RE: content guidelines checkpoint 3.1

> From:	Charles McCathieNevile [SMTP:charles@w3.org]
> 
> CSS2 and SVG both support defining fonts - SVG allows pretty much any kind
> of
> effeect you can imagine, and the text is still just plain text with the
> font
> applied via a style sheet. Although SVG is only a candidate recommendation
> the CSIRO viewer works with fonts.
[DJW:] 
I was trying to avoid getting into any depth of the issues of
downloadable fonts, but, besides a general ignorance of their
existence:

- HTML and SVG are hostile environments from an 
  intellectual   property view, as fonts are 
  relatively easy to extract from the site;  
  if I want to embed a font in an intranet product, 
  using the Microsoft WEFT tool, I have to run 
  WEFT for every single customer of the product, 
  in order to give their domain name permission to 
  host the font subset; that's so clearly commecially 
  unreasonable that its not even worth suggesting;

- My experience from PDF, which is relatively benign 
  from an IPR point of view, is that people still 
  forget that the fonts that they use are not
  standard, and fail to embed them;

- From an accessibility point of view, you cannot 
  assume that browsers support more than ASCII (some 
  parts of the worlds (invalidly) treat ISO 8859/1 as
  meaning their standard Windows font!) or that they 
  can switch fonts (for SVG, you must assume that the 
  source is displayed with no entities interpreted), 
  and certainly can's assume support for downloadable
  fonts - this means that you cannot safely use maths 
  symbols and   retain accessibility and you have to 
  compromise house style rules for accessibility (some 
  companies may find that difficult to accept);

- There is a lot of misunderstanding of the difference 
  between character sets and fonts (e.g. the abuse of 
  the Symbol font to display μ for m);

- Whilst SVG does support Indic languages (where glyphs 
  are not the same as characters), I don't think CSS 
  user defined fonts do so (an example of both the Symbol 
  abuse and, probably, trying to code glyphs is 
  http://eemaata.lekha.org/newiss/ (I found this when 
  trying to find CSS fonts before - they are rare on 
  the web!);

- Unicode fonts for some languages, are rare, and so are 
  tools to recode subsets to Unicode, if they exist at all, 
  leaving the choice of the Symbol font hack or   graphics;

- I'm not sure about this one, but when I tried to 
  fetch the above example, now, IE wouldn't render it 
  before it had the fonts, and hung trying to fetch them, 
  although was OK once they were cached - obviously it 
  would have displayed ISO 8859/1 gibberish in default 
  of the fonts because of the Symbol trick, as does NS 4.5;

- and, as pointed out above, very few content authors 
  even know that the facility exists (most authors copy 
  from other authors, rather than read specs).
-- 
--------------------------- DISCLAIMER ---------------------------------
Any views expressed in this message are those of the individual sender,
except where the sender specifically states them to be the views of BTS.

Received on Wednesday, 9 August 2000 15:13:40 UTC