Sep 27 21:47:20 *    moyogo has changed the topic to:  Fonts discussion forum; Discussion logs at | Text layout summit preparation discussion at 2100 UTC here

Sep 27 21:47:26 *    otaylor ( has joined ##fonts

Sep 27 21:48:03 behdad    hi otaylor

Sep 27 21:48:22 otaylor    Hey behdad

Sep 27 21:48:40 *    behdad grab some tea

Sep 27 21:50:02 glasseyes    moyogo: thanks

Sep 27 21:50:52 moyogo    arg... only simosx can add other people to the access list

Sep 27 21:52:39 *    hdu_hh ( has joined ##fonts

Sep 27 21:54:21 behdad    where is simosx btw?

Sep 27 21:56:35 behdad    otaylor: so, am I wrong assuming that you'll be hacking on Pango during the summit? :-D

Sep 27 21:57:05 otaylor    behdad: I'm not sure about Pango *hacking* - I'll definitely be around, but I'm so far out of it, that I'm not sure that I should be allowed near the code :-)

Sep 27 21:57:23 behdad    otaylor: will get your hands dirty :)

Sep 27 21:59:37 *    maiku ( has joined ##fonts

Sep 27 21:59:47 *    maiku has quit (Remote closed the connection)

Sep 27 22:00:02 glasseyes    welcome everyone

Sep 27 22:00:13 *    maiku ( has joined ##fonts

Sep 27 22:01:45 glasseyes    I proposed this IRC meeting so that we could discuss anything we need to deal with before we meet

Sep 27 22:02:15 glasseyes    let's start by introducing ourselves

Sep 27 22:02:50 glasseyes    Daniel Glassey : linux porter for graphite

Sep 27 22:03:16 *    eimai dejavu fonts developer

Sep 27 22:03:17 maiku    Mike Fabian : working at Novell/SuSE on internationalization

Sep 27 22:03:24 *    eimai Ben Laenen

Sep 27 22:03:32 moyogo    Denis Moyogo Jacquerye : DejaVu fonts, me too

Sep 27 22:03:46 hdu_hh    Herbert Duerr: OOo GSL ( graphic system layer), WINxx port and text technology

Sep 27 22:04:11 behdad    Pango and Cairo hacker, i18n junky.

Sep 27 22:04:34 avox    Andreas Vox: Scribus developer, working on Scribus's new layout system

Sep 27 22:04:38 pjrm    Peter Moulder, not attending the summit.  Working on text layout; in particular, working with avox on things needed for optimal line breaking and how it interacts with shaping.

Sep 27 22:04:40 *    fantasai (i=fantasai@ has joined ##fonts

Sep 27 22:04:42 avox    err, *text* layout system

Sep 27 22:04:45 behdad    "Behdad Esfahbod"

Sep 27 22:04:57 otaylor    otaylor: Owen Taylor past pango, gtk+ maintainer

Sep 27 22:06:01 behdad    yosch: ping

Sep 27 22:06:26 *    edtrager ( has joined ##fonts

Sep 27 22:06:42 glasseyes    hi edtrager, please introduce yourself Ed :)

Sep 27 22:07:29 edtrager    Hi, everyone, I made it through traffic and got home ...

Sep 27 22:07:36 mrdocs    Peter Linnell, Scribus possibly coming to Boston..

Sep 27 22:08:00 glasseyes    the page with proposed topics is on

Sep 27 22:08:07 edtrager    Ed Trager:, going to Boston

Sep 27 22:09:09 fantasai    Elika J. Etemad aka fantasai, Mozillan, W3C Invited Expert (CSS Working Group)

Sep 27 22:09:21 behdad    hi fantasai

Sep 27 22:09:21 fantasai    can't make it to Boston due to a conflicting CSSWG face-to-face

Sep 27 22:09:30 glasseyes    welcome fantasai :)

Sep 27 22:09:37 otaylor    And, judging, from your web pages, the owner of one of the coolest book collections ever :-)

Sep 27 22:09:52 otaylor    [ for script junkies, anyways ]

Sep 27 22:09:56 fantasai    haha, no those are mostly from the library :)

Sep 27 22:10:07 fantasai    and a few from Martin Heijdra's personal collection

Sep 27 22:10:11 otaylor    fantasai: Well, you have access to some nice libraries, then

Sep 27 22:10:26 fantasai    yeah, Princeton University's got a good system :P

Sep 27 22:10:40 glasseyes    there are 3 areas I think we can discuss in:

Sep 27 22:10:42 *    chahibi ( has joined ##fonts

Sep 27 22:10:54 fantasai    Access to the East Asian library is free, too

Sep 27 22:11:19 glasseyes    clarification of the proposed topics, and discussion on how to plan how to organise them

Sep 27 22:11:29 glasseyes    practical details

Sep 27 22:11:48 edtrager    Hi, Elika, glad you could join

Sep 27 22:11:48 glasseyes    and input from the people who won't be there on what we should do

Sep 27 22:12:26 fantasai    I would love to listen in to the presentations, if that's possible

Sep 27 22:12:44 fantasai    and if I can get an audio feed, I can take minutes

Sep 27 22:12:56 behdad    cool

Sep 27 22:13:02 pjrm    that would be great.

Sep 27 22:13:04 pjrm    For the ~dozen ppl not coming, it would be good if people's presentations could be made available on the web.  Possibly even in advance of the conference.

Sep 27 22:13:15 behdad    keithp says he can do that if he gets a non-blocked-enough ip there

Sep 27 22:13:15 edtrager    Keith said it is easy to do if he can get an unfiltered IP address at MIT

Sep 27 22:13:24 glasseyes    cool, we definitely plan to have an audio feed

Sep 27 22:14:02 fantasai    what are the hours for the presentations?

Sep 27 22:14:24 fantasai    9am - 5pm or something?

Sep 27 22:14:39 glasseyes    do we have details on what we'll have in the room(or rooms) e.g projector, whiteboard?

Sep 27 22:14:50 behdad    fantasai: yeah, that was my idea

Sep 27 22:14:53 otaylor    fantasai: I'd guess 10:30-4:30 or so, judging from past years

Sep 27 22:15:37 fantasai    hmm, looks like that needs a bit of discussion then :)

Sep 27 22:15:47 pjrm    what timezone (/usr/share/zoneinfo/America file) is Boston?

