Re: [As-users] XML image file format for AfterStep

Wed, 16 Jun 2004 21:52:46 +0200

Le 2004-05-24 16:39 (lundi), Sasha Vasko a =E9crit /=20
On 2004-05-24 16:39 (lundi), Sasha Vasko wrote :
> M.-A. DARCHE wrote:
> >
> >So could libAfterImage XML image format be valid XML, and always requi=
> >a single root element, such as <imgproc> ?
> I can make it optional, but not required.
> Thing is, libAfterImage handles xml in a kind of interactive mode. For=20
> example it processes first tree and displays the result, then it=20
> processes next tree and displays result, etc, etc. And it can actually=20
> get its input from stdin, which feels like a shell.

No matter how libAfterImage processes its input, if you can make it
able to deal with well-formed XML it's better.

Then we can modify all configuration files and examples to be written
using well-formed XML.

But I still don't get why you must deal with non well-formed XML. I used
to work on on-the-fly XML processing with a SAX XML parser, and it
sounds like what you are describing about tree fragments. But always
using well-formed XML never caused any problem.

> I think maybe it needs proper <!DOCTYPE > thingy, but I'm not sure how=20
> to do that stuff.

No, no need to specify any DOCTYPE for an XML document.

Specifying a DOCTYPE declaration is mandatory or strongly advised for
web pages, but this has to do with interoperability and accessibility,
not with XML well formedness and XML validity.

Next generation document types, namely RELAX NG and W3C XML Schemas,
don't use DOCTYPE declarations.

But not specifying any doctype doesn't mean that AfterStep should not
have for example an associated RELAX NG for defining how to write
<imgproc>, or any-other name, tree structure.



Marc-Aur=E8le DARCHE  <>
Association Francophone des Utilisateurs de Linux/Logiciels Libres
French speaking Linux and Libre Software Users' Association
As-users mailing list