RE: Is it possible to process architectural instances with (sgml-parse)?

Subject: RE: Is it possible to process architectural instances with (sgml-parse)?
From: "Didier PH Martin" <martind@xxxxxxxxxxxxx>
Date: Fri, 22 Jan 1999 17:29:20 -0500
Hi Eliot

The latest version James posted on his site contains the MIF backend.
However I am not sure about the architectural form.
Concerning the DCOM object named grovea it does not contain the "archParse"
interface member.

Maybe James could give a better answer to this. James, do the latest release
contains the architectural form modifications?

Didier PH Martin
mailto:martind@xxxxxxxxxxxxx
http://www.netfolder.com

-----Original Message-----
From: owner-dssslist@xxxxxxxxxxxxxxxx
[mailto:owner-dssslist@xxxxxxxxxxxxxxxx]On Behalf Of W. Eliot Kimber
Sent: Friday, January 22, 1999 2:32 PM
To: dssslist@xxxxxxxxxxxxxxxx
Subject: Re: Is it possible to process architectural instances with
(sgml-parse)?


At 06:44 PM 1/22/99 +0000, Norman Gray wrote:
>
>Greetings,
>
>Is it possible to process a document with respect to an architectural
instance
>using DSSSL's sgml-parse function (as implemented in Jade version "1.1.1"
>SP version "1.3.1")?

The version of Jade provided with the MIF back end
(www.isogen.com/JadeMif.html) provides this function. I don't know if James
has integrated with his version of Jade (I don't think he has, but I
haven't been tracking it very closely of late).

>Finally, am I correct in presuming that architectural form processing
>is something that happens as the grove is being built, so that there are
>no architectural transformations possible on the completed grove in DSSSL.

The grove you get from Jade reflects the architectural instance. You could,
of course, implement architectural processing in DSSSL (I have to various
degrees) if you needed it.  Note that in the full abstract architecture
processing model, you have groves for both the client document and all
architectural views in memory at once, with pointers in both directions
connecting corresponding nodes.  TechnoTeacher's GroveMinder tool
implements this, as does my PHyLIS tool (albeit crudely at the moment).
Some architectural processing, such as HyTime semantics, require this level
of architecture support--simply put, you use the architectural view to
figure out what things in the client document really are, but then apply
the data processing (e.g., link and address resolution) to the client
document, not the architectural instance.

>Just by the way (admission of confusion, here), am I correct in
>understanding `architectural instance' as the DTD which results from
>architectural processing, as distinguished from `architectural document',
>which is the document into which the original document instance is
>transformed by that processing?  HyTime's A.3.1.2 seems to define the
>latter fairly clearly, but I can't now re-find the definition of the
>former in there.

No--architectural instance is the document that reflects the architectural
mapping--its document instance is the architectural instance (the document
uses the architecture's DTD as its DTD).

If you have questions about using the MIF back end with architectures, or
run into any bugs, let me know.

Cheers,

E.
--
<Address HyTime=bibloc>
W. Eliot Kimber, Senior Consulting SGML Engineer
ISOGEN International Corp.
2200 N. Lamar St., Suite 230, Dallas, TX 75202.  214.953.0004
www.isogen.com
</Address>


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


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


Current Thread