Sep 27 22:15:53 fantasai    EST

Sep 27 22:16:02 fantasai    or Eastern

Sep 27 22:16:08 *    fantasai isn't sure what the file is called

Sep 27 22:16:08 edtrager    I suppose it is possible to both stream and record the audio at the same time, so the audio could be later posted on the web for latecomers or future reference?

Sep 27 22:16:12 otaylor    glasseyes: There will definitely be whiteboard, I'm not so sure about projectors, but I seem  to recall that the rooms actually had them built in

Sep 27 22:16:28 glasseyes    otaylor: thanks

Sep 27 22:16:39 otaylor    fantasai: I've just never seen people show up as early as 9 :-)

Sep 27 22:16:51 fantasai    heh

Sep 27 22:17:13 behdad    9 is wake up time :)

Sep 27 22:17:27 fantasai    so schedule for 10am, start at 10:30?

Sep 27 22:17:34 fantasai    :)

Sep 27 22:17:53 edtrager    Or schedule for 9:30, start at 10?

Sep 27 22:17:53 behdad    lets start stuffing topics into slots then adjust the slots

Sep 27 22:17:57 otaylor    edtrager: we can try. Audio rceording should be considerably easier than streaming, but anything is definitely going to be best-effort

Sep 27 22:18:32 fantasai    I'd prioritize recording over streaming

Sep 27 22:18:39 fantasai    not that both wouldn't be cool

Sep 27 22:18:46 behdad    so, Simon is arriving Saturday night.  others seem to be around for both days.

Sep 27 22:19:02 fantasai    Someone should take a bell, and hit it with a fork every time the slide changes :)

Sep 27 22:19:35 fantasai    "When you hear this sound *ding*, turn the page!"

Sep 27 22:19:44 behdad    is there anything someone wants to add to the agenda, or anything listed that needs to be removed/discussed?

Sep 27 22:20:04 glasseyes    I've asked Simon which bits topics he would be least interested in so we could do those on Sat (and he sends his regrets for not being here now)

Sep 27 22:20:10 edtrager    Hey, fantasai, that's a great idea!

Sep 27 22:20:11 behdad    fantasai: there may not be enough slides to be confusing ;)

Sep 27 22:20:12 *    avox wonders what to do if presenters jump a page back...

Sep 27 22:20:31 edtrager    Use a different "ding" sound for that?

Sep 27 22:20:46 behdad    fantasai: learned today, the best way to keep in sync is to let someone else handle slide navigation.  the speak then says "next slide please"

Sep 27 22:21:03 behdad    s/speak/speaker/

Sep 27 22:21:26 glasseyes    another thing with audio only - it would be good to be able to identify who is speaking

Sep 27 22:21:29 fantasai    that works too

Sep 27 22:22:14 pjrm    A bit harder for the speaker to search back to a particular slide (during a question) when someone else is operating the next/prev.

Sep 27 22:22:23 edtrager    I suppose the web page could identify current speaker while streaming, if you have that level of sophistication

Sep 27 22:22:43 fantasai    people should introduce themselves when the get up to speak anyway

Sep 27 22:22:49 pjrm    10:00 to 4:30 Boston (Atikokan,...) time is 15:00-21:30 UTC, btw.

Sep 27 22:22:53 otaylor    glasseyes: The list of topics looks pretty extensive, though I'm wondering if it would make sense to go in with a more focused set of goals, so we don't just all come in, say our piece, say "we should work together" and go back home

Sep 27 22:23:00 fantasai    or if someone else is introducing them, make sure to record that

Sep 27 22:23:22 glasseyes    otaylor: that sounds like a good idea

Sep 27 22:24:28 *    zwnj (n=zwnj@ has joined ##fonts

Sep 27 22:24:46 *    behdad is reorganizing the list of ideas

Sep 27 22:24:56 edtrager    On the list of topics, I've committed to a presentation called "Intern. Text Layout & Typography: The Big and Future Picture" which I would suggest occur early on on the first day (maybe even first?) so as to set the stage and have everyone on the same page.

Sep 27 22:25:25 glasseyes    fantasai: I expect it won't just be presentations - I would hope it'll be half presentation, half discussion during the presentation sessions

Sep 27 22:25:38 behdad    edtrager: that was my idea too.  kinda keynote

Sep 27 22:25:47 glasseyes    and have some purely discussion sessions as well

Sep 27 22:25:51 otaylor    glasseyes: It might be good to make the first item of discussion a fairly open discussion of "top issues that need fixing for text layout on oss" so we can keep that in mind through the rest of the weekend

Sep 27 22:25:55 fantasai    glasseyes: make everyone say their name before talking

Sep 27 22:26:01 behdad    glasseyes: yeah, we can allocate an hour, let presentatoins be as short as 20 mins

Sep 27 22:27:05 *    yosch is catching after a network timeout

Sep 27 22:27:10 yosch    hi everyone

Sep 27 22:27:21 edtrager    I am planning to send my presentation out to a "peer review group" ahead of time to see what I'm missing.  Hopefully that way I will be able to hit on all the major areas as an introductory, key-note sort of thing.  What does everyone think?

Sep 27 22:27:40 *    yosch is Nicolas Spalinger, SIL volunteer, OFL and Ubuntu fonts team

Sep 27 22:27:46 moyogo    edtrager: good idea

Sep 27 22:27:48 edtrager    People can suggest themselves or others if they want to do the "peer review" of the paper.

Sep 27 22:28:05 glasseyes    eimai, moyogo: what would your aims/goals of being there be from the fonts perpective?

Sep 27 22:29:48 edtrager    I also agree heartily with pjrm that everyone else who is going to prepare a presentation be able to post it a day or two ahead of time for everyone to read in advance.  That will make the discussions more informed, productive.

Sep 27 22:30:04 eimai    we would certainly do a talk about dejavu of course

Sep 27 22:30:10 glasseyes    edtrager: I think you should send it to everyone that will be there plus other interested folks

Sep 27 22:30:11 behdad    ok, I grouped the items based on the session type:

Sep 27 22:30:26 eimai    but we are there as the font developers

Sep 27 22:30:43 edtrager    pjrm: Who are you anyway?  I probably joined late and missed your introduction

Sep 27 22:31:12 zwnj    hi everyone

Sep 27 22:31:34 behdad    hi zwnj.  introduce yourself :)

Sep 27 22:31:59 moyogo    glasseyes: and advanced opentype features

Sep 27 22:32:28 glasseyes    behdad: I get some 'edit conflicts' looking at the page now

Sep 27 22:32:46 zwnj    i'm a developer of farsifonts package

Sep 27 22:32:50 behdad    ah

Sep 27 22:33:16 behdad    zwnj is Behnam Esfahbod, FarsiWeb.

