What is the head table checksum for collections?

Levantovsky, Vladimir vladimir.levantovsky at monotype.com
Tue Feb 3 21:35:34 CET 2015

Dear AHG members,

I would like to provide a quick status update on the issues related to font collections.

As you know, the W3C WebFonts Working Group is working on the new WOFF 2.0 specification with the plans to have the recommendation ready for promotion to the W3C Candidate Recommendation in Q2 2015. Given the renewed level of interest and increased popularity of the font collection format (formerly known as TrueType collections) - one of the goals for the new WOFF 2.0 technology is to support font collections as a webfont format, which the currently used WOFF 1.0 format does not do.

However, in the process of developing the spec to enable TTC support, the WG uncovered certain areas of the OFF spec that do not adequately cover the details specific to font collections. I reported earlier on one issue that was specifically related to calculating global checksums of the ‘head’ table if more than one ‘head’ table is present in the font collection format, and there are other areas as well (such as e.g. what tables can and cannot be shared in a collection, whether newly updated font collection format allowing use of CFF outlines should also allow mixed outline formats in a single collection, etc.).

Since addressing these issues will most likely require careful consideration be given to legacy implementations and what’s been done when existing TTC resources were created, I suggest that the AHG should make a recommendation to the WG11 to start a new work item and publish an amendment to the 3d ed. OFF spec (which is about to get the final approval at the upcoming MPEG meeting). Considering the fact that the font collections are seeing the renewed interest and gaining popularity among developers (which is evident considering the most recent changes we introduced in 3rd edition text to enable use of CFF outlines in a font collection format) – addressing the remaining items and eliminating the grey areas of the spec would be a welcome change.

If you agree with this assessment – I will add this as the AHG recommendation in our report to WG11, which will soon be presented for your review and approval. Your comments and suggestions for specific items to be addressed as part of this new work are always appreciated!

Thank you,

From: mpeg-OTspec at yahoogroups.com [mailto:mpeg-OTspec at yahoogroups.com] On Behalf Of 'Levantovsky, Vladimir' vladimir.levantovsky at monotype.com [mpeg-OTspec]
Sent: Wednesday, November 19, 2014 4:47 PM
To: mpeg-OTspec at yahoogroups.com
Cc: w3c-webfonts-wg (public-webfonts-wg at w3.org)
Subject: [mpeg-OTspec] FW: What is the head table checksum for collections?

Dear AHG members,

Today, during  the WebFonts WG call a question was brought up (see below) regarding support for font collections and the meaning of the checkSumAdjustment value for font collections (and how it should be calculated). We introduced quite a few changes in the spec relevant to font collections, this is probably another one that would benefit from clarification. Please respond to this email with your comments and suggested changes. I am copying both email lists on this thread so that the members of two groups can participate in the discussion.

Thank you,

From: Roderick Sheeter [mailto:rsheeter at google.com]
Sent: Wednesday, November 19, 2014 4:39 PM
To: WebFonts WG
Subject: What is the head table checksum for collections?

It is unclear what the head checksum means for collections.



To compute: set it to 0, sum the entire font as ULONG, then store 0xB1B0AFBA - sum.


The "entire font" isn't entirely clear for a collection. Does it mean the whole collection or the whole of the individual font? If the "entire font" is the individual font within the collection, what happens if the head table is shared among fonts within the collection?

One option that could make sense would be to define it as over the entire collection, setting it to 0 for *all* head tables in all fonts in the collection, doing the sum, and then updating all of them.

FYI, I asked some font experts at Google and the feedback I got was essentially that this checksum is largely ignored for collections and what value it should have isn't well defined. If we could get it better defined that would be great.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.aau.at/pipermail/mpeg-otspec/attachments/20150203/62d2acc2/attachment.html>

More information about the mpeg-otspec mailing list