On 14/06/2015, at 11:55 pm, Derek M Jones <de...@knosof.co.uk> wrote:
> Something to wake people from the dead:
Concerning student subjects in SE experiments:
people use the subjects they can afford. If you
use experienced software engineers, they expect
to be paid for their time.
One way for this to change would be if the professional
societies (like the NZ IITP, formerly NZCS) to recognise
"pro bono participation in experiments" as something to
be rewarded by say making it part of the requirements for
some industry certification.
The money to work with professionals certainly isn't
going to come from central government.
It also has to be said that if you want to tell how
readable alternative notations are, you don't WANT
experienced professionals, because then you will get
"how much does this look like Blub."
I mean, as far as I am concerned, Eclipse is vastly
inferior to NetBeans, which is vastly inferior to a
plain old Emacs-like editor. That's because I *have*
experience using Emacs-like editors that makes them
very easy for me to use, while I have never yet been
able to get Hello World going using Eclipse (but have
using NetBeans). You really don't want me as a subject
in an IDE experiment.
As for computing academics, I worked in Silicon Valley
for several years, and have been working on a program
that currently contains 340 thousand raw lines of C and
Smalltalk, and have learned a heck of a lot from it.
(Mostly about how bad I really am, sigh, and about how
putting together a good *build* process is harder than
it seems, and about how writing to standards like
C99 and Posix doesn't actually get you portable code,
and how every time a new style check is added old bugs
A colleague of mine has been working for years on a
high performance information retrieval system, with the
labour of several students as well as himself, so he's
faced team and distribution issues as well. He also
worked in the IR industry for several years, and is back
in industry for a sabbatical.
Another colleague has collaborated on the production of
an open-source library. As a principal. Admittedly,
it's of algorithmic interest, but it is actively maintained.
Yet another of my colleagues recently joined us from
Some others do more algorithmic stuff. My project has led
to some novel publishable ideas, but not enough to keep my
publication count as high as the university would like.
The open source library gets a new algorithm or two each
year, which is publishable. The information retrieval
engine gets performance boosts and applications to new
kinds of data which results in publications.
The key point again is society's reward structure.
Universities reward *publications*, not working code.
If you do something different every year, you get
more publications than if you maintain something for
several years. In fact, if you spend time writing
working code rather than publishing fluff, you will
effectively be *punished*.
Oh, here's another "reward structure" issue.
I'm involved in teaching a software engineering paper.
One year I gave the students a maintenance project to
do instead of a development project. They HATED it.
When students hate a project, they give you poor
evaluations. When you get poor evaluations from students,
you get a "are you really happy in your job?" questions
from the University. This result is pressure to give
students projects they *enjoy* rather than projects
that are *good for them*.
I still think maintenance is extremely important, so if
anyone has ideas for a software maintenance project that
might *not* cost me bad evaluations I'm very interested
to take suggestions.
You received this message because you are subscribed to the Google Groups "PPIG
To unsubscribe from this group and stop receiving emails from it, send an email
To post to this group, send an email to email@example.com.
For more options, visit https://groups.google.com/d/optout.