[rfc-i] Graceful degradation is key, was: Re: draft-hildebrand-html-rfc

Iljitsch van Beijnum iljitsch at muada.com
Sat Jul 14 02:34:09 PDT 2012

On 14 Jul 2012, at 2:13 , Tim Bray wrote:

> On Jul 13, 2012 1:06 PM, "Martin Rex" <mrex at sap.com> wrote:

> > It is vitally important that the document will be complete and fully
> > comprehensible when rendered in ASCII-only 72col/56lines format,

> No it isn't.  The people who really need that are a vanishingly small corner case. Why wouldn't we design for the mainstream?

I think both your positions are too restrictive. Yes, we need to make progress and not stick with something that made sense 40 years ago forever, but no, that doesn't mean we should only address the needs of those who graduated in the class of 2012.

The solution is to aim high, without premature optimization, but make sure we degrade gracefully. In some cases this is an easy win: reflowable paragraphs can easily be made to fit in the 72/56 format; the reverse, not so much. Providing bitmap alternatives for SVG and MathML elements is not that much extra work once we figure out a way to automatically display the highest quality version supported.

But in some cases we'd still have to limit ourselves to make sure we're compatible with devices (and people) that lack certain capabilities. So the difference between regular and italic text can't be semantically meaningful, because that difference isn't there on some character based terminals or in screen readers. Equations and images can only be additional information and not essential for understanding an RFC for the same reason. (Although for equations you could have an a=sqrt(b/c) formatted alternative.)

More information about the rfc-interest mailing list