[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [school-discuss] Static vs dynamic EduML

On Sun, 2002-02-10 at 17:15, Douglas S. Blank wrote:
> This will involve the parsers being able to handle tags it wasn't designed for
> (ignoring them at worst or dynamically looking for additional functionality
> remotely at best). 
> In some sense, the "final" EduML need not be agreed upon by everyone. If
> someone needs another tag (field), they can add it. If others find it useful,
> they can take advantage of it, else just ignore it.
> Maybe a method of *decentralized* data definitions (DTDs) and well-defined
> process of (automatic?) adaption is the way to go. 


	The use of such a "mutable" definition of EduML seems very tricky to
me. If you cannot rely on the data taking precise form, it will be very
difficult to write the agents (drivers).
	Maybe the XML namespaces could be used ? The way the openoffice format
was build is very intersting (http://www.openoffice.org). They used a
lot of the existing W3C schemas and assembled them in different
namespaces (SVG, MathML, HTML, FO, etc...)
	Also, why sould we adapt other formats ? As long as they are open
formats I think we should just use them in the definition of EduML. This
would considerably lower the cost of writing the applications because a
lot of open tools already exist for those formats.

> One idea that could go hand-in-hand with this is to build into EduML the idea
> of "executable content". For example, if one added a new tag, with it would
> come a URL pointing to a definition of functionality. Or the tag could have
> code associated with it (embedded in the EduML) to actually do the functions
> associated with it.

	I'm very favorable to Xiao Gang proposition to integrate dynamic
exercises. But the different attemps to mix code with XML that I've
already seen are not very encouraging. Maybe, the definition of the
exercies representation could be a fist step. And then, add the
definition of an exercise generator format. That format would certainly
be far more domain specific (you do not generate/correct an algebra
exercise the same way you correct a chemical one or an algorithmic
exercise). But the way you represent an exercise is always the same
(with some minor changes). Maybe it could be a good first step.

	Just some ideas... dont hit me ;-)


> -Doug
> > 
> > -- 
> > 
> > XIAO Gang (肖刚)                          xiao@unice.fr
> >                http://pcmath126.unice.fr/xiao.html
> > 
> -- 
> Douglas S. Blank,       Assistant Professor         
> dblank@brynmawr.edu,          (610)526-6501
> Bryn Mawr College,  Computer Science Program
> 101 North Merion Ave,  Park Science Building
> Bryn Mawr, PA 19010 dangermouse.brynmawr.edu