Re: Attn: Tom Dickey Re: lynx-dev superscript bugs

[David Woolley]

>       Polytonic Greek. The entities defined here are not intended for
>       the representation of modern Greek text and would not be an
>       efficient representation; rather, they are intended for occasional
>       Greek letters used in technical and mathematical works."

The entities are exactly equivalent in meaning to using UTF-8, and include
ALL the characters that can be represented in UTF-8, although not all have
symbolic names defined, as general entities, in the HTML entity
definition files.  At the time that the lookup table that was being
offered is being used the following steps should have already happened:

- document character set convreted to, effectively, UTF-16;
- general entities replaced by their equivalent numeric entities;
- numeric entities converted to UTF-16.

Any attempt to discriminate between general entities and UTF-8 is reading
semantics into the page that do not exist in the formal definition of
the language - not translating unknown entities is error recovery, but
Lynx does understand the entities.  It involves a software layering violation,
as the display lookup table should only use the final UTF-16 value.

Actually, the current policy is not to define more general entities, but
to encourage use of appropriate character sets *or* numeric entities;
otherwise it becomes rather arbitrary as to which characters have names
and which don't.

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden