[Demexp-dev] Re: demexp DTD and CDuce use

[Serge Leblanc]

On Mon, 2004-11-01 at 13:58 +0100, David MENTRE wrote:

> I have updated demexp DTD and XML example in
> demexp--development--0.3--patch-90. However, I have only done manual
> checks and not mechanical ones, so there is probably a lot of errors and
> mistakes. The produced file passes the dtd2cduce check (i.e. 
> "dtd2cduce . srv/demexp.dtd" gives a result).
>
> I would appreciate if you could review chapter 14 of
> srv/demexp-server-book.[dvi|pdf] and give me your opinion on it, as you
> are our XML expert. Do not hesitate to criticize the use of tags and
> attributes, if you find something not nice or homogeneous (of course,
> everybody on demexp-dev is invited to do so). I think that we should
> also write an XML Schema for our DTD, in order to refine the data types
> (e.g. ids are natural integers).

It's easier to build the DTDs or Schemas from the CDuce types than reverse so It's preferable to write the CDuce codes and in second step to write the DTDs or the Schemas.

> On the CDuce side, I have compiled and installed CDuce on my
> machine. I'm also starting to understand the cduce language.
>
> You once asked me what interface I would expect from the CDuce part,
> here is a start of answer:
>
>  - open an XML demexp file, for reading or writing;
>
>  - functions to read or write the XML document, one *part* at a time
>    (e.g. a participant, a question, etc.);
>
>  - close the XML document with necessary check or closing tag.
>
> I would like to avoid the load of the whole XML document if it is
> possible.

Don't misunderstand, CDuce isn't for the moment a XML repository  and it doesn't have file handle. Writing one part at a time isn't avalable without writing all the file.
But soon CDuce will manage a persist storage (http://www.lri.fr/~benzaken/stagedea.html).

--
pub  1024D/73791C2B 2002-09-30 Serge Leblanc <address@hidden>
 Primary key fingerprint: 8E0C 0D6D E026 A278 9278  BF4F 1A93 D552 7379 1C2B

signature.asc
Description: This is a digitally signed message part