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

Re: [school-discuss] Creating the 'Killer App'

On Sat, Feb 09, 2002 at 10:48:32PM -0800, Colin Dellow wrote:

> We've got the parts, why not connect them into a functional intranet?

sounds like the "integrated suites" from the mid 80's :-)

> Is there anyone else out there interested in this sort of thing?
> Developing it?  Testing it?

If one has total control over what software packages gets deployed, then it
would be interesting to have an integrated suite for education 

Otherwise, such a suite would  hold interest in so far as there are
"software drivers" that interface the "legacy or soon to be legacy" software
packages with the proposed "integrated suite".

How many drivers would be needed?  well, just like hardware drivers, we will
need as many software drivers as there are distinct legacy software to
interface with.

For example:  in a school that uses program XYZ to keep track of student
demographics and attendance , which itself interfaces with the school board
office and the minitry of education once a year in a format that changes
from year to year ;  we would want a software driver that automatically
extracts the student data to be used as classlists in our integrated suite
and then we would want to send back attendance reports and final grades at
the end of each term through a different software driver going the other way.

How many different permutations are there of XYZ with our open source
"integrated suite"?  Lots.  Now if we consider the easily predictable
continuation of the trend that there are/will be N good "integrated
suites" developped ; now we can multiply the number of software drivers
needed by N because each of those will need its separate driver to interface
with its unique data standard.  And then we may want to operate several open
source integrated programs in the same school because of some unique
features in each one ... and we need drivers between each of THOSE open
source programs as well.

you get the point:  we want, we need, we must have a common XML standard to
make the task of writing drivers (let's call them Eduml drivers for now)
more systematic and minimally repetitive.  This is not unlike deciding on
standards for railroad tracks before building locomotives ... sure we can
retrofit the locomotives, but it is more work and expense.

If we were sure that the proposed integrated suite would take over the
planet a la "windows", then sure, I'm in. Meanwhile, writing software drivers 
is not much fun and often for very small niche situations which is why I think
it has not been done much so far.  

If we developed our community of teacher/programmers' skills at software
driver construction, we would eventually get usuable integrated suites built
from what we have now, and then the confidence, and experience necessary to
simply replace age-ing legacy software with open sourced improvements in much
the way that some people replace NT servers with Linux ; in a way that no
one notices the difference, except for some occasional improvements.