Sep 27 22:33:19 otaylor    glasseyes: I'll clean that up, my mess

Sep 27 22:33:22 *    fantasai notes that there are a lot of Iranians here today :)

Sep 27 22:33:34 pjrm    edtrager: see the logs for an introduction to me and others you missed.  Err, except i've just looked at and can't actually see the logs (after a quick scan).

Sep 27 22:33:48 glasseyes    :)

Sep 27 22:34:06 edtrager    behdad: Are you and zwnj twin brothers :-)

Sep 27 22:34:11 behdad    fantasai: lets remove Latin from DejaVu fonts :-D

Sep 27 22:34:16 behdad    edtrager: not twin.

Sep 27 22:34:56 moyogo    behdad: :D

Sep 27 22:35:00 edtrager    behdad: cool!

Sep 27 22:35:21 eimai    behdad: I think you're patch won't be applied ;-)

Sep 27 22:35:25 yosch    edtrager: logs of previous IRC meetings are in separate archives pages

Sep 27 22:35:39 behdad    Behnam's in Iran though.  And yes, you really don't want the two of us commenting in the same bug!

Sep 27 22:35:52 glasseyes    eimai, moyogo: actually, you could you talk to keithp about how to get DejaVu back as the preferred font above Vera in upstream fontconfig ;)

Sep 27 22:35:53 behdad    eimai: no prob.  I'll rip them in Fedora ;)

Sep 27 22:36:20 behdad    glasseyes: why is that important?

Sep 27 22:36:21 chahibi    zwnj: it would be nice to rework Arabeyes font to support Persian and also hack its height

Sep 27 22:36:33 otaylor    Anyways, I'd suggest that we volunteer someone to organize the schedule and then just adjust as necesary at the start

Sep 27 22:36:35 behdad    glasseyes: we fix it in our dejavu-lgc-fonts package

Sep 27 22:36:44 chahibi    zwnj: But I guess you think the height for Farsiweb fonts is normal

Sep 27 22:36:47 otaylor    glasseyes: Do you want to do that?

Sep 27 22:37:01 eimai    glasseyes: yeah, I suppose discussions like that will take place in the dejavu talk

Sep 27 22:37:11 zwnj    chahibi: yes, that's in todo list already

Sep 27 22:37:15 behdad    otaylor, glasseyes: I can do.

Sep 27 22:37:20 chahibi    zwnj: Thanks :)

Sep 27 22:37:24 zwnj    chahibi: no, some of them do have problems

Sep 27 22:37:29 behdad    chahibi: the height for FarsiWeb fonts is broken, and Arabeyes fonts are ugly :)

Sep 27 22:37:43 glasseyes    I'll need to talk to the Debian font/dejavu  maintainers to see what they want (keithp is fontconfig maintainer so we can't change default)

Sep 27 22:37:46 *    otaylor hopes that discussions of font organization split-vs-mega will be kept strictly time limited. Maybe we need a cage :-)

Sep 27 22:37:59 behdad    hah

Sep 27 22:38:08 moyogo    yeah that issue could take forever

Sep 27 22:38:11 eimai    glasseyes: but everyone interested in dejavu future can join the dejavu meeting on #dejavu tomorrow at 18:00 UTC :-p

Sep 27 22:38:13 behdad    otaylor: there's a dejavu IRC meeting tomorrow discussing that

Sep 27 22:38:20 glasseyes    eimai: good point ;)

Sep 27 22:38:24 chahibi    behdad: Most fonts in Arabeyes may be considered as Artistic, fantasy fonts. I don't think they can fit well in a GUI or in a website text

Sep 27 22:38:43 behdad    actually it past my mind that: dejavu can ship separate font files, all having the same family name.

Sep 27 22:38:54 behdad    no reason to name the lgc variant "DejaVu LGC".

Sep 27 22:39:00 chahibi    behdad: May be except KacstOne but this one is different from Modern Persian typography

Sep 27 22:39:00 behdad    and fontconfig uses all available ones.

Sep 27 22:39:01 otaylor    OK, good that it will get discussed in advance.

Sep 27 22:39:15 behdad    chahibi: true.

Sep 27 22:39:20 glasseyes    I'll chase up Simon as we need input from him on what he wants to talk about from the Qt side

Sep 27 22:40:33 behdad    otaylor: very good point about testing.

Sep 27 22:40:37 otaylor    One thing I'd suggest to everybody to think about in advance is concrete achievable things that you'd hope to come out of the summit with.

Sep 27 22:40:46 edtrager    I recently discovered .  Uyghur Computer Science Association.  Some of the fonts appear very good.  Freely available, license not really specified.  I think all the Arabists should take a look and see what they think

Sep 27 22:41:05 chahibi    behdad: Roya and Nazli are wonderful but they look very small behind latin characters. DejaVu doens't have this problem

Sep 27 22:41:11 otaylor    behdad: I have to say that every time I see a change going into the Indic classificiation tables, I wonder "ok, that fixes one string. What strings is it breaking?" :-)

Sep 27 22:41:58 *    Pretor1ab has quit ("10-4")

Sep 27 22:42:06 behdad    otaylor: yeah, lemme add another item.  synching the indic shaper with ICU.  unfortunately I was under pressure to either commit LingNing's patches or go figure out the proper fix....

Sep 27 22:42:11 glasseyes    do we have a mailing list for discussing preparation and the general area that we can point people to?

Sep 27 22:42:14 moyogo    a lot of size issues could be fixed with proper BASE table stuff

Sep 27 22:42:18 behdad    otaylor: there's already one case that we did it wrong.

Sep 27 22:42:42 behdad    regardless, FarsiWeb fonts have too large font metrics

Sep 27 22:42:52 behdad    I suggest everybody subscribe to the wiki page for now

Sep 27 22:42:57 otaylor    glasseyes: The general GNOME summit preparation is mostly done on the wiki

Sep 27 22:43:10 edtrager    What if Fontconfig library could specify incremental or decremental size change "hints" for scaling fonts like Nazli and Roya to mesh with Latin defaults like Deja Vu Sans and Serif?

