) G2: looks fine
(21:27:38) elliss: Or DocsProject/Licensing/FAQ ?
(21:27:43) quaid: yes
(21:27:50) quaid: DocsProject/ first
(21:27:59) elliss: OK.
(21:28:41) ***quaid moves DocsProject/OPLDiscussion ->
DocsProject/Licensing/Discussion
(21:30:13) quaid: ok, I'll send out an action plan to f-dsco-l right now
(21:30:36) stickster: k
(21:31:10) MicroBob: quaid can you verify i am on that list please
(21:32:12) stickster: nman64: Can we get the webtest.f.r.c to also do a
"make rpm" and "make spec" for doc modules there
(21:32:35) nman64: That should be possible.
(21:32:43) stickster: nman64: This gives us a canonical location to send
FE reviewers for .spec and .src.rpm files
(21:32:44) quaid: MicroBob: affirm, you are
(21:32:51) stickster: nman64: Thans
(21:32:53) stickster: *thanks
(21:32:57) ***stickster shoots typist
(21:33:01) MicroBob: quaid thanks
(21:33:34) quaid: nman64: can you address the OPL-plan on f-dsco-list?
(21:33:58) nman64: I do not believe I am subscribed to that list.
(21:34:03) quaid: ok
(21:34:22) G2: How many nicks do you have MicroBob ? ;-)
(21:34:42) megacoder: "not enough", I think.
(21:34:50) MicroBob: g2 a bunch, i use 6
(21:34:51) quaid: nman64: low vol. list, can I sub you?
(21:35:00) stickster: nman64: One more note, we need those "make rpm"
and "make spec" targets in the docs-common/ module as well
(21:35:02) nman64: quaid: Yeah.
(21:35:14) quaid: nman64 at n-man.com
(21:35:15) stickster: Pigpile on Patrick
(21:35:57) nman64: stickster: The builder package is in CVS. Changes can
be made there. Additional configuration beyond the packaging can be
emailed to me or to the list.
(21:36:45) stickster: nman64: Oops, I had forgotten... let me do that
then
(21:36:48) stickster: Sorry
(21:37:57) quaid: anything you want to say/ask on packaging?
(21:38:34) stickster: Umm... "It's working well"?
(21:38:38) stickster: Until we break it again ;-)
(21:38:42) megacoder: native-compiled FOP in our future?
(21:39:27) quaid: lordy, I don't know what the status is
(21:39:36) quaid: I sent a tickler to f-devel-java-l last week but no
reply
(21:40:00) megacoder: I ask because we're building up more
infrastructure based on my hacked xmlto. Maybe we should fork in under a
different name?
(21:40:16) quaid: hmm
(21:40:32) megacoder: Fork "xmlto" to "fdpto" or something equally
silly.
(21:40:57) megacoder: Oh, well, let the idea percolate.
(21:41:03) quaid: let's first
(21:41:09) quaid: lean hard on the Java people to finish
(21:41:22) quaid: and then lean hard on twaugh to finally upstream your
changes
(21:42:15) MicroBob left the room (quit: "Leaving").
(21:42:21) elliss: OK, that FAQ is now up:
http://www.fedoraproject.org/wiki/DocsProject/Licensing/FAQ
(21:42:41) quaid: thx
(21:43:06) stickster: quaid: Made report to Matthias
(21:43:37) quaid: stickster: thx
(21:43:37) stickster: I think he thinks we're putting all our docs in
fedora-release... I clarified
(21:43:37) MicroBob [n=bob at pdpc/supporter/sustaining/BobJensen] entered
the room.
(21:43:42) quaid: ah
(21:45:15) stickster: I also gave him a quick rundown on how to see the
packaging in action
(21:47:16) megacoder: I like the FAQ, btw
(21:47:47) quaid: ok, another plea sent
(21:49:25) elliss: megacoder: Licensing FAQ ?
(21:49:43) megacoder: Thats the 1
(21:50:26) elliss: Cool. Looking at it, it needs messaging to explain
that the Wiki content is "documentation"
(21:51:48) elliss: I'll do that in a minute...
(21:53:18) elliss: Did we discuss the survey ?
(21:53:36) elliss: (last "priority one" on the schedule)
(21:53:51) quaid: haven't 
(21:54:13) ***quaid saves some updates to the schedule
(21:54:27) quaid: what I want to know is
(21:54:36) quaid: we know most of the stuff there, where are areas of
suckage are
(21:54:51) quaid: so, actions /= new stuff to write, build, etc.
