Re: Parameter entities vs. GI name groups

The problem I see with the PE rules is not that of writing a spec or
of implementing it, both of which are tasks which can or have been
done without humongous effort.  We're not talking DATATAG here.  It's
instead the problem of ever explaining the rules to the sentient being
on the street: we are rightly concerned about implementation costs, but
training costs and the general generation of ill will are problems
too.

I'd rather be saying to people "You can do it like that but it's hard
to understand" when they try something twisted, than finding myself
saying to them (or myself) "That would be quite reasonable but it's
just not allowed, for reasons I can't readily formulate" in a bunch of
other cases.

But we don't want to change this unless SGML changes too.

John Lavagnino

Received on Monday, 23 June 1997 15:36:31 UTC