Sep 27 22:43:11 otaylor    There should be directions, etc, there. Was that what you were asking?

Sep 27 22:43:13 behdad    for anything else we can use fontconfig list I suppose

Sep 27 22:43:19 glasseyes    otaylor: ok, thanks

Sep 27 22:43:42 otaylor    edtrager: I have some ideas floating around about per-script size adjustments, that maybe you could override on a per-font basis

Sep 27 22:44:38 moyogo    otaylor: would that work with BASE table?

Sep 27 22:44:48 edtrager    For the Summit, I'm *very* interested in how Graphite is going to fit into the picture.  Graphite seems to have many cool features, I haven't really had time to digest it all.  Is Daniel presenting on it?

Sep 27 22:45:47 otaylor    moyogo: well, you'd assume that all scripts in a single font are sized as they would need to be sized when mixing text

Sep 27 22:45:47 behdad    a concrete goal of mine is to add BASE support to HarfBuzz/Pango in the summit :)

Sep 27 22:46:40 chahibi    behdad, zwnj: A Malaysian guy did this ( )

Sep 27 22:46:57 otaylor    moyogo: So it's somewhat different issue from mixing different fonts.

Sep 27 22:47:18 chahibi    behdad, zwnj: the Arabic script characters and latin height are equal, but it has some problems with some fonts

Sep 27 22:47:41 behdad    anyway, seems like time to wrap up :)

Sep 27 22:47:47 otaylor    moyogo: (Note this means that the Arabic portion of a 12pt multiscript font will be significantly different from a 12pt Arabic font.)

Sep 27 22:48:09 otaylor    behdad: Yeah, we seem to be discussing stuf fohter htan summit prep

Sep 27 22:48:16 moyogo    otaylor: in theory, a BASE table can define MinMax for scripts from other fonts too, but that might be undesirable

Sep 27 22:48:17 chahibi    behdad, zwnj: when I wanted to make an Arabic language tutor, this font was the only GPL font that supported diacritic marks and had a reasonnable height

Sep 27 22:48:39 glasseyes    yes, I'll present on it, the current state of integration and the things I've been told that graphite can do that are next generation stuff

Sep 27 22:48:49 zwnj    chahibi: looks very good on screen

Sep 27 22:49:25 chahibi    zwnj: It also has Bitmaps to make readable in 12

Sep 27 22:49:28 pjrm    When people are making their (draft) presentations available on web in advance, where are they going to go?  As links from ?

Sep 27 22:49:30 *    avox is interested in a common FOSS shaper

Sep 27 22:49:33 edtrager    otaylor: I like the idea of fontconfig possibly being able to fix up (or at least suggest fix-ups) when fonts have good glyphs but maybe imperfect metrics ...

Sep 27 22:49:44 zwnj    chahibi: yes, saw that

Sep 27 22:51:35 otaylor    pjrm: Links sounds best. Or you can attach to the page if you don't have a good place to put them

Sep 27 22:52:02 glasseyes    could people take ownership of all the items - to chair the discussions, or to present

Sep 27 22:52:25 glasseyes    if we don't have anyone to do them we can't really do them ;)

Sep 27 22:54:27 glasseyes    is everyone here and involved happy to subscribe to the fontconfig mailing list? any objections to using it?

Sep 27 22:55:26 eimai    glasseyes: why not ?

Sep 27 22:56:03 glasseyes    eimai: I'd prefer somewhere that is desktop agnostic

Sep 27 22:56:27 pjrm    If we do in fact get a live audio feed, then it may be useful if there's a way for non-attendees to talk back (ask questions), e.g. if there's VoIP available, or a telephone in the room.

Sep 27 22:56:43 glasseyes    afaiu this is a cross desktop thing even though it is at the gnome summit

Sep 27 22:57:26 glasseyes    pjrm: as long as we have network there will be irc, but voip would be useful

Sep 27 22:57:26 eimai    it doesn't really matter to me, but I thought it would be good to use that mailing list once :-p

Sep 27 22:57:38 glasseyes    eimai: agreed

Sep 27 22:58:24 pjrm    glasseyes: will there be a live speech-to-text transcription on irc :) ?

Sep 27 22:58:53 pjrm    In any case, updating the web page once we know whether live audio feed is likely to be available would be useful, so that non-attendees know in advance whether to make that time available and organize voip stuff.

Sep 27 22:59:01 avox    pjrm: well, you have IRC and the audiostream...

Sep 27 22:59:19 edtrager    glasseyes: Good.  I'd really like to see how Graphite, Pango, and QT layout engines compare on technical merits.  What are the strengths of each?  What are the weaknesses?  On first look, Graphite looks like it has some powerful stuff in there, and both Pango and QT are looking at Graphite integration, right?  Graphite may not be as familiar to some as Pango or QT, so I'd kinda like to see a Graphite presentation early on on

Sep 27 22:59:19 edtrager     day 1.

Sep 27 22:59:59 *    yosch is looking at the remaining sessions with no presenters

Sep 27 23:00:21 hdu_hh    not to forget ICU's layout engine...

Sep 27 23:00:33 yosch    edtrager: would you do the OFL campaign bit? Can't be there but I'll help prepare some more material

Sep 27 23:00:55 edtrager    yosch: Yes, certainly.

Sep 27 23:01:07 glasseyes    do we know if anyone from ICU can or can't make it (I think Eric Mader was asked)?

Sep 27 23:01:18 edtrager    yosch: just put my name on it.

Sep 27 23:01:40 avox    edtrager: I don't think the merits in intl' shaping will finally decide if aa layout engine is used or not...

Sep 27 23:02:23 avox    from an app developer's POV it's more important how convenient the API is

Sep 27 23:02:24 yosch    edtrager: excellent, I'm updating the wiki now

Sep 27 23:02:56 otaylor    avox: I think it's pretty clear that any commonality will be well below the level of an app developers view

Sep 27 23:03:31 otaylor    avox: I can't see either GTK+ or Pango presenting an API change to their users, especially to something low-level-C

Sep 27 23:03:39 avox    well, *most* app developers :-)

Sep 27 23:03:54 edtrager    avox: I know, but since Pango already has Graphite integration (partial, not all Graphite features supported) and Since Daniel told me that QT guys at Akademy are interested in Graphite integration, it looks like Graphite may become an important element in the existing, and future, infrastructure.

