Question about Architectures and Versioning

[Peter Murray-Rust]

At 11:02 12/06/98 +0100, Michael Kay wrote:
>
>>How does one go about using Architectures to solve the
>following problem.
>>
>I don't know how to achieve this with Architectures (one of
>those concepts that is so off-putting because it has such a
>misleading name) but as a humble programmer I know how to do
>it with SAX: just write a DocumentHandler for your version 1
>document that receives the events for the version 1 document
>from the parser and generates the events for the version 2
>structure to another DocumentHandler that expects version 2
>events.

Yes. After struggling with architectures for 2-3 years I have come to the
conclusion that:
	- they are potentially extremely useful
	- we cause ourselves considerable grief if we ignore their message
	- they have not been well explained in public.
	- the syntax is not enlightening to newcomers
	- the terminology isn't easy either
	- the simple aspects of architectures are simple
	- there is no clear simple *implementation* of architectures that we can
use to 'learn by doing'
	
In practice I suspect that I use architectures by default (e.g. when I write:
<item type="float">1.23</item>
and then use a class FLOAT.class to process this

>
>Easier still if you use SAXON.

Yes. Maybe this is the implementation I am looking for.

My guess is that it would be valuable to implement some simple aspects of
architectures, particularly on top of SAX. David Megginson has a vision
here and has posted about his new tools (did it get to XML-DEV - I'm still
catching up). I think that a *communal* project to implement simple
architectures on top of SAX could be extremely useful. Perhaps we could
even contemplate using it to support Xlink - I don't know.

If it helps I would be delighted if someone wants to use JUMBO as a
platform for experimenting. JUMBO is offered as a public resource for
people to play with. Thus a SAX-SAXON-???-JUMBO composite might very well
give us a lot more insight into them.

I'd welcome a*simple* proposals along these lines. As several people have
already posted you don't have to have HyTime PIs and you don't have to have
meta-DTDs explicitly. As with all XML-DEV proposals it needs a champion. It
follows logically from SAX - since IMO any architecture is going to have to
be able to build on SAX - and I suspect it is orthogonal to much of what
else is going on. It may well help us with *implementing* namespaces.

So from my point of view, architectures are about defining an
implementation for common processing problems. If this is a fair statement,
then the sooner we address it the better :-)

	P.

Peter Murray-Rust, Director Virtual School of Molecular Sciences, domestic
net connection
VSMS http://www.nottingham.ac.uk/vsms, Virtual Hyperglossary
http://www.venus.co.uk/vhg

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)