Re: [school-discuss] Exercise 1 revised

[vernier@vc.bc.ca (Bruno Vernier)]

On Sun, Jan 27, 2002 at 11:05:23PM +0000, Gunnar Stefansson wrote:

> 1 You're not expecting student lists, or statistics, then?

not any more, I was convinced off-list to not use student names yet because
even if they are fake, others only half-listening to us here might think we
are playing with real student data.  not good.  Once we all have secure
https service going and have a trust/authentication/authorisation system, we
can use (imaginary or real) student data

> 2 Surely the userid needs to be a valid e-mail address someplace, no? 
>   (and not always on the same system...)

good point.  Looking at the full schemas of the XML standards, they all
provide for an email element for each user, as well as an "id" which steve
took to mean "login".  Keep in mind we are starting real basic here... but
it is good that you are foreshadowing one of the upcoming issues.

> 3 Although I am indeed employed at a real-world school, I want the
>   tutor-web to be real-world independent and only live in cyberspace.
>   Hope that won't offend anyone.  Thing is, the stuff on these pages
>   is taught in several (2 :-) schools and does not belong to them :-)
>   In particular there is no classroom on the tutor-web (but classes do 
>   use it and sometimes in classrooms)

I prefer the term "learning community" to "school" which is more inclusive
of any source of education-related data.  And here you are foreshadowing the
issue of intellectual property... thanks.  I am naively assuming that at
least for the duration of these exercise, anyone and everyone is welcome to join 
in.

> 4 My program currently just spews out a file, but starting with a line 
>   containing ...Content-Type: text/html; charset=ISO-8859-1... 
>   followed by your stuff.  Hope that's OK.  (couldn't get the perl cgi
>   libraries to work for me in this context :-(

Yes, from my end I see the all important:
<?xml version="1.0"?> which indicates we are dealing with XML data and I
don't see any HTML metadata anywhere...  although some purists would insist
the content type should be set to text/xml, i think.

> 5 The domain-providers only map the domain such as tutor-web.net, not 
>   tutor-web.net/cgi-bin/t-w so I need to give you a cludge - sorry.
>   I hope I will be able to change this URL after I find out how to...

> Anyway, I think this may work:
> 
> 	http://saegreifi.raunvis.hi.is/cgi-bin/t-w
> 

Excellent issue:  this is exactly the sort of obstacles I expected in
exercise 1:  https not working, cgi permissions, incorrect headers, invalid
xml, and confusion over what to include or not.  In 2 days, we have already
had an example of each at least once!!!

> Is that what you want?

oh yes! 

this exercise 1 and its later variants (exercise 3 ...) later, are in most cases, the
only thing most of us will need to do:  indeed, once the raw data is
available, then exercise 2 and its variants can be written by only a handful of programmers
as software packages downloadable or usable via the net by all participants.

without it, one cannot participate in interopability.  In the commercial
world, you are going to increasingly here the phrase: 

"You need to upgrade to the next version of whatever_ware in order to have your learning
community interoperate ... because the upgraded version is (IMS or SIF or
Cancore or SCORM or ARIADNE or whatever) compliant"

Bruno