Sep 27 23:04:14 glasseyes    I know that the debian i18n meeting a few weeks ago was videod pretty easily - it wasn't streamed but the files were available pretty quickly - would we be able to find a camcorder and tripod from someone that'll be there?

Sep 27 23:04:16 avox    maybe

Sep 27 23:04:27 otaylor    avox: Speed, robustness, i18n support, advanced typographic support, ability to fit easily into existing high-level apis, etc.

Sep 27 23:04:33 avox    OTOH OTF is more commonly known and there are more OT fonts arounf

Sep 27 23:05:20 glasseyes    avox: graphite is mostly for situations where otf is insuffient, but I'll get to that in the presentation ;)

Sep 27 23:06:28 yosch    avox: and fonts can be both OT and Graphite (some SIL fonts have AAT as well)

Sep 27 23:06:36 *    avox knows

Sep 27 23:06:59 avox    layout engines should support both

Sep 27 23:07:18 otaylor    The hardest part of graphite integration seems to be deciding whether to use the graphite tables or OT tables from a mixed font :-)

Sep 27 23:08:44 edtrager    avox, otaylor: And I think Graphite could become very important for certain scripts like Myanmar and Lanna and Khmer where maybe OT is not quite good enough.  Or maybe OT is good enough, but the flexibility of Graphite is very appealing, so maybe its better to do it with Graphite.  Plus let's not forget that Graphite is an Open Source technology, whereas somebody else not in the FLOSS community is and will continue to call t

Sep 27 23:08:44 edtrager    he shots on OT and AAT technologies.

Sep 27 23:09:59 otaylor    edtrager: Well, there is only so far that you can swim upstream against what people are making for fonts. I can't really see using anything but OT for major scripts. But yes, sure, the more fonts that can be supported, the more scripts that can be supported, the better :-)

Sep 27 23:11:57 avox    what licence does Graphite have?

Sep 27 23:12:48 yosch    avox: LGPL (and CPL)

Sep 27 23:14:26 *    fission (n=fission@pdpc/supporter/active/fission) has left ##fonts

Sep 27 23:14:51 edtrager    Also I think Graphite -- maybe I'm wrong here -- but the SIL people do think about issues in a whole bunch of scripts and the Graphite manual says it supports all features needed for all known human scripts.  So just understanding the breadth and depth of such a feature set can give everyone perspective on future development avenues.  People can say, oh, Graphite supports that ?  Hmm., we can do that pretty easily in Pango o

Sep 27 23:14:52 edtrager    r QT to, but don't yet.  So I'm curious what Graphite's feature set is when it claims to be able to support "all known human scripts" (not exact quote .. my paraphrase)

Sep 27 23:17:07 glasseyes    yes, that is the kind of thing I'm hoping to present

Sep 27 23:17:32 glasseyes    I'm not sure if I should do that on Sat (early) or on Sun when Simon, the Qt guy, gets there

Sep 27 23:18:49 glasseyes    everyone, if you have comments on any of the topics please add to the wiki page, and subscribe to it or keep an eye on it

Sep 27 23:19:11 yosch    looking at the remaining session ideas again: any of the Dejavu team members want to give an overview of the current state of free design tools and what may need improving?

Sep 27 23:19:12 edtrager    glasseyes: Great!  Originally I thought early Saturday would be best for the Graphite presentation, but you are right, having QT there would be good.

Sep 27 23:20:21 mrdocs    one thing which is not listed, but I think relevant is font testing.. will they print with commercial printing gear..

Sep 27 23:21:18 edtrager    yosch:  The "Architecting a unified text layout engine for FLOSS systems" topic for me is most likely going to be part and parcel of my "International Text Layout & Typography : The Big And Future Picture" presentation.

Sep 27 23:21:34 mrdocs    AFAIK there is no common testing framework for open fonts, but some standardized tests could be done with existing tools

Sep 27 23:22:01 glasseyes    otaylor: would printing be part of "How to test layout engines"?

Sep 27 23:22:36 avox    edtrager; I would be interested to perr review your talk

Sep 27 23:22:37 edtrager    yosch: But I will also prepare some stuff on the "Next generation font dialog" subject as well, mostly as an update to what I've already got on

Sep 27 23:22:47 avox    s/perr/peer/

Sep 27 23:23:16 mrdocs    IOW "Can I use these fonts without worrying they might blow up on press - potentialy costing thousands of $currency? "

Sep 27 23:23:34 edtrager    avox: My pleasure. Do I know your email?

Sep 27 23:23:55 yosch    edtrager: yes, it makes sense. I can update the wiki if you want.

Sep 27 23:24:10 avox    edtrager;

Sep 27 23:24:10 edtrager    mrdocs: The idea of a common testing framework is good.

Sep 27 23:24:42 mrdocs    The quality of free fonts is highly variable :S

Sep 27 23:24:55 rillian    mrdocs: don't know if you've noticed, macosx now runs a font lint and warns you if there's a problem when you install a new font. unfortunately it doesn't give any technical details

Sep 27 23:25:22 mrdocs    rillian: I have a 10.4.x upgrade on the todo

Sep 27 23:25:22 rillian    (rillian is Ralph Giles, who works on Ghostscript. not coming to the summit)

Sep 27 23:25:26 edtrager    yosch: Can you add "common font testing framework" while you are at updating the wiki?

Sep 27 23:25:26 moyogo    mrdocs: i think we could benefit from common testing

Sep 27 23:25:34 yosch    mrdocs: back at the LGM you mentioned you had some process in place to test, was using Scribus or other tools?

Sep 27 23:25:34 moyogo    mrdocs: at least some kind of review system

Sep 27 23:26:10 fantasai    hm, there was a lot of discussion recently  in the CSSWG about embedding fonts in web pages

Sep 27 23:26:31 fantasai    two issues that came up were IPR issues and security

Sep 27 23:26:42 yosch    edtrager: otaylor and zwnj also have a test-related item there? Should it be merged?

Sep 27 23:26:43 fantasai    just to throw that out there

Sep 27 23:27:28 zwnj    yosch: i think you meant behdad, right?

Sep 27 23:28:02 mrdocs    edtrager moyogo I'll put down some thoughts in the next few days at and link it to the wiki

Sep 27 23:28:24 edtrager    fantasai: I agree "Embedded Fonts in Web Pages" should be added as a great topic for discussion.

Sep 27 23:28:34 yosch    zwnj: yep sorry glanced over the diff too quickly

Sep 27 23:29:23 fantasai    edtrager: the CSS2 spec already has syntax for doing that, but the only browser that supports it is IE, and only for .eot format fonts

Sep 27 23:30:02 yosch    edtrager: the wider topic should be open fonts on the web (a new set of open core fonts, metadata flagging in browsers, cross-platform behaviours, etc)

Sep 27 23:30:40 yosch    fantasai: yes the trouble is all the attempts to do something like this are still proprietary and non-cross platform

Sep 27 23:31:35 behdad    yosch: pango test suite, layout testing, and font testing are different problem.  of a common theme though

Sep 27 23:31:51 glasseyes    fantasai: what is IPR?

Sep 27 23:32:01 fantasai    Intellectual Property Rights

Sep 27 23:32:15 edtrager    fantasai:  I can imagine having FLOSS browsers be able to download required Open Source fonts for web pages on the fly from known URLs.  Forget about IE's .eot.  What about just having firefox go to a URL and downloading a font file as specified in web page?  To get fancy, Firefox could then open the file and check the license ... OFL? OK GPL? OK.  License not clear? ==> Popup warning message to user.  Anyway, it would be gr

Sep 27 23:32:15 edtrager    eat for the FLOSS community.

Sep 27 23:32:22 fantasai    font foundries that sell their fonts don't want them to be copied all over the web

Sep 27 23:32:58 fantasai    edtrager: yea, we need browsers with downloadable .ttf support or something

Sep 27 23:33:10 yosch    yes, I'm thinking that a browser extension telling you if the font called by the css is open or not (flagging the metadata) and allowing the user to act on it is a great way forward

Sep 27 23:33:14 fantasai    there are licensing bits in the ttf format

Sep 27 23:33:21 rillian    fantasai: aren't they already all over the web in pdf files?

Sep 27 23:33:39 fantasai    rillian: yes, but you can't reuse that

Sep 27 23:33:49 glasseyes    I've added a quick list of places where text layout is used to the wiki - folks please review and refine

Sep 27 23:33:50 rillian    because of subsetting I guess

Sep 27 23:33:54 fantasai    rillian: embedding them in web pages means making them available for download

Sep 27 23:34:06 fantasai    rillian: that and being embedded in the PDF

Sep 27 23:34:18 *    rillian sighs

Sep 27 23:34:28 fantasai    rillian: for web pages, embedding means linking, basically :)

Sep 27 23:34:36 fantasai    the other issue is security

Sep 27 23:34:41 edtrager    fantasai: Yes, so to be on the safe, conservative legal side, browsers like Firefox could just refuse to install "embedded" fonts where the TTF licensing data was not crystal clear.

Sep 27 23:34:54 fantasai    because fonts can contain programmed code

Sep 27 23:35:10 fantasai    our Microsoft rep says they have a collection of fonts that were designed to crash your system

Sep 27 23:35:17 *    mrdocs puts on BOFH hat and does not want any fonts downloaded on my users machines locally

Sep 27 23:35:36 fantasai    BOFH?

Sep 27 23:35:55 yosch    fantasai: you want the user to actually be able to lookup the website and grab the real font which he can then use on the rest of his system not just the browser. Creating a new format is less than ideal.

Sep 27 23:36:05 rillian    fantasai: Bastard (System) Operator From Hell

Sep 27 23:36:27 mrdocs    Bastard Operator from Hell, otherwise known as dictatorial sys admins....

Sep 27 23:36:34 pjrm    i gather that font distributors generally try to get their customers to enter into a contract (EULA), in addition to any restrictions of copyright (on the font program, not the letterforms afaik), trademark (on the font name), and design patent (rare).  Does "tracking IPR" include embedding info about licenses and known legal restrictions?

Sep 27 23:36:34 mrdocs    they do exist...

Sep 27 23:36:45 glasseyes    ok, I need to get to bed - an early flight back for some more aKademy - has someone else got the whole log and can post it to the wiki? Looks like the discussion is hotting up so I don't want to stop it ;)

Sep 27 23:37:02 glasseyes    just one thing, shall we meet again here before Boston?

Sep 27 23:37:13 glasseyes    if so lets plan a time/date on the wiki

Sep 27 23:37:23 mrdocs    why not same time next week ?

Sep 27 23:38:29 *    avox doesn't like apps enforcing IPR ...

Sep 27 23:38:39 edtrager    So sounds like the plan is for everyone meanwhile to adjust the wiki as they see fit, adding value to what has already been adjusted during this past hour.  I second same time next week.

Sep 27 23:40:45 glasseyes    ok, I'll check if Simon can make then, it would be good to discuss stuff with him so I'll propose other times if he can't make the same time next week

Sep 27 23:41:59 hdu_hh    edtrager:  I have some ideas/comments/requests on the topic of a unified layout engine

Sep 27 23:42:11 hdu_hh    Having worked with some layout engines I often miss some features application developers really need: text justification, device independent formating, layout compatible metafile replay capability (especially when the original font is not available)

Sep 27 23:42:25 hdu_hh    Will these features be discussed in your talk? Can I assume that there is a consensus that they are a requirement for a unified layout engine?

Sep 27 23:43:21 avox    I could probably also add to that... Scribus has some special neds

Sep 27 23:43:24 avox    needs

Sep 27 23:43:29 pjrm    avox's presentation topic doesn't currently convey the bit that i'm interested in, namely optimal line breaking and hyphenation and justification and how these interact with shapers: the correct glyph string depends on what line breaks are chosen, whereas choosing line breaks requires knowing width information from shapers, preferably including info on how much that width can vary for justification purposes by using glyph substitution or kashida

Sep 27 23:45:11 edtrager    hdu_hh:  Please write down all of your ideas ... Hmm, thinking fast here: Let's make "Architecting a unified text layout engine for FLOSS systems" an open discussion.  I'll volunteer to chair/moderate since my name is next to it already on the wiki.  How does that sound?  I'm sure many people have good ideas and expertise and can voice their ideas.

Sep 27 23:45:14 avox    yes, that's one of those needs.

