RE: About Constructions rules

Subject: RE: About Constructions rules
From: "Didier PH Martin" <martind@xxxxxxxxxxxxx>
Date: Thu, 15 Jul 1999 09:59:13 -0400
Hi Brandon,

Brandon said:
As far as (sgml-parse) goes, I believe it returns a singleton node
list containing the root node of the new grove.  You should be able to
then pass this in to (process-node-list) (or something like that),
possibly within a special mode, to do what you want.  Didier, did you
envision something that wouldn't be possible with this type of
approach?

Didier says:
I did some exploratory work for:
a) OpenJade - find news ways to add more powerful and simpler constructs,
add what is missing in the current implementation.
b) DSSSL-2 - start the thinking process about the next specs. See how we can
improve the specs from what we learn and know.

I found that the query construction rule is missing in the current
implementation and before doing any implementation I started a thinking
process to see if we can improve the construct and by the same way get
simpler constructs or scripts.

So, it is exploratory work.

There is also an other part. To propose a presentation to XML-dev this
summer. The presentation content: Show how to process a collection of
documents with DSSSL and topic maps.
a) The document collection to be processed is encoded in a topic map
document (i.e. an ISO standard: ISO 13250 in its last rounds of approval).
b) A DSSSL script process the topic map document, for each <location>
element contained in a particular topic, it uses sgml-parse and
(process-node-list) to transform the document into something printable or
displayable.

The Topic map in question represents a table of content, the table of
content then acts as the document used for batch processing a collection of
documents. To take a concrete example, at Bombardier, they have a huge
collection of documents for an airplane or a train. This huge collection of
documents is structured with a table of content. The table of content and
the documents themselves are presented in a IETM environment. the whole idea
is to structure the table of content with a topic map document, process it
with a DSSSL script and then de-facto process the whole collection of
documents. In this case, the whole collection is based on the same document
architecture and then could be processed with a single DSSSL script.

So, this is the idea and part of the marketing operation we do in OpenJade
(When I say marketing do not forget that this is voluntary work for an Open
source project). The presentation could show that using two ISO standards we
can simplify the process of creating big documents from a collection of
documents using topic maps and DSSSL.

But, working on this presentation, I got all these questions about the query
construction rule and some "what if" scenarios. But for the presentation,
I'll stick to actual DSSSL constructs implemented in Openjade and then won't
use the "query" construction rule.

During this process came a crazy idea to my mind. Let's imagine that a query
construction rule would call a new construct which is in fact a dialog
(adapted for each platforms) and that this dialog ask you for a particular
document (XML or SGML) then returns a node-list (so this new construct is
part of the query-expression). The node-list contains a singleton object
(i.e. a sgml-node) then we can do some processing on this node list like
processing all its elements. Then, in this case, we expanded the notion of
query not only to be applied solely to the original source document but also
to the user ;-). But this was part of my field of dreams :-)

Then, actually, I see no problems with the sml-parse and (process-node-list)
construct to do what I wanted to do. But I still explore if this could be
done in a better way and sharing with you help me think clearer and force me
to precise my thoughts. So, thanks for that.

In the Construction rules document, I'll update the document to reflect the
original intents of the spec designers by describing the query construction
rule as a dynamically constructed pattern match expression. I have now to
find a good example showing that using a "query" construction is in some
context is better than the "element" construction rule. I have to admit that
I lack some imagination for the example and that some help would be
appreciated. Does someone as any idea?

Thanks a lot Brandon,
regards
Didier PH Martin
mailto:martind@xxxxxxxxxxxxx
http://www.netfolder.com


 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist


Current Thread