[Edu-sig] open source admin in academia? (editorial)

[kirby urner]

On Mon, Jul 19, 2010 at 3:30 PM, kirby urner <kirby.urner at gmail.com> wrote:
> I'm becoming more aware of the fact that one
> reason universities need to charge those
> tuitions is to pay licensing fees to private
> vendors who provide them with such basic
> services as the ability to store and schedule
> classes, record student enrollment and grades,
> record instructors etc.  The catalog needs to
> be published on-line.  There might be a lot
> of extended education options, e.g. non-credit
> courses open to anyone willing to sign up.
>

Of course it takes time/energy to develop such software
no matter who is doing it.

If a university can afford a system architect and to pay
developers, fine.

I know Reed College had an ad in the paper for
Open Source Developer (PHP centric).  But that doesn't
mean the fruits of this labor are shared with a wider
community (might not be relevant).

"Open source" may just mean that the tools themselves
are open (e.g. a LAMP stack), not that anything developed
is going to escape the silo.

> Some of these proprietary programs are pretty
> old, lack features departments need, and so
> various intermediating applications grow up
> around the edges to fill in the gaps.

I interviewed this system architect from a large community
college and he talked about how their in-house people
used to run everything to do with admin (courses,
enrollment, scholarships, instructor compensation...)
using FORTRAN on a mainframe.

Over time, components were modernized, moved to
other technologies.

Just before he left, the school signed on with a major
vendor.  He said this was a result of some political
wheeling and dealing and that the in-house people
were still using their own systems, but stuffing data
into the vendor product to keep the politicians happy
in some way.  The vendor product was quite lame in
the opinion of most staffers.

> Maybe the big dino system doesn't record
> student evaluations for example, or keep track
> of which courses are in the pipeline, but still
> haven't found a place in the sun.
>

This is a real life situation I'm facing.  To make up for
what's missing in the vendor product, they have a
one-of-a-kind custom application written in FoxPro.

FoxPro has been a rather popular language in the
Microsoft world, though Microsoft has tended to be
ambivalent about it (competes with Access, is in
so many ways better than Access).  The decision
was to not commit to any VFP 10 (no more releases),
while putting most developer tools into Codeplex
(the "shared source" repository).

Some FoxPro developers decided to code a development
environment that was rather similar, in Python.  That's
Ed Leaf and Dabo.  I've been to a couple of his talks,
shared Vietnamese food in Washington DC that time.
In any case, the concepts are all familiar if you do
RDBMS.  Xbase, originally developed in connection
with some JPL satellite project (interesting lore) had
it's own non-SQL way of talking to tables though --
with SQL grafted on later.

> One would think that universities in particular,
> which pride themselves on having advanced
> knowledge of state of the art skills, would band
> together in various consortia to pool resources
> and "eat their own dog food" as it were.  A
> school that teaches medicine actually practices
> medicine (the "teaching hospital").  Shouldn't
> schools that teach computer science and
> business administration actually walk the talk
> in some way?  Maybe many of them do, I don't
> actually know.
>

That seems a boldly correct statement on the
face of it maybe, but I've been listening to the
counter-arguments.

One smart exec I know put it this way:

"a university's main mission is to prepare a large
number of students for entry level positions
in various professional walks of life,

NOT to write sophisticated software that tries to
compete with Microsoft Word -- it takes a
veritable army to write industrial grade code,

and who's got that kind of time or resources
within academia?"

> To outsource something so core to one's business,
> to pay licensing fees while not having the power
> to make design modifications, just seems more
> than a tad on the ironic side.  It's like a bank
> outsourcing everything it does around money.
>

As another co-worker put it, universities won't lean
on something so nebulous as an "open source
community" if that means there's no one on the
hook to hold accountable if something goes wrong.

This is the chief advantage of having a vendor:  if
the system breaks, there's someone specific to call.
In the eyes of your supervisors, you've done all you
need to do:  report the problem and keep the pressure
on.  People on some other payroll are responsible.

One alternative is to get into a finger pointing war
as to which component is to blame and who
the maintainer might be.  This is the stereotyped
picture of the open source world, fed by some
vendors.  If something breaks, no one knows who
to contact.  You're dead in the water without a
service contract.

