Re: Lisp for the Twenty First Century: audio talk
- From: Joubin Houshyar <Sun_of_27@xxxxxxxxx>
- Date: Mon, 8 Dec 2008 18:54:12 -0800 (PST)
On Nov 28, 1:15 pm, Mark Tarver <dr.mtar...@xxxxxxxxxxxxxx> wrote:
QUOTE
I learned this from my colleagues, and in dealing
with programers from other companies, service providers, data
centers,
sys admins, API gateways, and duties of field tutoring. I didn't
think
i have very qualified expertise in what i do, but the reality i
realized is that most are far lesser than me, and that is the common
situation. That they have no understanding of basic mathematics such
as trigonometry or calculus. Most have no interest in math
whatsoever,
and would be hard pressed for them to explain what is a “algorithm”.
UNQUOTE
Actually a lot of this stuff is disappearing from UK CS courses. A
lot of programming is what Slobodan refers to as boilerplate - not
technically or math'lly demanding and requiring a factual grasp of IT
conventions and not a deep understanding of CS.
The reasons for the current state of affairs are fairly complex, but
in context of this gathering (c.l.l.) there are a few points worth
considering:
There is a huge disconnect between the ivory tower of Computer
Scientists and the vocational practitioners. The bulk of "IT"
programming is, factually and sadly, boiler plate. The precise reason
for that, as far as I have been able to determine, is that computer
science, as a discipline, has failed to solve the (very difficult)
problem(s) that face the (business) software practitioner. On one
hand, we have (at this point) requirements for ubiquitous information
processing systems, from toasters to governments. On the other hand,
we have a "science" that being very generous is at best an arts and
crafts discipline. (Open source => Guilds).
Why is there boiler plate?
Because no one from academia has managed to articulate a practical
path for the practitioner in the manner that has been accomplished for
mechanical, electrical, chemical, civil, ..., disciplines. The boiler
plate is because it is apparently either more appealing (or perhaps
easier?) to produce yet another dissertation on data structure
manipulations (or invent another variant on lisp) than to devise a
methodology for mapping the open ended set of domain semantics to the
(asserted) closed set of information processing architectures. Boiler
plate is a by product of what the best *practitioners* devise to
tackle this real world problem: frameworks. (There are as many
variants on the MVC in framework land as there are on lisp in language
land.)
And why aren't the kiddies learning math and physics?
Because the paying jobs are (were ..) in the business end of IT, and
believe it or not, managing the complexity (of the silly but present
minutia) of assembling business systems requires the brain power of a
higher percentile set. (Not the top, but still, dummies need not
apply.) Most of the best programmers I have worked with over the
years have been engineers or physicists who had jumped the boat to
"professional" programming. Certainly none of them utilized their
advanced knowledge of quantum mechanics or number theory to solve the
problems at hand. But their inherent mental agility and acuity did in
fact serve them quite well. Why? Because, believe it or not, the
problem domain is hard, for a variety of reasons. (And no:
"elegance" and "conciseness" are not necessarily helpful in context.)
The challenge to CS academics is to solve this problem. We don't
really need another language. Doing quicksort in two lines is indeed
fine, but that is not the problem that the practitioner is facing.
The problem is: how do we develop a system that is in essence the
CRUD pattern but which must be mapped to a domain that is either
subtly or substantially unique, in a barely sufficient time frame with
a highly uneven talent pool?
Boiler plate, regrettably, is the best solution at the moment. CS
academics design languages; IT stars design frameworks.
An anecdote: one week ago I met a guy who told me he had just been
laid off 15 minutes ago along with 18 of his co-workers. He was a
programmer and was browsing Border's with me. I got on to asking him
about the state of the UK video game industry and he told me it was
declining. A lot of graduates were coming out with Java and frankly
their maths and physics wasn't up to speed. What was needed by people
like Sony, he said, were maths/physics savvy people with C++ skills
and the UK was not producing them.
Now this was an industry which until recently, the UK had a commanding
presence. So we're heading into a recession with a declining skills
base where the real workers are coming from China and India and the
Western students are too often not competitive because their
scientific education is degraded. I don't see the West as maintaining
its technological dominance of the last 300 years. I also see some
very dark clouds ahead.
This is obviously necessary, if you are willing to shift your point of
view. If you are sitting on the top of the pyramid of the
international oligarch system, having wild disparity between West and
East is hardly desirable. (For a national of the West, in your case
UK, that would be an advantage ..) So clearly the strategic decision
to dumb down the education in the West (an undeniable fact) while
promoting educational excellent in the East, is a good way to address
the disparities, and over time build up a fairly uniform global cadre
of (just barely enough) educated workers to work the global
plantation. (Rest assured that as soon as the East has been yanked
forward from its historic rut, you will not be seeing hyper educated
Chinese, Indians, etc. Too much education is not really a desired
quality in workers, regardless of your coordinates.)
Mark
/R
.
- Prev by Date: Re: Coding challenge: Can you make this function more elegant?
- Next by Date: Re: Coding challenge: Can you make this function more elegant?
- Previous by thread: Coding challenge: Can you make this function more elegant?
- Next by thread: Learning lisp
- Index(es):
Relevant Pages
|
