Re: specification for Mozilla's SVG in OpenType proposal

On 6/02/13 7:49 AM, Sairus Patel wrote:
> I'd like to take a step back for a moment.
>
> At first glance, this proposal seems to have been developed with no
> reference to the discussions that have been taking place on this list
> re. various design decisions.
>
> For example, one of the first design decisions was whether all glyphs
> were to be represented in a single SVG doc, each glyph in its own
> doc, or a hybrid or chunking approach (multiple SVG docs each
> containing one or more glyphs). I recall the consensus being the
> first option, and this was reflected in the Adobe draft proposals.
> However, Mozilla's proposal uses the third option. Is this a
> deliberate choice and preference, or simply the way Mozilla's
> implementation happened to have done it, unaware of (or perhaps
> predating) discussions on this list?

I do remember having those discussions, although I don't remember that 
we came to a conclusion on them.  I'll try to dig up the minutes. 
*looks*  Here they are:

   http://www.w3.org/2011/10/27-svg-minutes.html

The implementation was probably done without reference to the 
discussions here; I take the blame for that for not remembering them and 
forwarding them on to Edwin when he came to implement.

> Another example is the proposed new glyphchar attribute. This goes
> against the principle that the SVG glyph descriptions are purely
> graphical, and that all semantics and metrics are to be taken from
> the usual OT tables ('cmap', 'hmtx', etc).

The glyphchar="" attribute does not replace the cmap table, it is just 
an alternative to specifying the glyph ID on that element.  (You don't 
stick both a glyphchar="" and glyphid="" on the element to define that 
mapping.  You just use one or the other.)

 From my POV I'm not sure how important it is to have glyphchar="", as 
glyphid="" does the job just as well, and since you're creating the 
whole font you of course know the cmap and the purpose of each glyph, so 
I don't think it is a burden to have to specify each glyph's contents by 
ID.  roc: does glyphchar="" buy you much convenience?

> So this proposal seems to have taken the discussion back to the
> drawing board in at least a couple of ways. This is fine if reasons
> are offered that weren't discussed before, but I don't see any.

The spec would benefit from some justifications being included as notes.

> I share Vlad's concern re. sufficient advance notice (the Adobe Tech
> Summit is this week).

Sorry for distracting you from that.  Note that there is nothing being 
decided at the SVG F2F this week.  I just wanted to bring up the fact 
that we had finally written down our views of where we think the 
proposal should end up, and to highlight the remaining differences 
between the two.

> That said, thanks for jumping back into the discussion, Cam. I look
> forward to some meaningful discussion both at the F2F (Thu Feb 7, 2
> pm Pacific Time is the information I've been given) and subsequently
> on this list.

Yes, that's the time.  (That's 9am for us, which is the start of our 
day.  We may end up starting a little later than that as we get the 
teleconference equipment set up etc.)

Thanks,

Cameron

Received on Wednesday, 6 February 2013 04:32:06 UTC