- From: Thomas Lord <lord@emf.net>
- Date: Wed, 05 Aug 2009 10:37:08 -0700
- To: Oliver Rigby <oliverrigby@gmail.com>
- Cc: www-font@w3.org
On Wed, 2009-08-05 at 10:29 +0100, Oliver Rigby wrote: > On Tue, 04 Aug 2009 16:49:02 -0700 Thomas Lord wrote: > >Then to a future implementer. An EOTL bug which > >results in accidentally rendering an EOTC font > >is quite easy to imagine. Perhaps Moz. is above > >such a mistake yet the possibility is relevant > >to evaluating the quality of the proposed Recommendation. > > This is not a reasonable concern and it can not practically happen. It seems an odd claim in that EOTL is explicitly designed as a format which a certain EOTC processor will "just happen" to process. I fully believe you if you say that someone who sets out write from scratch a strict EOTL processor has a very low probability of writing code that will accidentally process EOTC. I also believe that the ways of the world are such that future implementers are not all that likely to write their EOTL processor from scratch. They will grab a library or extract code from some other system. It is not so far fetched to imagine someone searching for libraries, finding one advertised as "supporting EOTL", and winding up with an EOTC processor. And, I also believe that the ways of the world are such that at some point, some programmer will be going through a list of bug reports. One will say: "Here is this font file. It works in IE but not in your browser." Perhaps the font file will come from a popular generator of EOTL that contains a bug. Thinking "be tolerant in what you receive" the programmer quite plausibly might relax some of the checking in in the EOTL processor - winding up with a processor for some EOTC files. If it were made very clear by the stakeholders that there are no objections to supporting EOTC, sans any enforcement features, the concern would go away. An EOTL Recommendation would still be controversial - people might not want to be risk being stuck implementing the "quirks" of IE. The controversy, though, would shift in balance quite a bit - the argument for EOTL would be very strong. -t
Received on Wednesday, 5 August 2009 17:37:49 UTC