- From: Chris Lilley <chris@w3.org>
- Date: Wed, 2 Dec 2015 17:04:23 -0500
- To: WebFonts WG <public-webfonts-wg@w3.org>
Hello,
Minutes are at
http://www.w3.org/2015/12/02-webfonts-minutes.html
and below as text
WebFonts Working Group Teleconference
02 Dec 2015
See also: [2]IRC log
[2] http://www.w3.org/2015/12/02-webfonts-irc
Attendees
Present
Rod, David, jfkthame, Sergey, Ken, Vlad, Chris
Regrets
Chair
Vlad
Scribe
ChrisL
Contents
* [3]Topics
1. [4]open action items
2. [5]remaining public comments
* [6]Summary of Action Items
* [7]Summary of Resolutions
__________________________________________________________
<scribe> Scribe: ChrisL
<RSheeter> Chris would you like to scribe or shall I?
I was ready to (but thanks!) I'm fine this week
(discussion on DPub IG and recent font discussions, licensing
for online/offline/epub)
Vlad: sorry for the delay in getting the changes done. Now
ready for review
[8]http://www.w3.org/Fonts/WG/track/actions/pendingreview
[8] http://www.w3.org/Fonts/WG/track/actions/pendingreview
RSheeter: have reviewed the spec changes
jfkthame: found a few minor typoes, nothing major, will send to
list
action-184?
<trackbot> action-184 -- Vladimir Levantovsky to Add test for
valid collection to spec & cts -- due 2015-09-09 --
PENDINGREVIEW
<trackbot> [9]http://www.w3.org/Fonts/WG/track/actions/184
[9] http://www.w3.org/Fonts/WG/track/actions/184
Vlad: if it rejects a collection, it still passes which is
bogus so we add a conformance requirement to accept valid font
collections and do something with it
... just a placeholder pn the CTS plan
[10]https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent#mustL
oadFontCollection
[10] https://www.w3.org/Fonts/WG/wiki/TestPlan20-UserAgent#mustLoadFontCollection
(it is fine)
close action-184
<trackbot> Closed action-184.
action-187?
<trackbot> action-187 -- Vladimir Levantovsky to Incorporate
jonathans hmtx suggestion into spec -- due 2015-10-14 --
PENDINGREVIEW
<trackbot> [11]http://www.w3.org/Fonts/WG/track/actions/187
[11] http://www.w3.org/Fonts/WG/track/actions/187
[12]http://dev.w3.org/webfonts/WOFF2/spec/#hmtx_table_format
[12] http://dev.w3.org/webfonts/WOFF2/spec/#hmtx_table_format
Vlad: several changes there, lets review all of them
jfkthame: Think I am satisfied by these changes, it seems ok
Vlad: original draft had hmtx transformation mandatory, now it
is optional as discussed at f2f
close action-187
<trackbot> Closed action-187.
action-188?
<trackbot> action-188 -- Vladimir Levantovsky to Update spec
for flags, weith glyf and loca treated specially for historical
reasons -- due 2015-10-20 -- PENDINGREVIEW
<trackbot> [13]http://www.w3.org/Fonts/WG/track/actions/188
[13] http://www.w3.org/Fonts/WG/track/actions/188
Vlad: as discussed at f2f, stepping back from original hard
mandate on glyf and loca transforms, allow encoder to be more
flexible.
<RSheeter>
[14]http://dev.w3.org/webfonts/WOFF2/spec/#conform-transformedL
ocaMustAccompanyGlyf
[14] http://dev.w3.org/webfonts/WOFF2/spec/#conform-transformedLocaMustAccompanyGlyf
Vlad: so we made it possible to not transform, using a special
flag. if both set to 1, tables are not transformed.
... so gives us 0 through 2 and space for other transforms
... added to spec and placeholders to CTS plan
RSheeter: lgtm
Vlad: special case is section 5.3, added condition to say
actual transfrm 0 is optional, conditional on the other table
ie do it to both or neither
ChrisL: seems clear to me
close action-188
<trackbot> Closed action-188.
action-189?
<trackbot> action-189 -- Vladimir Levantovsky to Clarify about
shared hmtx tables, can only transform if all glyf tables match
-- due 2015-10-20 -- PENDINGREVIEW
<trackbot> [15]http://www.w3.org/Fonts/WG/track/actions/189
[15] http://www.w3.org/Fonts/WG/track/actions/189
Vlad: depends on checking for all fonts in a collection, to
look for shared tables
... change discussed at f2f
... this is also relevant to action-190
action-190?
<trackbot> action-190 -- Vladimir Levantovsky to Add conf reqt
on at and ff to test for non-transformable shared hmtx with
non-atching metrics in the two glyf tables -- due 2015-10-20 --
OPEN
<trackbot> [16]http://www.w3.org/Fonts/WG/track/actions/190
[16] http://www.w3.org/Fonts/WG/track/actions/190
Vlad: not sure how to test without making it very complicated
... see comments on
[17]http://www.w3.org/Fonts/WG/track/actions/190
[17] http://www.w3.org/Fonts/WG/track/actions/190
ChrisL: we don't test a lot of invalid bytestreams that the
decoder still has to deal with.
