Open source: dangerous to computing education? (opensource.com)

I did find one of Guzdial's points a bit silly (the worry that FOSS will convince parents that there's no money in programming and their kids should go be doctors or something instead), but that's a minor part of his argument. As he says in responding to DeKoenigsberg, "You're absolutely right that students cannot work on any commercial software project that they choose, but they can choose to get involved in any open source project. What I was claiming is that the actual, empirically-measured demographics of commercial developers is much more diverse than the empirically-measured demographics of open source development. You're right that the potential is much greater in open source. The reality is that it's not happening."

That's completely true, and should concern us.

Information is unlike traditional renewable and non-renewable resources
which diminish as they are consumed and suffer from over consumption.
Information does not diminish as it is consumed. Instead information grows
with wider distribution and consumption. With wider distribution and
consumption the growth rate of knowledge increases. Unlike traditional
resources rationing information does not preserve this precious resource,
but ratter as access to knowledge is reduced fewer contributors are able to
pass on their precious knowledge and knowledge is lost. Knowledge and
information does not stand in isolation.

Much of how our Information Technology has developed and will develop is
direct result of the particular nature of information. Much has changed
since the early years of mainframes, to Unix workstations, to personal
computers, to web enabled technology. Accessibility and low cost
distribution are the hallmarks of success. Unix workstations provided a
common OS interface expanding its users and contributors. Personal
computers with its common OS put computing power into the hands of the
people, enabling low cost software through the massive customer base. At
the same time proprietary software companies restricted access to the
knowledge required to build their increasing complex computers systems.
This gave advantage to the largest corporation which could assemble the
largest pool of knowledge and contributors. In areas where communication
and storage standards were not defined the largest corporation had ability
to define proprietary formats and insure these formats were propagated
through the large user base. However with the traditional proprietary model
the knowledge required to understand the systems inner workings was a
deeply held secrete, which met outside contribution was highly restricted.
The growth and economic benefits of Information Technology has presented an
increasing demanded for engineers and scientists. Deep knowledge of
proprietary licensed systems is restricted, driving contributors to explore
open systems with either optional sharing or mandatory sharing, BSD
licenses or GPL licenses. Proprietary technology may come and go, but
openly shared technology which comes out of BSD and GPL licensed solutions
will remain.

Information Technology businesses require access to engineers and scientists
trained in the deep inner workings of Technology on which Information
Technology is based. Information Technology spans virtually all industries;
industrial and mobile control, manufacturing automation, finance,
distribution, sales, communications, consumer electronics and entertainment.
I believe Universities of Science and Technology have an important role to
play meeting the business demand for skill engineers and scientists.

But the larger issue is that Guzdial, as someone whose (meta?) expertise is in education,
seems to have less useful information on what actually should be taught.

It made me recall what Dijkstra had to say about ten years ago (how he is needed
more than ever today to dispel the fog of nostrum-chasing in CS). From his "1283"
correspondence
(http://www.cs.utexas.edu/users/EWD/transcriptions/EWD12xx...):
-----
"At my former university it was a firm principle of the informatics group that we would not
teach our students how to use industrial products. The main reasons at the time were
•     the quality of industrial products was never up to academic standards, and
•     the market being as fickle as it is, the industrial product was of volatile significance only.
Later I learned how much the purposes of the University and those of industry can diverge.
Universities, believe it or not, are interested in education, but I learned of industries that were
not interested in education at all, neither in an educated work force, nor in an educated
customer base. On the contrary, they preferred a docile, brainwashed work force and
undemanding customers hooked on their products.
Another remark is that it is the task of a "leading University" to lead. In particular this means
for us that we should give society not what it asks for, but what it needs. This issue is
particularly acute for CS because their society asks for snake oil, for more of the same,
though we all know that it hardly works.
-----
If Guzdial is truly interested in education, then like the sciences that CS both emulates and
transcends (let's not forget that to Dijkstra, computer science was the most difficult branch of
applied mathematics -- the poorer mathematicians had better remain pure ones) it should
adopt the customs of transparency, collegiality, and critique are the bulwarks of winnowing
the surviving truths of intellectual inquiry from the chaff. Almost none of this is possible when
*using* proprietary software. Assuming the best possible interpretation that Guzdial is
merely defending the future industrial employment of his students, we should ask to what
ends? In a blog post two years ago, he speaks
(http://www.amazon.com/gp/blog/post/PLNKJ4EHR75FBSRY) of a recruiter describing a
desperate corporate need for programmers willing to work "far more than 40 hours per
week." No criticism, or even comment, on this appalling admission of planned exploitation
in the name of competitiveness. Note that one often unstated advantage of free software is
that it self-regulates to some extent against exploitation -- we can well do without
the race to the snake-oil that the proprietary companies feel is their only leverage to
outflank one another.

