[plt-scheme] Perplexed Programmers
Richard Cleis wrote:
> This assumes that the system was never automated. I find that hard to
> believe.
Of course, I was speculating based on very little actual evidence.
Besides, there are other systemic factors at work here (more below).
But depending on the history of the school district, there may have been
multiple previous systems involved - e.g. at the individual school
level. Some of those systems could have been manual, or run using
primitive systems (Excel is always fun).
See http://en.wikipedia.org/wiki/Los_Angeles_Unified_School_District for
some info about the school district in question. It's big, and
allegedly badly run in general. Because of the effects I described, a
badly run organization (or just a messy and disparate one) will not
usually find automation projects to be straightforward.
> If this is true, then why the 95M$ price tag? That's 2000 dollars per
> case (48k "certified employees".)
Certificated employees are just the ones with e.g. educational
certifications. The article mentions 90,000 employees towards the end,
and their union claims a similar number of members.
But I wasn't making any claim either way about the fairness of the price.
Deloitte was originally paid around $55m[1] (before overruns). The
system is built on SAP, and SAP was paid about $12.2 million according
to [2].
> I repeat, then: for what is the 95M$ paying?
I'm sure it'd be interesting to find out. But at that level --
considering that the school district would apparently rank #250 on the
Fortune 500 if it were a public company -- things don't work quite the
way they do in smaller scenarios, as article [2] hints at.
Such things don't help make the software development situation any saner.
That all said, it's easy to underestimate the costs of big-organization
software development when you take into account the whole cycle from
requirements gathering through successful deployment. A lot of
highly-paid people are involved in such a project, including lawyers,
salespeople, and management. Companies like SAP and Deloitte are used
because of their experience with similar-sized projects. There's a big
risk mitigation element there, and the consultants take on significant
risk, of being the one to be blamed and sued when something goes wrong,
which is starting to happen now.
> Hmmm. We have such stories, but I think we can make the typical
> approaches work better... and I am a traitor to my colleagues for
> saying so. :)
Well, just to clarify my interest in this, I find it rather galling to
see such debacles being attributed to "perplexed programmers". There
are indeed likely to be some programmers of questionable merit working
on such a project, but whose fault is that?
I think these problems are largely outside the scope of HtDP, for example.
Anton
[1]
http://www.latimes.com/news/opinion/la-ed-lausd12jun12,0,4609377.story?coll=la-opinion-leftrail
[2] http://www.excellenceinlaschools.com/news/news?id=0030