RE: WOFF and extended metadata

On Tuesday, June 08, 2010 10:54 AM Sylvain Galineau wrote:
> 
> Finally, I do not want the metadata discussion to delay the review
> of the rest of the specification. I would even suggest that we
> keep the format in a non-normative appendix in the 1.0 version of
> the spec as I'd rather see WOFF the container standardized as
> soon as possible. 

Sylvain, I don't think this discussion introduces any delays, as far as W3C Development process is concerned. I believe we already reached the consensus on the WOFF container format and its structure, and any potential changes to XML content of public metadata block are not going to affect the container itself. In order to speed up the standardization process, we should work towards creating a first Working Draft for public review but it doesn't mean we should finalize everything now - we can continue the metadata discussions and make further changes while the document is offered for review by the community. I doubt the metadata discussion is holding us back in any way.

> A future version of the spec - or a different
> document - could define a normative metadata format. In the absence
> of sufficient information on the use-cases font vendors will
> want, allowing for the rendering of arbitrary XML in the entire block
> may be the rational thing to do. 

Our goal as a WG is to make every possible effort to produce a high quality document that satisfies the needs of the industry at large (both font vendors and UA implementers). I don't think that producing a document that is deficient and is known to require frequent revisions would be prudent for this WG. You were the one who advocated that the WG should work towards creating a document that doesn't need to be revised soon, and it was your suggestion that initiated the discussion about extensible metadata block. We are almost there, and we have already reached the consensus on many issues - why is this sudden change of heart?

> And remain so until a plurality of
> font vendors agree on a stable format. As much as we want to, it's
> entirely possible that we are not yet able to complete this part of
> the effort.

I believe that we as a group (and you in particular) have already stated that defining a stable format may not be practical or feasible for a variety of reasons (e.g. regulatory) - this is why we agreed to introduce an easily implementable extension mechanism for public metadata block. Once implemented, it eliminates the need for us to try and foresee the future and to nail all the XML elements and attributes now - a simple extension mechanism and clear understanding of the rules for arbitrary XML content may all we ever need to do to define a metadata block that will survive multiple generations of tools.

Regards,
Vlad

Received on Tuesday, 8 June 2010 16:07:02 UTC