RE: WOFF and extended metadata

> From: Erik van Blokland [mailto:letterror@gmail.com] On Behalf Of Erik
> van Blokland
> Sent: Tuesday, June 08, 2010 8:14 AM


> You're widening the scope of your criticism and shifting around. 

That's a mischaracterization. First, this is not 'criticism'. It's
feedback about a technical design. If you want me to implement it
you will have to provide me with the information I need to figure out
how it's going to work, what it does, what it doesn't do etc. And I
will definitely try to find holes and issues. It's my job. 

Second, given the total lack of context and uncertainty around the
feature, I'm not 'shifting around'. The conversation started with
people expecting just about any kind of random XML to be rendered
because 'we have no idea what's going to be in there'. Given that
file metadata is neither new nor specific to fonts, I was certainly
more than willing to narrow this down considerable. We've made
progress there but every time I ask for data, input or any kind of
context to ascertain whether the proposed solution fits all I get
is "well, how do you expect us to know ?" But as soon as I suggest
simplifying it further since we know so little I'm told that 'well,
obviously you have to have this because of all these well-known problems 
with OpenType'. Followed by no examples or references.

(Yes, I am paraphrasing for emphasis).

In such a context, I am unable to estimate how stable this format
will be in practice.


> The only failing use cases are for things that you suggested to add.

What use-cases ? What failings ? There are no use-cases. If there were,
why expect any kind of random XML to be needed ? And which failings did
I suggest exactly ? I certainly didn't suggest language-matching name
and values independently. 

> Dropping the meta data will greatly reduce the support the proposal
> has at the foundries and it would move the discussion back a year. I'm
> not sure that's a good way to go.

We're not dropping it. There is and will be a metadata block. It will
contain XML. The exact format of that XML - if one is needed - would
remain informative until market practice tells us it's ready.

>
> The elements in the current proposal are polished, 

We're not talking about them. But given that changes are being proposed
as to how they support localization, it seems we're still polishing them.
And that's good.

> have all the approval and support from all sorts of stakeholders

Implementors are not stakeholders ?

> There is no discussion about whether or not to include them.

We shall discuss the issues the members of this WG deem relevant to
carry out their work. And make the changes we collectively deem desirable
in order for this specification to be standardized and implemented.

Received on Tuesday, 8 June 2010 22:47:06 UTC