Re: ERB decision, 31 October 1996

Sign me up as also not getting what Charles is driving at here.  When 
I use text entities, it has almost always been in support of authoring,
with a hub document along the lines of 


With the entities so named typically being subject to constant revision;
so it would be surprising for the to document stay unchanged between two runs
of the parser; in fact, since the parser is mostly used for validation
and typesetting, you usually run it *because* you've been changing entities.  
I don't *think* this practice is exotic and weird... like Eve, I feel like
I'm missing something obvious.

In case it's not obvious, the ERB (and I think this group) do *not* think that
in general the SGML external text entity semantics are obviously correct for 
delivery of distributed network documents; the most recent change in the spec 
language was precisely to avoid making browsers deal with them.  But it's also 
not obvious that the semantics are not correct.