RE: Comments on the WOFF Last Call

Your explanation below does resolve (most of) the issue I raised. It does, however, have the unfortunate side effect of having an "ID" attribute that is not of "Type ID". This, it would seem to me, will likely lead to much confusion on the part of authors. At least, you have made the text be clear about the intent of the "ID" attribute in this case. For this reason I consider my comment resolved.

Steve Zilles


> -----Original Message-----
> From: Chris Lilley [mailto:chris@w3.org]
> Sent: Tuesday, February 08, 2011 6:39 AM
> To: Stephen Zilles
> Cc: www-font@w3.org
> Subject: Re: Comments on the WOFF Last Call
> 
> On Wednesday, January 5, 2011, 6:35:22 PM, Stephen wrote:
> 
> 
> SZ> 2.       In the uniqueid element in the Metadata section, there
> SZ> is a note:” The id attribute of the uniqueid element, and of
> SZ> several further metadata elements defined below, is not required
> SZ> to conform to the rules for the XML type ID; its form is at the
> SZ> discretion of the font creator or vendor.”
> 
> Yes.
> 
> SZ> It does not seem
> SZ> appropriate for a W3C spec to advocate violation of XML well
> SZ> formedness,
> 
> Indeed it would not! That was not the intent at all.
> 
> You seem to have read this text as "This attribute is of type ID, but its
> syntax is funky and not well formed" while what we were trying to say is
> "This attribute is NOT of type ID, so don't be misled by the name; legal
> values will not necessarily conform to the constraints on the ID type".
> 
> The attribute is intended to hold identifiers, such as stock numbers or
> build numbers, which are probably pre-existing and which may well start
> with numeric characters. For example
> 
> id="123-4FDW-6.27"
> 
> which would not be a legal value of type ID, but could easily be a legal
> identifier.
> 
> The spec states this:
> 
> "it is intended to allow vendors to reliably identify the exact version of
> a particular font."
> 
> and thus, in practice, the string "1.0" is likely to be fairly common.
> 
> SZ> especially since the uniqueid element is empty and
> SZ> could contain the unique ID as its value or the some other
> SZ> attribute than “id” that is less constrained might be used.
> 
> There is no constraint on an attribute *called" id.
> 
> There are constraints on attributes of *type* ID, either declared as type
> ID in the DTD or schema, or inherently of that type (for example, xml:id).
> 
> 
> SZ> Also,
> SZ> the conformance requirement summary for Metadata says, “The
> SZ> decompressed data MUST be well-formed XML” which seems to make the Note
> irrelevant.
> 
> Hopefully it should not be clear that the allowed values for @id do not
> make the content non-well-formed.
> 
> Indeed, one of the tests performed by the WOFF checker is to extract the
> XML and check that it  is well formed.
> 
> 
> Having hopefully clarified the intention, the fact remains that the wording
> we used to express it has been demonstrated to be be open to
> misinterpretation, so we need to clarify it. The following is suggested;
> would it be sufficient to resolve your comment?
> 
> Note:
> The id attribute of the uniqueid element, and of several further metadata
> elements defined below, is not of type ID. Its form is
> at the discretion of the font creator or vendor.
> 
> Tracker, this relates to
> ACTION-65: Respond to Stephen Zilles re. uniqueid
> Chris Lilley
> 
> --
>  Chris Lilley   Technical Director, Interaction Domain
>  W3C Graphics Activity Lead, Fonts Activity Lead
>  Co-Chair, W3C Hypertext CG
>  Member, CSS, WebFonts, SVG Working Groups

Received on Tuesday, 8 February 2011 16:51:02 UTC