[clean-list] Re: Matrix timings (Marco Kesseler)

[Monique Wittebrood & Marco Kesseler]

>> From: "Marco Kesseler" <M.Kesseler@aia.nl>
>
>>These issues indicate that it would be A Good Thing to open
>>source the Clean compiler. Unfortunately it seems that the code
>>generator (which is written in C, even in 2.0) will not be published.
>>Exploiting processor parallellism would certainly require some
>>changes in that part of the compiler.
>
>The problem with this open source projects are:

[...]

With respect to your arguments against open source:

- Clean 2.0 is not exactly an example of a project with great progress. And
I don't know if many open source projects are stagnating. To me, it does
not seem they do, but if they do, one has to determine whether they are
because they are open source, or for other reasons. Some companies are in
trouble too.

- No programmer is truely 'average'. I do not know exactly where I stand -
I have dome some Clean implementation work in the past -, but in my current
professional life, a part of what I do is extending, modifying and
rewriting code that has been written by others. This is not always
well-documented code. I am sure that this is true for many software
developers. Building a system from scratch is rare. Furthermore, I do not
believe that the Clean system is fundamentally more complex than for
example Linux or OpenOffice. And yes: you must know what you are doing.

- You do not have to download many different Clean versions if you do not
want to. You can stick with the one distributed by the people in Nijmegen
if you like.

- About reading the code, I happily - and wholeheartedly - agree with the
comments of rthappe (remember my remarks above about modifying code). Clean
research itself is about building comprehensible systems. The Clean
compiler written in Clean had better be comprehensible. I would certainly
find it interesting to see whether it is.

- Extending the Clean compiler is not only about fixing nasty bugs. I
intend to publish some of my wishes on this list in due time. Most of these
have nothing to do with bugs. None of them have to do with the GUI support.
But even then, and with all respect: fixing subtle bugs is not God's gift
to a select group of people. An important part of computer science is about
enabling others to fix your bugs.

- In my view, dealing with GUI issues would better be solved via a better
interface with the rest of the programming world (one of my wishes) that
already has plenty of tools for GUI development. How do I get this
interface if I can't get to the sources?

- and another reason in favor of open source: big companies are not going
to rely on software of a small company, let alone a research team if they
do not get some sort of support guarantee.

>Take an evening  and read some discussion
>in the net about Linux. You will soon recognise that nobody knows what
>actually Linux is, nor does anybody know what way Linux should go in the
>future. And for real software projects (and I think Clean is a real software
>projects) the Linux community is a kindergarten.

I am no Linux fanatic either. I use Macs at home, write Windows software in
my professional life, where we use AS/400's, Linux and NT (still) for some
heavier server duties. XP is a nice step forward I think, but mostly for
the home user, so it seems. .Net is nice for developers, but we still have
to see how it holds up against other technologies. For those who do not
know where Linux is going: do you know this for Microsoft, Intel, Sun, or
Apple? Now that MS has 'improved' its licencing scheme, some big companies
are considering moving certain tasks to Linux. Other big companies back it
via their products. You can say much about Linux, but classifying the Linux
community as a kindergarten for real software projects is well,... not
entirely realistic.

regards,
Marco