Gregory Murphy wrote:
>I've tried rast2sgm on DOS, but there are a number of problems:
> - It won't port as is to Unix, and I hate futzing with other people's
I hate having to fool with anyone else's code, but I tried being portable,
so most changes are probably related to #include file names.
> - It processes output from rast, not sgmls, so that means one more
> layer of processing which I'd rather avoid.
rast2sgm does use the output of SGMLS. At least the one I wrote
> - It leaves some of the rast output uncoverted, for instance, the
> attribute value keywords.
True. It looked doable, but I never got around to adding the code.
Actually, I started using OmniMark. My need for an external
program to do expand end tags went away because OmniMark
doesn't need them and can easily write an OmniMark script to
put end tags back in.
Actually, not too long after I put rast2sgm up, I got E-mail from
someone else who had modified the SGMLS or one of its
associated utilities. I had assumed that it would eventually
end up in the SGML FTP area. Unfortunately, I don't recall
the person's name and the chances of still having E-mail that
old laying around is not very good.
>The nice people at SoftQuad pointed out to me that I could obtain the
>results I desired by importing documents into Author/Editor, and then
>exporting them. But somehow I can't picture myself whiling away the
>midnight hours importing and then exporting text after text after text...
And you can do this with OmniMark. it is about a 5 or 6 line program
and can be done as a batch job, so you can do tons of them in a
single run. (Well, now that I went off and pulled the code back in,
it is 2 lines without attributes and probably a dozen to include
[log in to unmask]