[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
On Mon, Nov 16, 1998 at 07:52:04AM -0500, Doug Loss wrote:
> I realize that it's perhaps a little soon for you to have thoughtfully
> considered it, but do any of you have opinions on my proposal of two
> parallel development efforts for a GNOMEified gradebook and a
> comprehensive data model?
The main author of Gbook announced his intention to redesign the program
presumably to fit the XML paradigm ... which is what I think we are
leaning towards as a structure for a comprehensive data model.
I feel like taking a month off to learn about comprehensive data modelling,
before taking on the leadership role you propose. If I were to continue
doing this on my own (with lots of student help as usual), I would start by
writing down a DTD (document type definition) ... sort of like creating
fields and relations for a relational database... for every situation
brought up in this mailing list so far. Then I would translate my data
(question bank, comment/PLO bank and student lists) to use these tags, and
change my three programs (taq,jaq and jar) to slurp in and spit out the new
data form. Hopefully, this would confirm my current intuition that XML
markup is the way to go for this project.
In the unix philosophy spirit, I would want to create or promote the
creation of a quick and dirty prototype of the idea of using tags (learning
objects?) and try it out with email input and output... with procmail
filtering. E-mail is another common denominator that I think we can safely
assume everyone using this system will have access to.
So for example, I should be able to update the Gbook by sending an email
from home with the appropriate tags:
<COMMENT> good job, Bruno
While this is inconvenient to write in full, it serves to test the idea that
tags can be used to update any compliant program in a simple to understand
way. Then we can work on "email macros" like
UPDATE vernier it11 assign1 A+ good job, Bruno
which would be parsed to the above tags for input to Gbook for example.
Finally, having proved the concept and made the lowest common denominator
work (namely email input), we can encourage a thousand input methods to
flourish (barcode, voice recognition, gesture recognition, direct Gbook
entry , whatever) which would all spit out the appropriate tags.
For output, an email request macro like
SEND vernier it11 reportcard
<INSIGHT>(the set of all such request macros is what I think another person here
referred to as part of the comprehensive data model. I think I am starting
would cause Gbook or jar or any compliant program to look into the database
with all the tags in it, and spit out the relevant info via email to the
authorized requester. Initially, we can verify that it sends the correct
raw tags, and then go wild and display it in any imaginable way conceivable.
did I make any sense?