(21:54:55) quaid: because we are trying to do that
(21:55:02) quaid: maybe actions == more resources?
(21:55:25) quaid: like, should I take those results to the Foundation
board and ask them to help us to do ... what?
(21:55:38) quaid: the more concrete we can be, the better.
(21:56:27) elliss: We need writers, but that needs more than resources
(21:56:53) elliss: People have to want to do it
(21:57:44) quaid: they could focus recruiting on documentation for us,
for example
(21:57:54) quaid: maybe do something special at FUDCon
(21:58:23) quaid: also
(21:58:32) quaid: if we had more technical resources, we could build
wicked infrastructure
(21:58:51) elliss: I think that's key in a way
(21:58:51) quaid: and that would attract people (promotion of same,
etc.)
(21:58:56) elliss: I keep looking at other projects, and one thing
strikes me...
(21:59:31) elliss: The one's with a lot have content have webpages, not
formal documents 
(22:00:24) stickster: We've had quite a few volunteers pop up in the
last week or two... where are they coming from? (Not that we're not
overjoyed of course)
(22:00:51) G2: not sure. I have a load to add to the list
(22:00:51) quaid: stickster: I'm not sure if it's more than usual or
not, but what is diff is this
(22:00:55) quaid: they are getting active right away
(22:00:56) elliss: Publicity from test releases ?
(22:00:59) quaid: v. disappearing
(22:01:11) stickster: quaid: Yes, it's been a real kick
(22:01:25) elliss: I think our breakthrough has been to assign/suggest
work fast 
(22:01:28) quaid: I counted this mon. v. the last few months, it wasn't
a big upswing in self-intros as much as an upswing in continuity from
the intro'd
(22:01:57) quaid: elliss: thank stickster for that idea 
(22:02:00) stickster: Hugo geared up quickly, did a bunch of work
already this past week
(22:02:03) megacoder: I think it's a threshold problem: success will
breed success; documents breed documents.
(22:02:11) quaid: we're all so ingrained to the FLOSS way of "make it
yourself"
(22:02:18) stickster: megacoder: I think you may have hit it on the head
(22:02:24) quaid: that we tell people to look around, etc. and they
disappear
(22:02:27) elliss: Not yet at the tipping point :)
(22:02:34) G2: need to dash. Chat tomorrow guys
(22:02:38) stickster: quaid: And that is probably precisely where we
need to depart from std. thinking
(22:02:38) quaid: everyone just about replies, yes tell me what to do,
I'll find my own way later.
(22:02:40) quaid: G2: cheers
(22:02:45) G2 left the room ("See you all tomorrow.").
(22:02:55) quaid: right, one reason people come to docs is to -not- get
treated that way
(22:03:02) quaid: "The kinder, gentler sub-project,"
(22:03:17) stickster: quaid: Agreed, I think I would have lost without
early help from several individuals
(22:03:26) stickster: s/lost/lost interest/
(22:03:43) megacoder: I expect a surge of activity after FC5 and we've
got stuff to show.
(22:03:47) quaid: stickster: exactly, and you are the sort who _does_ do
well just leaping in
(22:04:28) quaid: man, I hope we can get relnotes translated for all
languages for FC5
(22:04:40) quaid: megacoder: I still owe you some Make-fu, right?
(22:04:53) megacoder: Right now, how to we distinguish FDP from tldp?
(22:05:05) megacoder: Specifically, what make-fu?
(22:05:20) quaid: megacoder: F /= TL?
(22:05:31) megacoder: Huh?
(22:05:33) quaid: I mean, are distinguishing characteristics are == 
(22:05:39) quaid: * Fedora-centric
(22:05:44) quaid: * Fedora-only
(22:05:52) quaid: * provides Fedora best practices, canonical
(22:05:56) quaid: right?
(22:06:01) elliss: Also, perhaps, we build docs to *ship* with Fedora
(22:06:05) quaid: yes
(22:06:11) quaid: and they are updated for the release
(22:06:18) quaid: and tied to the release, v. being generic Linux
(22:06:25) quaid: megacoder: xml2pot and po2xml, basically
(22:06:55) quaid: megacoder: I know you can figure it out, just going to
see if I can find some stuff from the RHEL packaging to kickstart and
make us more like downstream (For a change)
(22:07:26) megacoder: I've added "make po" to the Makefile.common, and
an "apply_po" verb to fdpsh, so we are probably OK unless you've got
something special in mind.
