RE: WOFF and extended metadata

On Wednesday, June 02, 2010 6:59 AM Tal Leming wrote:
> 
> I've given this a proposal a lot of thought and played out it out
> across different scenarios. On the one hand, it allows for extensions.
> That is good in light of the > 1.0 issues that were discussed. On the
> other hand, it doesn't allow for localization. The existing metadata
> elements do allow for localization where appropriate, so I think that
> the extension block should as well. It also doesn't follow the same
> pattern as the existing metadata, which seems odd to me. I think these
> problems can be solved. Below is a rough draft of an idea that I think
> retains the strengths of the proposal and addresses the problems.
> 
<snip />

Thank you for your efforts to develop a common extension format. I agree that localization capabilities are important and your proposal addresses them within the scope of the original idea of enabling simple extensions mechanism.

> 
> Granted, this is verbose. It could be simplified into something more
> like this:
> 
> 	<extensions>
> 		<group>
> 			<name lang="tag">string</name>
> 			<item>
> 				<name lang="tag">string</name>
> 				<value lang="tag">string</value>
> 			</item>
> 		</group>
> 	</extensions>
> 
> However, that would break the similarities with the top-level
> structures. This could make things confusing. For example, in working
> thorough this, Jonathan, Erik and I noted that the WOFF metadata spec
> should probably say more about localization fallbacks in the top-level
> structure. If the extension structure is the same, we can have one
> global set of localization notes. If the extension structure is
> different, we have to break things apart. Clarity seems better than
> saving some lines of XML.
> 

Considering that XML will always be compressed, having few hundred bytes used for some extra lines of XML shouldn't be a problem. Clarity of the spec and consistency between top-level and extension structures are important indeed.

> Liam also brought up the idea of two separate blocks: one that UAs
> would show and one that UAs would not show. I'm curious to know what
> others think of this.
> 

I thought we agreed that any arbitrary XML (not defined in the WOFF spec) could be added to the metadata with the caveat that it won't be visible to a user via UI provided by browsers. We should explicitly mention this in the spec but I am not sure (meaning I simply don't know) if there is something else we need to do about it. Let's discuss it during the call.

Thank you and regards,
Vlad

Received on Wednesday, 2 June 2010 14:01:48 UTC