[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [school-discuss] SIF/XML tools for Data Transfer.
The only question I see with regard to the database in schools deals
with security. XML and OpenSSL creates a tunnel. Also, with regard to
simultaneous connections, I would want a Certificate of Authority (CA)
styled after the Globus Grid. If a need for connecting to multiple
geographic databases exists, then the database vocabularly will need to
fall under a standard.
I start to wonder if we'll see US Federal standards required at the
school level as we're starting to see at the State and Local Government
level. The NCLBA has already demanded a uniform reporting standard. How
will states roll-up their testing results from districts with disparate
database vocabularies? This evolving digital convergence doesn't seem to
be going backwards. IMHO, DC has progressed faster than someone of us
thought. Simple queries and reports in plain text may soon become
Maybe, I'm getting ahead of myself but as students get access to
computers in schools, how do we know they won't hack the main databases?
Or have someone hack them on their behalf?
Let's examine the needs in Data Transfer for School programs.
My OpenAdmin thingy is designed to run on a single Linux server for a
school division. Currently, for me, it's running on a medium performance
server (P800/1GBRam/SCSI) in one division and on Xeon class (way more than
required) in another. Several other related applications could run on the
same server and access the databases directly.
OA will export to several CSV formats and I'm working on making it export
several XML data object formats as outlined in SIF. It can import
basically only a Sask Education format for updating provincial numbers. As
part of installation there is a very poor import script to import CSV into
the student demographics.
So how can developers work together to build larger school applications?
Well, of course, they could help expand OA to add on the features that are
important to them (or fix the ugly stuff that I've created...<grin>). This
would entail knowing a bit of perl and learning a bit about the data
structures / table structures. I've tried to keep it as simple as I
could (although it looks like there will have to be some coming
up... it's just too ugly in some of the scripting).
Another alternative would be to build your own application (that hopefully
doesn't duplicate functionality unnecessarily), and then share the data
that already exists in OA. How could this sharing come about?
1) Shared access to the same database. This will raise issues with file
locking, although this could be implemented. The accessing applications
could use the mysql tcp port remotely as well.
2) The other application could have its own datastore, and simply query OA
for this information. This could be done via an XML datastream using SIF
Data Objects (or simpler variants if needs be).
3) Some other loose data coupling method.
Following the Unix way, we should build small discrete tools that can
interoperate via "plain text". If we can substitute "XML" for "plain
text", we have very much the same thing. We could then "build up" a larger
administrative system with smaller interoperating modules.
In terms of school priorities, this could be:
1) Student Demographics - a data store of student characteristics.
2) Student Attendance - a datastore of attendance records.
3) Discipline - records of poor student "decisions". <grin>
4) Report Card/Gradebook system - student assessment/evaluation.
So, for example, if someone built a custom report generator, I could feed
this software using an XML datastream of "StudentPersonal" dataobjects for
its local datastore. Now obviously much less efficient than a straight SQL
query to MySQL, but would still get the job done.
So, the basic question is... Is there currently (or will there be over the
next year) a need for applications to be fed an XML datastream from a
student demographics / attendance / discipline / gradebook / report card
If not, we don't really need things as complicated as a full ZIS system
and probably not even a BZIS(Baby ZIS) or tZIS(tiny ZIS), mZIS(microZIS)
If there IS, then a simple mZIS might be useful for controlling subscribe
and publish behavior for a few dataobjects like demographics, report
Does this strike an e-motion-al chord with anyone? (Motion = hard thinking
about data + writing code to manipulate it). Errr.... maybe too much
coffee here. Anyway you get the idea.
Open Admin Nerd