(22:07:33) ***quaid is feeling the pain this week on his own imposition
about per-lang sub-dirs, as he updates internal docs-common usages.
(22:07:38) quaid: oh
(22:07:39) quaid: sorry
(22:07:41) quaid: didn't notice :)
(22:07:42) megacoder: Of course, I'll look at anything.
(22:07:53) ***quaid bends over for megacoder's kick in the arse
(22:07:57) megacoder: Just don't get weird or something.
(22:08:02) megacoder: Oops, too late.
(22:08:04) quaid: not I!
(22:08:15) stickster: quaid: don't fret, the per-lang is a good idea...
Tommy and I have agreed, though, that the rpm-info per lang is a Bad one
(22:08:20) quaid: ok
(22:08:26) megacoder: ditto.
(22:08:28) quaid: I know it's good
(22:08:37) quaid: just feeling the effects,tha's all
(22:08:40) quaid: so
(22:08:52) quaid: can we tell Fedora Trans that we have .po files they
can pick up?
(22:09:25) megacoder: My point about distinguishing Fedora is that we
need momentum and FC5 should do that.
(22:09:29) quaid: ah
(22:09:55) quaid: megacoder: the translators need a way to get .po back
into .xml, is that function there?
(22:09:57) megacoder: Anybody with a Makefile can "make po". And even
commit it to CVS, now.
(22:10:11) stickster: Yes, new look is just superficial (but
crowd-pleasing), there's also a lot under the hood with this release,
way more than FC3->4
(22:10:20) quaid: yeah, I think we're going to manually carry .po files
into i18n.redhat.com
(22:10:34) megacoder: PO->XML isn't done via a make target, but there's
a separate script to fold it back into the XML.
(22:10:37) stickster: quaid: jargon alert... can you clarify?
(22:10:43) quaid: ok, that's what that is, thanks
(22:10:51) quaid: stickster: ok
(22:10:56) quaid: I forgot to report on a meeting I had
(22:10:57) quaid: here
(22:11:00) stickster: "manually carry"
(22:11:04) quaid: last week, which I sent IRC log to list of, etc.
(22:11:10) quaid: stickster: ideally
(22:11:30) quaid: the Web apps at i18n.r.c 
(22:11:41) ***quaid thinks
(22:11:59) quaid: 1. we need to move said Web apps to cvs.fedora from
elvis.r.c (i18n.r.c)
(22:12:06) quaid: trans CVS is still on elvis.r.c
(22:12:14) quaid: why do we need to do that?
(22:12:28) quaid: A: to allow the Web app to hook our .po files from our
make targets
(22:12:51) quaid: the webapp runs make, grabs .po, makes them avail to
trans, and uses the script to put the .po back in.
(22:13:09) quaid: to do that on separate machines takes resources we
don't have.
(22:13:26) quaid: bgroh wrote all that, and I put up a ticket to sopwith
et al to look at taking it over from him.
(22:13:31) quaid: along with the whole CVS move
(22:13:38) quaid: 2. Meanwhile
(22:13:57) quaid: we can give Fedora Trans access to our *-en.po files
(22:14:12) quaid: and they can import them into CVS on elvis.r.c
(22:14:21) quaid: and make them avail via the webapp
(22:14:43) quaid: so we get many more translations for FC5 relnotes (and
press releases) than just the heroes doing it by hand.
(22:14:54) quaid: I don't want to expose this to all guides yet, because
it's manual.
(22:15:09) quaid: we will have to manually commit the trans'd .po files.
(22:15:15) quaid: we == Fedora trans
(22:15:58) quaid: it's 0530 for aalam , so he's probably asleep, but I
believe that Fedora Trans team leaders understand all this as-is.
(22:16:18) ***quaid finishes thinking
(22:16:21) quaid: make sense?
(22:17:33) stickster: yup
(22:18:17) stickster: In other words, we're just adding one new make
target to allow people to pick this stuff up
(22:18:25) stickster: and translate
(22:18:25) quaid: afaict, yes
(22:19:12) megacoder: There must be a way to incrementally
create/maintain/merge .po files during the document life, or is every
update a fresh start?
