At 3:55 PM 9/4/93 -0400, John C Klensin wrote:
basically assume that composing and display agents are going to be
required for all functions except debugging (and nerd-nerd
communication), then it is rational to make the language richer and let
it support a broader range of specifications.
Seems to me like there are three goals:
1. Provide enough "rich" to be useful.
2. Make the language easy to implement, so that it gets widely implemented.
3. Keep the result readable (writeable) without a parser (composer).
It would seem to me that 1 & 3 both argue for a more powerful language.
(The more power you have, the fewer directives you need to do what you
want, and so the easier the raw text is to read.)
So I don't see that saying that 3 is wishful thinking necessarily frees you
to use a more powerful language; the power restriction comes from 2, not 3.
Anyway, I don't think it matters at all whether enriched uses << or <lt>.
The marginal differences in acceptability to non-MIME users and difficulty
to MIME software authors are so tiny as to be beneath notice.
Steve Dorner, Qualcomm Inc.