No telcon tomorrow

Hello WG,

I am traveling this week and won't be able to join the call tomorrow.

We are really close to getting the spec ready for promotion to Candidate Recommendation. There are only few nits left to do, and one of them is the one I asked you to comment on below. I think that while it makes a perfect sense to require that an encoder MUST use known table tags when processing an input file I don’t see a good reason to replicate the same requirement on the UA side - whether an entry in the table directory is encoded with a known table tag or as a combination of "custom flag" + "table tag" has no bearing on the outcome of the decoding process  - the resulting decompressed font file will have the same table entry and the same table data. Imposing a restriction on the UA side to only allow known tables encoded with known table tags seems like an artificial limit that doesn't add anything useful and only complicates the life of UA vendors that need to comply with an unnecessary conformance requirement. I am tempted to completely forego any known table tags related requirements on the UA side and only make it an AT case.
Comments?

Thank you,
Vlad


-----Original Message-----
From: Levantovsky, Vladimir 
Sent: Thursday, December 03, 2015 5:45 PM
To: public-webfonts-wg@w3.org
Subject: RE: WOFF-ACTION-193: Known tags must use known tag format

Folks,

Following up on our telcon discussion yesterday - mandating the use of known table tags requires new conformance test cases. 
I have added the "MUST" requirement for encoder to use known table tags but I am somewhat reluctant to do the same on the UA side and would like to explore and hear your opinion on the following:
 - Does it make sense to allow decoders to process an input font even if known tables are encoded *without transform* using custom table flag with the known tag that follows? Seems like this would be harmless and easy way to preserve backward compatibility.
 - Similarly, does it make sense to mandate that only those tables that have transform defined be encoded using known table tags? This would likely be a safer way to deal with transformed tables but, hypothetically, would limit our ability to introduce new transforms for new font tables yet to be defined in the future.
 - Should we simply replicate the "MUST use known table tags" for the UA without any regard for transform flags value?

Thank you,
Vlad


-----Original Message-----
From: WebFonts Working Group Issue Tracker [mailto:sysbot+tracker@w3.org] 
Sent: Thursday, December 03, 2015 5:19 PM
To: public-webfonts-wg@w3.org
Subject: WOFF-ACTION-193: Known tags must use known tag format

WOFF-ACTION-193: Known tags must use known tag format

http://www.w3.org/Fonts/WG/track/actions/193


Assigned to: Vladimir Levantovsky




Known tags must use known tag format

Received on Wednesday, 9 December 2015 05:20:19 UTC