Re: [seul-edu] EduML Questions

[Bruno Vernier <vernier@vc.bc.ca>]

On Sun, Feb 06, 2000 at 05:00:18PM -0600, Eric Sandeen wrote:
> Perhaps this is a bit late, as EduML has been around in its near-final
> form for a while, but I've been looking at it from the perspective of
> using it for a data format for a gradebook application.  I wound up with
> some questions and suggestions...

EduML is ours; it is never too late to improve it :-)

> First off, what is the rationale for deciding whether to use an
> attribute, or an element, for a piece of data?  In some places, this is
> duplicated - in the <PERSON> tag, for example, much of the data is
> present as both an attribute, and as an element:

> Is there a reason for having the name data in both places?

Precisely because of the debate over which way is best, I went post-modern
and made both ways valid.  Converting from attribute to element is among the
easiest things to do with XML tools ... if it looks like the
metadata/user-content dichotomy becomes a standard, we can eliminate the
variants from EduML.

> (One rationale I've seen for choosing between attributes and elements,
> is that attributes should be "metadata" and elements should be "data" -
> i.e. attributes are for internal housekeeping type data, whereas
> elements are for the application's true chunks of data - stuff that the
> user really cares about.  However, it seems that the attribute vs.
> element data debate has provoked a lot of discussion, and I'm not sure
> anyone is clear on the distinction...)

> Anyway, not trying to throw a wrench in the works, just wondering what
> the scheme is in this respect.

XML is still young, so very young...

> I've seen different teachers use different grading schemes - some simply
> gave each task some point value, then computed (earned points)/(possible
> points) for the final grade.  Some gave percentages for each task (each
> task essentially worth 100pts), with a possible multiplier ("this test
> is worth double").  Some teachers just gave symbol grades (A, B, C...)
> and some used the category weight scheme described above.

and I have seen yet other schemes ... which is why I wanted to provide a
meta-schema in EduML and let the programmers implement specific schemas as
needed.

> I had brought up the Session/Course/Class hierarchy before, and I think
> it would be helpful to implement this in EduML...

Thank you for your contribution, which brings up something I've been meaning
to do:  should I put EduML in sourceforge so that you and others can modify
it directly as needed?

> Also, it seems a little bit odd to have the <TASKS> outside of a <CLASS>
> section - while some tasks may be duplicated across different classes,
> there are likely to be many differences, as well - perhaps the <TASKS>
> should belong to a specific class...

I agree with you;  I am running into this very problem with my current
programming of a markbook/report-card maker in zope.  I am now leaning
towards your way; make the tasks unique to each class, and people can import
other tasks from other "public" classes (oh boy, this terminology can get
confusing to programmers :-)

> Ok, well - that's a lot of questions.  I've put together a mock-up of an
> XML file at http://lager.dyndns.org/xml/demo.xml that implements some of
> this if you're interested.  I'm not an XML guru, and some of it might be
> plain wrong, and I might be reinventing the wheel.  Some of it might be
> worth adding to EduML, though - let me know what you think!

I checked it out and it looks like EduML to me (the next revision that is :-)
Thanks for you work,

Bruno