Re: Alternative syntaxes for the prolog

Steven Pemberton <steven.pemberton@cwi.nl> writes:

> ...
>
> ... any other character [other than namestart, comment start, and
> whitespace] is available to signal the start of a prolog.
>
> ...
>
> but there is no functional reason for the "ixml", so better:
>
> [version "1.1"]
> (version "1.1")
> <version "1.1">

The observation that there is no functional reason for labeling ixml
grammars with the string "ixml" makes me think.

I wonder how you feel about title pages in books.  Waste of paper,
aren't they?  Books got along just fine without title pages or tables of
contents or running heads or page numbers for hundreds and hundreds of
years.  If anyone wants to know when a book was published, or who wrote
it, or what its title is, then surely the library card catalog will tell
them.  And what's more, there only has to be one record in the catalog,
not one for every copy of the book.  So we can avoid the tedious
situation in which every single copy of the book has to carry the same
information, at a massive cost in redundancy.

The same holds true, I think, for ixml files.  After all, if a user
didn't already know that a file contained an invisible-XML grammar, why
would they be looking at the file?

-- 
C. M. Sperberg-McQueen
Black Mesa Technologies LLC
http://blackmesatech.com

Received on Monday, 4 March 2024 19:26:22 UTC