* don't try and mix the two paradigms, user-agent and typing-by-hand
* consider a migration path from simpetext to richtext as user agents
becomre more common. This simply means that, if this is a preferred
migration path, simplemail concepts should map directly onto richtext ones
if possible.
(And maybe even further, being able to map richtext onto some kind of
SGML-for-mail DTD, so users can choose how complex they want to be in mail).
As a richtext implementor, I must say that I agree with Lee here 100%. If
it is possible to map simpletext (or whatever it is called this week :-) onto
richtext, a MIME implementor that wishes to go for full richtext support can
use the same formatting back-end for both formats, with just separate lexical
front-ends, reducing his or her workload immensely. Similarly, a smart
simpletext formatting back-end may be able cope with a subset of richtext
with a suitable front-end, providing a gradual migration path for the
MIME implementor.
Hear, hear! I was just going to ignore this whole discussion about
richtext and simplemail (because I've decided that I'm basically happy
with unformatted ASCII and I expect that's what we are going to have
for a long time from now...) and now suddenly this idea comes up of
having several formats that share an abstract formatting model.
Since it seems that SGML completely lacks such a model (leaving it up
to the DTD definer what the semantics of tags are) and other formats
vehemently disagree about the model (e.g. formatting codes
bold/italics/... versus intention codes emphasis/definition/...),
perhaps it should be a good idea to start thinking about such a model
before starting to define a format...
Delighted,
--Guido van Rossum, CWI, Amsterdam
<Guido(_dot_)van(_dot_)Rossum(_at_)cwi(_dot_)nl>