Sep 27 23:45:34 *    yosch has quit (Read error: 60 (Operation timed out))

Sep 27 23:46:02 avox    edtrager: sounds good. so you would hold an introductory talk for the big picture and we would add our views?

Sep 27 23:46:06 *    otaylor has quit ("Leaving")

Sep 27 23:46:42 glasseyes    edtrager: yes, that sounds like a good idea

Sep 27 23:47:03 glasseyes    edtrager: I think that will be the most important discussion

Sep 27 23:47:37 avox    do we have to include Simon fomr Qt into that?

Sep 27 23:47:54 edtrager    avox:  Sure, I'll do an introductory talk.  The question is when is the best time to then have the "Open Discussion on Architecting a unified text layout engine for FLOSS systems" discussion?  Maybe later after we have all heard about vertical Pango and Graphite and Scribus, etc?  But not too late, cause we want to have plenty of time to hash it out.

Sep 27 23:47:55 pjrm    Could any of that discussion take place on the wiki, so that ppl have more time to consider things?

Sep 27 23:48:22 glasseyes    yes, I think so, we could maybe have some preliminary discussion on Sat, but everyone should be there for that

Sep 27 23:49:27 *    eimai has quit ("going to bed now")

Sep 27 23:49:38 edtrager    So maybe Sunday is the day reserved for really hashing it out about the "Unified Text Layout Engine"?  Sunday would allow our QT colleague to attend and participate.

Sep 27 23:50:03 *    thorsten (i=thb@nat/sun/x-5c24370f69fa994a) has left ##fonts

Sep 27 23:50:56 edtrager    I suppose that a big initial part of this "Unified Engine" discussion is first just listing and categorizing all the features that people would love to see available.

Sep 27 23:51:14 glasseyes    starting the discussion off on the wiki sounds like a good idea

Sep 27 23:51:19 hdu_hh    I won't be there but having the whole day on this essential topic sounds like a good idea

Sep 27 23:51:25 edtrager    glasseyes:  Yes, agree

Sep 27 23:51:58 pjrm    It's a big topic, and there are a number of difficult areas (such as what code to start from).  There are bound to be some initial disagreement about some bits.  Having things on the wiki allows finding out things in advance (e.g. looking up existing source code) and having time to consider arguments and possible solutions.

Sep 27 23:52:02 avox    hdu_hh: I think I missed your introduction?

Sep 27 23:52:54 edtrager    hdu_hh:  But maybe you will be listening on the audio?  Should this meeting also have some kind of IRC thing going on too, so remote people could at least have the theoretical possibility of asking questions or voicing their opinions?

Sep 27 23:53:16 hdu_hh    avox: I'm the OOo GSL+text guy

Sep 27 23:53:22 avox    ah yes, thanks

Sep 27 23:53:29 hdu_hh    edtrager: IRC sounds good

Sep 27 23:53:31 glasseyes    ok, edtrager, could you start off a wiki page to start the discussion and link to it from the main wiki page?

Sep 27 23:54:14 *    yosch ( has joined ##fonts

Sep 27 23:54:20 avox    hdu_hh: maybe we could even meet for a cup of tea before Boston :-)

Sep 27 23:54:28 pjrm    GSL=

Sep 27 23:55:00 hdu_hh    avox: can't make it to the Boston tea party ;-(

Sep 27 23:55:21 avox    I know, what about HH or Lübeck?

Sep 27 23:55:44 yosch    hdu_hh: me neither, hopefully there will be another one :-)

Sep 27 23:55:59 hdu_hh    avox: yup, I'm a Hamburger and aspire to become a Whopper ;-)

Sep 27 23:56:03 edtrager    glasseyes:  OK.  Hey everyone, I've got to go for awhile.  I'll leave my connection open and come back in a little while to see where things are.

Sep 27 23:56:38 yosch    edtrager: OK bye

Sep 27 23:58:16 yosch    I've updated the wiki a bit

Sep 27 23:58:41 hdu_hh    got to go... good night everyone

Sep 27 23:59:06 *    hdu_hh ( has left ##fonts

Sep 27 23:59:06 avox    nite

Sep 28 00:00:05 glasseyes    same here, I'll post the log I have and edtrager will add to it if there is any more discussion, g'night and thanks for being here everyone

**** ENDING LOGGING AT Thu Sep 28 00:00:37 2006

* glasseyes has quit ("good night/afternoon/morning")

Below is the remainder of the discussion after glasseyes logged out (posted by edtrager):

yosch      anybody interested in issue of fonts on the OLPC platform? AFAIK the OLPC building is just the next building to the summit facilities, sounds like a good opportunity to invite them / see what direction they're going...