... happy to leave this one as an untestable (FF)
Vlad: so if we reconsider, just close the action with no change
... can keep open looking for new data
RSheeter: just close it
close action-190
<trackbot> Closed action-190.
close action-189
<trackbot> Closed action-189.
open action items
[18]http://www.w3.org/Fonts/WG/track/actions/open
[18] http://www.w3.org/Fonts/WG/track/actions/open
remaining public comments
Vlad: we still have some open issues from publiuc comment, need
to close before requesting Candidate Recommendation
... issues with how some things are worded, c-like structures
<RSheeter>
[19]https://lists.w3.org/Archives/Public/public-webfonts-wg/201
5Nov/0000.html
[19] https://lists.w3.org/Archives/Public/public-webfonts-wg/2015Nov/0000.html
Vlad: from Frédéric, some issues are marked as deferred
... some changes impact existing implementation, hence
deferred. Nice to have it clean and pure but we have to
consider the cost
... his second comment was related to uint128 datatype
... would improve spec but is a breaking change
ChrisL: does it make the spec clearer?
Vlad: no
... he wants to eliminate the chance for overlap
... not a simple improvement. Invalidates all existing fonts
RSheeter: prefer not to invalidate all deployed fonts
jfkthame_: might have been better, but benefir no longer wirth
the breakage cost
Vlad: feel the same way
RESOLUTION: do not change how uint128 works because it would
break all deployed fonts
jfkthame_: he is wanting pseudo code, not a format change
RSheeter: I like that in a spec
<scribe> ACTION: RSheeter to come up with pseudo code for the
uint base128 description [recorded in
[20]http://www.w3.org/2015/12/02-webfonts-minutes.html#action01
]
<trackbot> Created ACTION-191 - Come up with pseudo code for
the uint base128 description [on Roderick Sheeter - due
2015-12-09].
Vlad: table tags comment
... he is asking why allow a known table to be encoded as
custom tag
jfkthame_: have always wondered
<RSheeter> we all want to know ... who is then left to answer
:D
jfkthame_: we no longer rely, with the flag bits, on table tags
to see if it is transformed or not
... it is all defined by flags. That one would trigger impl
changes, does not affect existing fonts
ChrisL: would disallowing it have any impact
RSheeter: our code already does it that way
... Vlad do you have a secret encoder?
(laughs)
RESOLUTION: accept change, known tags must use known tag format
not custom tag format
jfkthame_: so that need s a new decoder test. in theory this
affect backwards compat, in practice it will not
Vlad: next deferred one is why the decoder should verify the
checksum
... need to find out exactly what he is asking there
ChrisL: is it his question that is unclar?
Vlad: no, need to follow up his references. discuss in email
ChrisL: does the mathml get rendered properly?
Vlad: no!
RSheeter: prefer to say "it is C integer division" in words
<RSheeter> even better specify what that actually means
<jfkthame_> or write it as "4 * floor( (numGlyphs + 31) / 32)"
Vlad: alternative is to keep original and add it is C notation.
Division is C integer division
<RSheeter> I like that
ChrisL: like that too
<scribe> ACTION: vlad to insert "4 * floor( (numGlyphs + 31) /
32)" [recorded in
[21]http://www.w3.org/2015/12/02-webfonts-minutes.html#action02
]
<trackbot> Created ACTION-192 - Insert "4 * floor( (numglyphs +
31) / 32)" [on Vladimir Levantovsky - due 2015-12-09].
Vlad: last comment was related to defining flag components,
bitshify vs explicit ref to bit numbers, like OT spec does
... seems reasonable
RSheeter: prefer the bitshift version better, but not a strong
opinion
... issue is they say bit 1, is that 0 or 1?
Vlad: OT spec uses 0 to 15, so it is clearly defined
RSheeter: clearer to me, but not necessarily the entire world.
either is fine
RESOLUTION: accept the comment on bit numbers
Vlad: last question is already addressed. lsb vs lsb
(adjourned)
Summary of Action Items
[NEW] ACTION: RSheeter to come up with pseudo code for the uint
base128 description [recorded in
[22]http://www.w3.org/2015/12/02-webfonts-minutes.html#action01
]
[NEW] ACTION: vlad to insert "4 * floor( (numGlyphs + 31) /
32)" [recorded in
[23]http://www.w3.org/2015/12/02-webfonts-minutes.html#action02
]
Summary of Resolutions
1. [24]do not change how uint128 works because it would break
all deployed fonts
2. [25]accept change, known tags must use known tag format not
custom tag format
3. [26]accept the comment on bit numbers
[End of minutes]
__________________________________________________________
--
Best regards,
Chris Lilley
Technical Director, W3C Interaction Domain
Received on Wednesday, 2 December 2015 22:04:24 UTC