Path: news.mathworks.com!not-for-mail
From: "Paul Mennen" <nospam@mennen.org>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Submission of p-code
Date: Wed, 13 Feb 2008 00:01:04 +0000 (UTC)
Organization: Sigknowledge
Lines: 42
Message-ID: <fotc00$arf$1@fred.mathworks.com>
References: <24497039.1202489169219.JavaMail.jakarta@nitrogen.mathforum.org> <foif46$17n$1@fred.mathworks.com>
Reply-To: "Paul Mennen" <nospam@mennen.org>
NNTP-Posting-Host: webapp-03-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1202860864 11119 172.30.248.38 (13 Feb 2008 00:01:04 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Wed, 13 Feb 2008 00:01:04 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 869812
Xref: news.mathworks.com comp.soft-sys.matlab:450992


"Helen Chen" <helen.chen@mathworks.com> wrote in message 
> You are welcome to submit your p-code as long as it is
> accompanied by the m-file. As someone else posted on this
> thread, our community value is for learning and teaching
> others how to use MATLAB. If you can't see how the code is
> written, the file is of little use. 

I disagree with this sentiment. Sometimes there can
be value in sharing p-code. For example I have submitted
a plotting routine (called plt) in pcode format and it has
been on the file exchange for years now. Many users of my
plotting routine have emailed me to say that plt has
revolutionalized the way they use Matlab, so I think it
has value. Once it was even chosen as "pick-of-the-week"
and in fact the latest version is far more useful than
when it earned this honor.

Many .m files are included in the plt release, but the
goal of these files is to teach one how to use plt, not
how it works internally. 

I have have no commercial interest in promoting plt and so
there is no impediment to distributing the .m source.
In fact when a user complained about this I often sent them
the source, and usually lived to regret it. There are
several reasons for this. Foremost is that most attempted
modifications are misguided. Sometimes the desired feature
is already available or there is an easier way to accomplish
the desired end. (I'd rather hear from the user about
what they would like to plot and how they want it to look).

Secondly, even if the user has a good enhancement idea,
they are rarely successful at implementing it. This is
because I developed plt over several years and it has grown
into a complex, and tricky to modify piece of code (even
for me). I think this is not unusual for advanced GUI
driven interfaces. Once they modify the code, it inevitably
doesn't work and they come back to me to explain why. (So
again I would prefer to make the modifications myself).

~Paul