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

Re: Large Scale Integration

On Sun, Nov 15, 1998 at 02:08:53PM -0600, David Kirtley wrote:

> Attendance
> Report Cards
> Class Grade Sheets
> Standardized testing results
> Participation in assorted programs and activities
> Bussing and Address information
> Heath and Immunizatoin Information
> Physical Data required for sport participation
> Student Scheduling
> Class Scheduling
> Course Planning (for students graduation requirements)
> And probably a hundred more that I have not thought of this morning.

> This is a different area than the ones I have seen in any forum of discussion.

this is what makes this mailing list particularly meaningful to me too :-)

> Any ideas?

If we use SGML or XML style markup for the data mentioned above, and agree
as to the names of the tags (META-DATA?), then anyone and everyone can try
their hands at writing ultra simple to ultra sophisticated programs handling
this data and could switch from one program to the other without having to
rethink or convert their data format.  This would allow amateur programmers
like myself to try out my ideas without forcing those who want to check out
my program to follow my particularly arrangement or sample of data... but to
use their real everyday data.  We would only need to convert our existing
proprietary or uniquely arranged data into this commonly agreed format once.


if we agreed that sport participation data would be marked up thus:
<PARENTAL_NOTE>be aware that he starts off slowly due to asthma
condition 11 oct 1998</PATENTAL_NOTE>

with this data file,  a server can send email to parents wanting to make
sure that their note was slurped in to the system, AND an html page on the
school's sport page can be automatically produced (showing only the
learner's public name not the parental notes), AND the coach could receive a
roster for taking attendance and noting points scored AND the coach would
get a printout of all the parent's notes in nice readable fonts.


The way I see it, markup tags allow us to be completely open ended about
what form the data is in;  links, memos, keywords ... whatever... 
we can impose the existence of certain tags in certain order (DTD document
type definition) but we don't have to.

the LEARNER tag, for example, can be an email address, a full name, a login
name, a SQL reference to a database, whatever ; and we don't need to ever
agree on what KIND of data in in between the LEARNER tags, we just need to
agree to use a set of LEARNER Tags to enclose whatever we each want to
consider to be a learner.  In my case, I'll be using Linux login names from
the /etc/passwd file because I put all my students on there even if they
don't use their accounts (I use the * to protect the account from being

I know some of you prefer to work with binary database files because of
increased speed... and you still could ... programs making use of such binary
files would import the text markup, and export text markup as a beginning
subroutine and end subroutine respectively (whenever original data changes)
and would use the binary data for all internal algorithms.