Lists Home |
Date Index |
On Thursday 06 March 2003 11:00, Rick Jelliffe wrote:
> > Hmmm. I think they should have just put the information in attributes and
> > been done with it. People are free to ignore attributes they're not
> > interested in.
> Err, except that then the originators would say "oh we don't think this is
> important enough for non-humans" in which case it would have been be left
> out and Daniella would be left hopeless.
But I don't think that justifies having a middle ground, let alone using
comments as that middle ground.
Why bother marking information as potentially useless? If you're already
going to the effort of including it, why not put it in an attribute somewhere
rather than in a comment?
> > What happens to this .Net config file if you strip out the comments?
> > Presumably something breaks
> Why presumably? Who know what the user wants to make of the information.
Well, the example given didn't look human readable, implying that some kind
of machine processing was performed on the comments by whatever is supposed
to read it...
> Maybe a future XML should have a few extra kinds of tags with clearly
> specified statuses in this regard: for example, a kind of tag used for
> "computer information whose format is variable and which may not be
> maintained through the document set" or "data that uses some simple
> datatype" or "information only present to make it difficult to convert into
> ASN.1" and so on.
Ahah! Now I understand!
This support for hiding meaningful information in comments where no schema
can mandate it is a bid by the infoset-haters to discredit anything that
transports infosets! :-)
> Rick Jelliffe
A city is like a large, complex, rabbit