See also: IRC log
<trackbot> Date: 12 May 2010
<scribe> Scribe: Chris
<erik> Hi everybody.
<scribe> ScribeNick: ChrisL
hi erik
<erik> I muted, it might have been me. sorry.
is the noise better?
http://www.w3.org/Submission/2010/SUBM-WOFF-20100408/
JK: Do we do the Abstract as well?
CL: Yup
... Is the invertibility reuirement listed in the spec?
JK: Checking ....
<sylvaing> 'lossless' doesn't occur again in the document
issue: invertibility needs to be a conformance requirement
<trackbot> Created ISSUE-1 - Invertibility needs to be a conformance requirement ; please complete additional details at http://www.w3.org/Fonts/WG/track/issues/1/edit .
CL: also, "Such metadata does not affect the rendering of the font in any way" is a requirement
JK: Restated in section 6
SM: Please note in the abstract that private block does not affect rendering either
AT: Yes, add that to the abstract
JK: Third sentence is useful or not?
AT: How do we define replacing? May well replace TT/OT on the Web
SM: WOFF is not really a format
AT: Analogy between multimedia
formats and font formats, like codecs
... WOFF is a container
"This document specifies WOFF, a container for sfnt type fonts such as OpenType"
AT: that is better
<cslye> Nice clarification, I think.
(general agreement, no objections)
<sylvaing> +1
JK: Fine by me
VL: Intro explains what sfnt is and is clearer on container
<jdaggett> ChrisL: VL? VL perhaps?
maybe change "directly based on the" to "a container for"
CS: Is the compression optional?
VL: Need to be clear on what the user agent is to do
CL; a converter that outputs all uncompressed is conformant 9if ill advised) but a UA that fails to decompress is not conformant
CS: Can we only talk about the format?
CL: No, need conformance for UA as well
AT: So if some data can be compressed, a user agent MUST (rfc2119) be able to decompress it
VL: So a UA must decompress
compressed data
... can recommend to authors too
<cslye> A UA must be prepared for WOFF data to be either compressed or uncompressed?
issue: add conformance for uas and generators re compression
<trackbot> Created ISSUE-2 - Add conformance for uas and generators re compression ; please complete additional details at http://www.w3.org/Fonts/WG/track/issues/2/edit .
CL: Its implied but not explicit
AT: Table directory section
interchanges WOFF file and WOFF font. Needs to be
consistent.
... also distinguish between sfnt fonts and the WOFF made from
it
... for the original sfnt fonrt, say 'font' that makes it
clearer
... distinguish between font and WOFF file
VL: Agreed
AT: So replace WOFF font by WOFF file throughout
JK: No problem
<scribe> ACTION: Jonathan to change WOFF font to WOFF file throughout [recorded in http://www.w3.org/2010/05/12-webfonts-minutes.html#action01]
<trackbot> Created ACTION-3 - Change WOFF font to WOFF file throughout [on Jonathan Kew - due 2010-05-19].
AT: and change "compressed" to "may be compressed"
CL: Does collection imply table ordering is maintained?
JD: Specified later
VL: Easier for implementors and
users to make it look similar to OFF spec. Data types first
then file structure then specific chapters on each part
... easier for people familiar with OT and OFF
JD: Section 2 has data types defined
VL: Maybe a more elaborate verbal descriotion before the header details etc
AT: Don't folow, seems to folllow
OT pretty much
... To me, seems quite logical from a tree approach
VL: OK its fine, but WOFF header itself defines fields not described but defined elsewhere. Improve on forward referencing
CL: Another way is to link keywords to their definitions
VL: prefer to see every field in the header described right there
JK: What further description would you want? e.g. for offset
SG: So this is moving the content around, not changing it. Can we review the content first?
