Re: Escher could have drawn it (Re: XML Schema and Entities)

[Peter Flynn <peter@silmaril.ie>]

At Tuesday, 26 June 2001, "Al B. Snell" <alaric@alaric-snell.com> wrote:

[PIs]
>Can they now be considered semi-deprecated?

I hope not...you'd cripple the publishing industry and a large number of
applications if they were blacklisted. Why would anyone want to deprecate
them?

I wrote:
>> The whole point about PIs is that they are NOT part of the
>> structure of the document.

I should have made it clear that I meant the textual structure, in the
way that a heading or a paragraph is. They're just a convenient
method of storing information for passing on to a subsequent process
in a way which does not upset the parser.

>If they're not described in the DTD/Schema they can't be validated and
>nice XML editors can't suggest which ones are legal in any situation.
Is
>this wise?

Yes, absolutely. The identifying token which comes between the first 
question mark and the first white-space is something the user makes up,
or which is defined by the application. There is no such thing as 
a list 
of legal values: they are intended to be transparent to editors and
other XML-aware processors unless explicitly programmed otherwise
as part of a specific application.  They don't need to be in any DTD or
Schema because they are a part of the document syntax, so they're
built into XML at ground level. A "nice" XML editor can of course pop
up a menu of values allowed by a specific application if the app 
designer
adds the relevant stuff in macros or whatever technique the editor uses.

The GRiF editor, for example, used (uses?) PIs extensively to allow the
user to add abritrary font styling even across element boundaries,
just like
a wordprocessor does, by storing the font changes in PIs. That meant 
they
were available to subsequent edit sessions so the document could be 
prettified for editing or printing, but had absolutely no effect 
on the normal
processing of element/attribute/entity markup.

///Peter