(22:19:22) quaid: same CVS?
(22:19:25) quaid: would that do it?
(22:19:37) quaid: and they use the .po files in /cvs/docs only?
(22:19:46) quaid: i.e., we host our own translations, forevermore
(22:19:56) quaid: oic
(22:19:59) quaid: you mean, procedurally
(22:20:07) megacoder: XML->PO is a complete overwrite every time. They
probably have a process for handling incremental updates.
(22:20:10) quaid: ideally, they use diffs
(22:20:21) quaid: yes
(22:20:24) quaid: but it's been line-based before
(22:20:36) quaid: you may recall this as being one of the original
reasons for a good xmldiff tool
(22:20:42) megacoder: OK. Not my problem, then.
(22:20:49) quaid: I don't think so
(22:20:59) megacoder: Then I'm in favor of it.
(22:20:59) quaid: well, I don't know for sure 
(22:21:01) quaid: heh
(22:21:04) quaid: what I mean is
(22:21:17) stickster: The line based thing works because of the PO
format, right?
(22:21:20) quaid: we may be able to put more smarts in on our side,
since we're doing the .po generation
(22:21:27) quaid: I think so
(22:21:41) stickster: Well, an xmldiff tool may not help, because the
element is equivalent to a line in PO
(22:21:45) quaid: I need to get bgroh to explain it to me.
(22:21:57) stickster: Have you run xml2po?
(22:23:17) ***stickster very slowly catches up to megacoder's logic
(22:23:36) quaid: no
(22:23:58) stickster: I'm just doing it and looking at the output, i.e.
PO formta
(22:24:00) stickster: *format
(22:25:48) stickster: Each block-level XML element gets a msgid, and the
translation (msgstr) is indeed blank... but it would be a simple thing
to compare the set of existing msgid's to the old set and diff, which is
probably how the i18n.* system does it... i.e. a new package doesn't
reset the counter to 0 every time, I think
(22:26:06) stickster: If the msgid hasn't changed, no need to translate
(22:26:41) stickster: I'm sorry if I am being slow here, I may not have
the gist of the real issue
(22:27:28) quaid: right
(22:27:32) quaid: but when it changes
(22:27:36) quaid: I think they do some sort of diff
(22:27:43) quaid: such that, for RHEL docs
(22:27:56) quaid: we're supposed to not C-c C-q (or M-q) to fix the
block in Emacs
(22:28:16) quaid: but leave extra long or short lines in <para/> so they
whole block doesn't have to be visually diffed
(22:28:21) megacoder: Step 1: You have an XML file.
(22:28:21) megacoder: Step 2: make po.
(22:28:21) megacoder: Step 3: add translations to .po file.
(22:28:21) megacoder: Step 4: touch the original XML file.
(22:28:21) megacoder: Step 5: make po.
(22:28:28) megacoder: Step 6: see all previous content to PO file
replaced by blanks.
(22:28:54) stickster: Right, you'd have to have an intermediate diff in
there
(22:29:09) megacoder: As long as they know to do this step manually...
(22:29:32) quaid: manually by the webapp?
(22:29:45) quaid: I think the webapp replies upon the smarts being in
the Makefile
(22:30:31) megacoder: I see we need more details of their process before
we let them roar off with a fresh batch of po files.
(22:31:06) quaid: ok
(22:31:20) quaid: hmm
(22:31:25) quaid: how to engage ?
(22:31:49) quaid: I'll ask bgroh to respond to a thread on f-trans-l
(22:31:51) megacoder: Maybe they have a twiki that describes the process
for newbies
(22:34:00) megacoder: We're running a bit long; I've got to drop out.
Anything else for me ATM?
(22:34:02) quaid:
http://webtest.fedora.redhat.com/docs/translation-quick-start/en
(22:34:04) quaid: somewhat
(22:34:14) quaid: thanks
(22:34:16) quaid: </meeting>

-- 

Stuart Ellis

stuart at elsn.org

Fedora Documentation Project: http://fedora.redhat.com/projects/docs/

GPG key ID: 7098ABEA
GPG key fingerprint: 68B0 E291 FB19 C845 E60E  9569 292E E365 7098 ABEA
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://listman.redhat.com/archives/fedora-dsco-list/attachments/20060208/c2ffd009/attachment.sig>