Re: ERB meeting, 30 October 1996
>All this trickery does indeed get very fragile. There is a proposal
>that will probably be considered at the combined ISO/ANSI meeting
>the week before SGML '96 that proposes a neat solution using
>different closes as well as opens for start- and end-tags. Among
>other things, it requires that EMPTY or CONREFed elements use the
>STAGO and the ETAGC for their "start"-tag. For example, if STAGC
>were "/>" and ETAGC were ">", then "<x/>..<y>..</x> would be an
>x element with content that included an empty y element.
A good beginning (as WG8 members know, I have often requested that the
overloaded delimiter roles be divided). So I would strongly support a
revision along that line. However, if you only make empties use STAGO+ETAGC
you can't distinguish ETAGC from empty-tag-C, which is essential. EMPTY
should have its *own* delimiters, or you pointlessly rule out perfectly fine
syntaxes, including the one we just approved!
That is, the delimiters should be broken down to something like this, which
imposes no backward-compatibility limitation, but opens the road forward
completely rather than partially:
STAGO STAGC for start-tags
ETAGO ETAGC for end-tags
MTAGO MTAGC for EMPTY/conref/etc.
Likewise, the "I can't use curly quotes" problem should become a non-issue:
The language-bias issues from punning on "+" meaning "repetition" and "+"
means "addition", etc. imply the need for PLUS to be divided into INCLUSION
vs REQ; and MINUS to EXCLUSION vs. CAN-OMIT-TAG; and so on.
Basically, any time a delimiter appears in more than one 8879 production,
there's a limitation on the range of definable syntaxes, and thus a
potential problem requiring review.