At 2:35 PM +0000 3/15/99, Charles Lindsey wrote:
I don't like the idea of an 'initial version' - it implies that a more
complex version will follow later. Trouble is, people who implement the
'initial version' will be none too happy if a 'revised version' follows
too soon afterwards.
I think this is a good way for standards to proceed: Get consensus on
an initial version that is a simple as possible and fixes the minimal
subset of the problem.
If the standard succeeds and gets deployed, great. If there is then
interest in fixing an additional part of the problem, this can
usually be done with minimal incremental cost, and with the benefit
of the operational experience gained from deployment of the initial
version.
But in fact, what I suggest is not significantly more complex than what is
proposed already (which is itself an 'uncomplex' as anything I have seen).
It is already the case that the implementer has to keep a count of '>' at
the beginning of a line, and then to do the line wrapping without those
characters, and finally but them back after the wrapping.
What I am proposing merely requires a count of the SPACE characters at the
beginning of the line (with some allowance for TABs), to be counted AFTER
the '>' have been removed, but then treated in precisely the same way
(i.e. count them, remove them, wrap the line, and put them back).
What you suggest does sound useful. But I would still prefer to
advance Format=Flowed in its present, minimalist state, and see if it
gains acceptance. If it does, and there is interest in doing more,
I'd be happy to pursue extensions such as your proposal or proletext
or something else. (Note that your suggestion does conflict with the
space-stuffing rule.)