VL: Listed in the agenda, a few changes of layout
<erik> moving on +1
<jdaggett> moving on +1
SG: So the questions of sequencing are orthagonal to the actual content
AT: Happy to review Vlad's proposals
SM: Table data order can be
swapped
... sequence of chunks is not enforced
AT: Table directory specifies
this
... allows tables to be ordered in different ways
VL: Padding needs to be specified, as well as allowed order
AT: SFNT allows padding
JK: Blocks are mandatory, metadata must be after the sfnt and so on
AT: "followed by" could be strengthened: "immediately followed by"
CL; Or add "in that order" to the end of the sentence
AT: Third para of section 2 could be strengthened
JK: Happy to rephrase like that. No particular reason to enforce it, but no benefit to shuffling around either
AT: Want to exclude black hole padding.
<scribe> ACTION: Jonathan reword "The main body" per these minutes in section 2 [recorded in http://www.w3.org/2010/05/12-webfonts-minutes.html#action02]
<trackbot> Created ACTION-4 - Reword "The main body" per these minutes in section 2 [on Jonathan Kew - due 2010-05-19].
JK: Does say balck holes in severl places, eg "up to three bytes" so padding to longword boundaries only
JD: Good to review for black holes and put something in the intro to that effect
AT: Unadressed blocks are not
permitted except for longword padding
... helps explain how the structure works
... implementers may not be the same folk who do font
rasterisation. web server folks
SM: Lack of holes should be reflected in conformance
<Vlad> That was Sergey
CL: Conformance on the WOFF file itself, orthogonal to UA conformance
SM: Data structure should be
enforced, metadata out of place should mean the font is
rejected
... else mno UA consistency
SG: UAs need to all accept and reject the same things
AT: As its a container format, strengthening relationship wth sfnt. Waht to do with invalid sfnt?
SM: Talking of invalid WOFF files
only
... can enforce from the beginning
SM: Flavor can be either TT or OT but later says it can be something else
CL: Yes, in the table its 'for example'
AT: TT and OT are the most typiczal
Cl: AAT? Graphite?
<jdaggett> ChrisL: those formats are defined by tables, not by the format tag
AT: glyph imaging, OTTO, or line
layout streams wchhis is OT layout, TT, SIL Graphite or
others
... so flavor is really the glyph flavor
... line layout flavor is not specified in the header
CL: So does it need two fields?
AT: No
... may have multiple tables in the font, so its up to the
layout engine what it uses
JD: no need to over specify, defer to the OT and OFF specs
AT: So remove excess wording from tables into the prose
<cslye> Gotta run, everyone. Thanks.
<scribe> ACTION: Jonathan to make proposals to the mailing list for points raised on 12 May call [recorded in http://www.w3.org/2010/05/12-webfonts-minutes.html#action03]
<trackbot> Created ACTION-5 - Make proposals to the mailing list for points raised on 12 May call [on Jonathan Kew - due 2010-05-19].
AT: Could we all review the whole spec for next week and sent in a prioritised edit list
VL: Also, can we classify proposed changes as general, or technical (specific sections) or editorial (readability etc)
(general agreement)
<erik> okidoki. have a good trip!
adjourned
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/9(/9)/ Succeeded: s/VK/VL/g Succeeded: s/low/llow/ Succeeded: s/dat/data/ Succeeded: s/VL: Lack/SM: Lack/ Succeeded: s/AT;/AT:/ Found Scribe: Chris Found ScribeNick: ChrisL Default Present: [Microsoft], +1.443.895.aaaa, +0845397aabb, +1.781.970.aacc, +1.408.536.aadd, +45.04.2.aaee, ChrisL, +1.212.381.aaff, [IPcaller], +45.04.2.aagg Present: [Microsoft] +1.443.895.aaaa +0845397aabb +1.781.970.aacc +1.408.536.aadd +45.04.2.aaee ChrisL +1.212.381.aaff [IPcaller] +45.04.2.aagg Agenda: http://lists.w3.org/Archives/Public/public-webfonts-wg/2010May/0045.html Found Date: 12 May 2010 Guessing minutes URL: http://www.w3.org/2010/05/12-webfonts-minutes.html People with action items: jonathan[End of scribe.perl diagnostic output]