In fact, his most confident criticism seems to be that programs written by an aberrant
demographic are themselves aberrant. This is transparent nonsense. We are stronger as a
society with diversity. Presuming an innate superiority of a program written by a (diverse
demographic assortment) rather than (narrow demographic assortment) is irrelevant, and
the worst sort of vague identity politics triumphalism. Of course, having (insert excluded
demographic) participating in free software development who would otherwise feel unwelcome
is a benefit to the community, and that challenge is perhaps the only useful call to arms from
Guzdial' contribution.

Meanwhile, I call on Guzdial to address the real exploitation of the commercial software
sector. Norman Matloff's criticisms of the industry (Debunking the myth of a desperate
software labor shortage, etc) would be a good start. Matloff's uncovering of the vast turnover
in IT employees (with only 18% remaining in the field twenty years post graduation)
is a dirty little secret so apparently embarrassing to the field that it is never even mentioned
in polite CS company.

First, we have "concern" for all those people getting into computing education via open source. Well, I think a lot of people got into computing education via the microcomputer revolution - something which was a proprietary/homebrew/hobbyist hybrid - and the result of that may well have been a male predominance in education and industry, too. Why the special "concern" about open source now?

Then, we have a deliberate "compare and contrast" exercise featuring an initiative to "broaden participation in computing" and a blog post (actually extensively reviewed previously in various communities) describing what one might refer to as the "scale of diversity" (or "diversity of diversity" if you're being clever) in various open source projects: the Linux kernel top table is very male; other projects are very female. I guess the implicit argument is that top-down things are better than self-organising things, and people shouldn't trust those open source types to "do the right thing".

Finally, we get the cheap shot: "I learned that it is not acceptable to criticize religion, Santa Claus, or open source development". I know of various people who are working actively to encourage broad participation in open source development; such people clearly accept the deficiencies of their communities; many such people aren't actually "open source" people at all, but are "Free Software" people who don't claim methodological superiority over how certain companies write software - the old debate about quality - but do insist that openness is preferable. Given that there is a debate in many places about diversity, such remarks are disingenuous if not plain juvenile.

I get the impression that the various Free Software and open source movements are a thorn in the side of organisations like the ACM ("We're trying to convince students that talk is also valuable in computing.") given that people have somewhere else to turn if they're interested in developing their expertise further. When I was a subscriber to Linux Journal and they were playing fast and loose with their subscriber list (as is the American style), I'd get mailshots from the ACM (or maybe the IEEE, or maybe both). I suppose these organisations know where to get their new recruits.

By regurgitating the same old "geek folklore" soundbites from Linus Torvalds as if they were the banner under which everyone in any kind of open source or Free Software community could be bundled up and labelled as a single thing, just as one might do if "open source" were perceived as a competitive threat to one's organisation and had to be treated like another single corporate entity, I sense another attempt to demonstrate relevancy by the high-priesthood of journals and professional bodies. I just hope that the articles published in their journals entertain more rigour than their apparent outreach attempts.

> Students can get started in software development at a company by doing
> tasks that aren't directly about writing software, but are about the whole
> enterprise. These legitimate peripheral tasks serve as a stepping stone
> into the process, like writing documentation or running subjects in
> usability testing

When a computer science student interns at a company, it's almost always as a software developer. Companies assume that most engineers cannot write good English. If they wanted documentation written, they would hire an English major.

Similarly, usability testing is done by another group of people. What engineers find usable or unusable is no guide to how the general public will react. Also, it's cheaper to hire people without qualifications to do work that doesn't require qualifications.

The only environments I know of where newcomers write documentation and do quality assurance are... open source development and academia. So this argument is not only wrong, but the very opposite of the truth.

With regard to his comments about community participation... This guy apparently has not heard of kernel janitors, or all of the groups that exist to promote women in open source.

He implies that open source projects value "talk" (i.e. theoretical beauty and elegance) less than commercial projects. Whatever planet this guy is from, it's not a planet where actual commercial development takes place. If he had worked, even for a single month, at a company that was in the middle of shipping a product, he would soon learn how foolish his comments are.

He also doesn't understand how people get paid for developing and maintaining open source software. Given his astonishing ignorance of the rest of the world outside academia, this should come as no surprise.

I'm sorry if this post comes across as an ad hominem, but Guzdial's post is just incredibly misinformed. He is as far from his region of expertise as I would be if I started spouting off about (say) algebraic topology.

Among other things I know what a mean calculator is and what it illustrates. The "best" options probably depends on what you are trying to achieve. If you want to understand theoretical CS then 99.98% of commercial development is probably not worth anything.

I can see the value of "commercial" experience if you want to get paid $$$$$. I seriously doubt it helps if you want to publish something in an ACM journal, where you probably do need mathematics and do not need eye candy.