{{{rillian yosch: seems like i18n is an even bigger problem for them }}} {{{rillian although they also don't have to support everything right away, they can do localization }}} {{{rillian s/bigger/pressing/ }}} {{{pjrm Are there interesting scripts needed? Anything outside of unicode, by any chance? }}} {{{chahibi yosch: jg is the right person to ask }}} {{{pjrm (may be relevant to graphite, if so) }}} {{{yosch pjrm: with their Nigeria plan, they're going to have a lot of fun with all the languages there :-) }}} {{{chahibi yosch: He questioned about what Fedora lack for Arabic support, I told him that the main problem is a fully featured Arabic font }}} {{{yosch chabihi: I talked to jg briefly at GUADEC, just wondered what happened since then. I'll send him a note about the layout summit. }}} {{{pjrm says 250 languages spoken in nigeria }}} {{{yosch chahibi: well, at some point Scheherazade and Lateef will problably get OFL-ed: }}} {{{chahibi yosch: Great. I hope the height problem will be solved }}} {{{yosch more according to the Ethnologue: I guess it will depend how OLPC and the local ministry of education sees things }}} {{{behdad yosch: isn't Wed oct 4? }}} {{{behdad not 3? * behdad has quit ("Leaving.") }}} {{{yosch behdad: ? }}} {{{avox behdad might be gone, but he's right }}} {{{avox :-) }}} {{{yosch avox: yep, after seeing the wiki diff I understand what he's referring to. Will fix the wiki }}} {{{chahibi yosch: have you ever contacted SIL font designers? }}} {{{chahibi yosch: the Arabic fonts look good except the lam alif ligature (لا) }}} {{{yosch chahibi: yes I have good contacts with them }}} {{{moyogo for Nigeria, they should first focus on the official languages }}} {{{chahibi yosch: if you can, please tell them about this problem. * moyogo has quit ("live's short, sleep more") }}} {{{yosch Ok, I can do that but I'm not a Arabic specialist, you should probably be the one describing the problem: feel free to send a note to the designers, contact information is available on the website }}} {{{yosch moyogo: some languages may be unofficial to some, but crucial to others. The OLPC i18n L10n strategy will have to scale to bring culturally appropriate technology to every child. }}} {{{chahibi yosch: Ok, done }}} {{{yosch chahibi: cool }}} {{{chahibi yosch: I remember SIL Arabic fonts don't work well with Openoffice }}} {{{yosch chahibi: is it the RTL features of OOo that need some work? }}} {{{chahibi yosch: to be more specific, Diacritic marks in Arabic shaping }}} {{{chahibi yosch: Some fonts work well, others don't, the diacritic marks are not displayed on the letter but after the letter breaking the word }}} {{{edtrager I find OpenOffice has some issues with diacritic marks in other scripts too. The bug I've seen for Thai is that you can type a diacritical mark like a vowel or tone mark and it doesn't appear. But if you minimize/restore in order to force a repaint, it then is there. I think it is a line height problem especially in mixed text documents. But this is different than the Arabic diacritical marks problem }}} {{{avox is OOo using their own shaper? }}} {{{zwnj yosch: which feature SIL fonts use for positioning? GPOS or the old TT table? }}} {{{yosch AFAIK there are plans for beefing up the OOo rendering CTL engine with hopefully things like font features and split cursors at some point (heard at the OOoCon 2006) }}} {{{yosch zwnj: which ones? }}} {{{yosch some fonts have OT as well }}} {{{zwnj yosch: Arabic ones, Shahrzad and Lateef }}} {{{chahibi edtrager: with some fonts, at least on Linux, "كَتَبَ" is displayed as " ك َت َ بَ " }}} {{{mrdocs chahibi: which ones work correctly ? }}} {{{zwnj chahibi: mozilla has same problem in the native (non-pango) engine }}} {{{chahibi mrdocs: Free fonts: Farsiweb fonts and ( }}} {{{zwnj chahibi: we don't have any problem with oo.o using farsi-fonts }}} {{{chahibi zwnj: Farsiweb fonts and are those I know don't have problem with openoffice }}} {{{edtrager chahibi: Yes, that's pretty bad. More the reason I say for a unified layout engine. The hard thing for inexperienced users is that it makes no sense. For example, a font might work OK for browsing the web w/ Firefox, but then doesn't work in OOo. * ccxvii has quit () }}} {{{chahibi mrdocs: most of those don't work }}} {{{mrdocs chahibi: good to know }}} {{{chahibi edtrager: Yes, DejaVu Sans and Nazli work Ok on Firefox without Pango, MS fonts don't }}} {{{edtrager chahibi: I remember too that fonts dont' work especially when trying to add diacritics in OOo }}} {{{mrdocs chahibi: is it the fonts, freetype or the layout engine ? }}} {{{chahibi mrdocs: No idea }}} {{{yosch zwnj: AFAIK the OT GPOS (along with AAT for OSX) but I guess you should check with an fc-match -v }}} {{{chahibi edtrager: Yes, same problem in KOffice. But not in Gimp for example/ }}} {{{avox would be interesting to know which app uses which shaper (pango, harfbuzz, Qt, ICU, ...?) }}} {{{edtrager chahibi: Right, Koffice uses QT engine which, in my experience, seems to have a lot of issues with Arabic. Gimp uses Pango which works pretty well. OOo uses IBM's ICU text layout written by Eric Mader. }}} {{{chahibi edtrager: OpenOffice Arabic PDF export is really bad }}} {{{mrdocs chahibi: which viewer ? }}} {{{chahibi mrdocs: Kpdf and Acrobat Reader }}} {{{edtrager avox: See a presentation I gave in 2005. A little out of date by now, but start around slide 22 I think : }}} {{{mrdocs Ok then... }}} {{{avox edtrager: tnx }}} {{{mrdocs our experience with PDF viewer is: 1. Adobe reader 2. gsview + GS 8.54 3. kpdf 4. other xpdf based readers }}} {{{edtrager My experience with OOo PDF export is that it is crucial to double-check whether your text is really set to the font that you think it is. Sometimes you'll be typing along in English using a Latin font like Vera, then you switch to Arabic and OOo is still showing "Vera" as the font, but it's really whatever Fontconfig hands it (I guess), so then you go to export PDF and its all wrong. Doesn't matter which viewer. But if yo }}} {{{edtrager u go back and highlight and specify a good (Arabic or whatever non-Latin) font manually, it works (for me, at least) }}} {{{chahibi it is all messy in Arabic }}} {{{chahibi The PDF quality is also bad }}} {{{edtrager However, my experience is limited mostly to just small snippets of Arabic mixed in English. I've only studied a little bit of Arabic, so I can't do much more than that. }}} {{{chahibi gtg }}} {{{edtrager chahibi: Bye }}} {{{chahibi edtrager: There is a typo in Arabic here : :) }}} {{{chahibi اللغة العربية }}} {{{chahibi not }}} {{{chahibi اللفة العربية }}} {{{mrdocs zz time }}} {{{mrdocs nite all }}} {{{avox nite Peter }}} {{{edtrager chahibi: Hey, thanks! Just like I said, I've only studied a little bit. when I did that slide, I knew even less than I do now. }}} {{{chahibi :) }}} {{{edtrager chahibi: Recently I bought a music CD by "Elissa" ... and my daughter (12 years old) just makes fun of me; says I just like to listen to all those Arabic love songs ... }}} {{{edtrager mrdocs: Good night }}} {{{chahibi edtrager: :) }}} {{{chahibi edtrager: Arabic songs sound different }}} {{{edtrager chahibi: Different from what ? }}} {{{chahibi edtrager: different from western languages }}} {{{chahibi edtrager: But modern songs give more importance to the video clip than to Music, really stupid }}} {{{edtrager Hey everyone, it's now just about 3 hours since the IRC meeting started, so I assume it's officially over, right? I need to go now too. Good night /day/afternoon everyone. }}} {{{chahibi bye }}} {{{avox cya }}}

Attic/TextLayoutSummitPrepLog20060928 (last edited 2013-12-03 22:42:56 by WilliamJonMcCann)