>From one of Microsoft's own memos:

As far as forming a partnership with a third-party
is concerned, we've heard from a number of
large FoxPro customers that this would make
it impossible for them to continue to use
FoxPro since it would no longer be from
an approved vendor. We felt that putting
the environment into open source on
CodePlex, which balances the needs of
both the community and the large customers,
was the best path forward."

http://en.wikipedia.org/wiki/Visual_FoxPro

That's another stereotype of open source I'm
afraid:  a hodge podge of older / used technologies,
maybe on their way out, every dime extracted,
and so now given away to the community for the
die-hards to "maintain" for free (good luck to 'em).

Is that what's happening with MUMPs I wonder?

> I realize not every college or university wants to
> reinvent the wheel around something so basic,
> but I do wonder to what extent there's some
> open source sharing going on, around these core
> utilities.  Are universities so competitive they
> won't share?  So does that mean they all pay
> the same licensing fees to use the same
> private vendor offerings?
>

Putting on my idealistic hat again, I'm imagining
universities as throbbing centers of innovation.

Rather than simply point students to Facebook,
Youtube, Blogger and Flickr as ways to build one's
ePortfolio (as I was hearing about at the recent
AAPT meeting (physics teachers)), the university
itself could have it's own social networking tools.

Student organizing and collaboration would be
all that much easier because some of the brightest,
freshest minds were doing custom project
development in-house.  A senior thesis may be
increasingly something multi-media that needs
to run (as in execute).  The possibilities, for an
art scholar, would depend in part on what the
art school might provide, in the way of electronic
infrastructure.

Music schools need good instruments.  As a
metaphor, that works in computing.

Hot new ideas would germinate in the university
(like Linux did) and then feed the larger community.
The liberal arts perspective means lowering
barriers to entry across the board.  MIT's
OpenCourseware is indicative of this commitment.

> I remember Zope / Plone and SchoolTool.
> http://en.wikipedia.org/wiki/SchoolTool
>
> Is there something even more comprehensive
> that's out there, suitable for college and university
> use?  Does it come in modularized components?
> Is it an over-the-web database?
>

Nothing has come to my attention so far.

There's no GLOBAL U app written in Django, ready
for download and customization, complete with
Students, Courses, Sections, Instructors, TAs,
Scholarships, Supplies, Catalog, Users, Security...
all the myriad relational tables and modules it'd take
to turn this into a complete system, with maybe
PostgreSQL for a back end (or another).

I do think it'd be a boost for a university's reputation
to have a lot of self sufficiency around its core
business.  Students could learn about the guts of
the very systems that are used to run the university.

Of course the actual data is protected in various
ways (open source does not mean open data), but
with pseudo-data students could work on enhancing
and documenting in a collaborative environment.
The mandate to "follow your curiosity" should extend
into the heart of whatever system you're into, no?

Learning how a university works is a great lesson
in microcosm management, and could be a key
to community development across the board,
given how schools are akin to villages or towns
(with sprawling network components, given
distance education).

Learn about the guts of a university in your formative
years, and maybe you'll become a system architect
for some semi-utopian oasis in a beleaguered world
(yes, more inspiring rhetoric).

> Or do few if any universities really eat their own
> dog food?
>
> Like I say, I'm new to this business, just trying
> to get oriented.
>
> Kirby
>

My tentative conclusion so far is a lot of universities
were among the first to have mainframes and these
were put to use to run the universities, a way of
paying their own way (mainframes were and are
quite expensive).

What's happened more recently though, is as these
first generations retire, more core functions are being
outsourced to external vendors.  Large cultural tides
are at work.

Another conclusion I've reached, and maybe this is
well known in management circles, is that it would
behoove large (and smaller) institutions to chronicle
in-house lore, with an emphasis on the choices of
technology.

I'm getting some hits on Google (how self-documenting
is Google (the company)?).

This would be retrospective / historical information
and not just "eyes only" to a few executives.

A liberal arts institution, and/or a government of/by/for
the people, might aim to be especially transparent in
its operations as a matter of self definition and long
term accreditability.

Kirby

More ruminations:
http://controlroom.blogspot.com/2010/07/book-of-lore.html