Re: Octave reputation

[Jordi Gutierrez Hermoso]

I will respond to each one of the people who wrote to me. I tend to be
opinionated, and this may come off as flamish or trollish, although I
candidly state that this is not my intention. It may also be highly
offtopic, in which case please point it out to me, and I'll stay put
and promise not to follow up with further remarks on this subject.

One small note, regarding free beer vs free speech. Since I firmly
believe that the only freedom that matters is linguistic, not
alcoholic, in all that follows I say "free" only in its freedom
sense. If I need to say "free of charge", I will say "gratis", a Latin
word that is nevertheless found in English dictionaries.

Now I respond, very long-windedly:

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote:
> Jordi Gutierrez Hermoso wrote:
> > ... very few people will even think it's immoral to use an
> > unlicensed copy of Matlab.
>
> That may be your perception, but trust me, there are many people who
> do think it is immoral,

There may be many, but judging by the amount of people who copy music
around or give proprietary software for their friends to copy, or even
by the general coolness of copyright infringement as popularised by
warez groups or the Pirate Bay, there are more people who do not
believe this kind of copyright infringement to be immoral.

The viewpoint that copyright infringement is acceptable is even more
widespread in places like India, China, Russia, Eastern Europe, and
Brazil along with the rest of Latin America, and even here in my
native México. I personally don't know anyone who will not buy copied
music CDs off the street or who thinks that anyone who buys such CDs
is committing an immoral act, or at least immoral enough to not do it.

I understand where you are coming from, though. As a US government
worker in a wealthy country where copyright infringement is less
common (but still common), it's easy to believe that the few people
who openly commit copyright infringement are committing an immoral
act. Indeed, after I came back to México after four years of
university education in Canada, I too held such beliefs. However, all
attempts to justify my beliefs of the time to my compatriots were met
with ridicule and contempt. Eventually I came to the conclusion that
my fellow Mexicans were right, that copyright law isn't keeping up
with current means of reproduction of bitstreams and that sharing what
is essentially an inexhaustible product once it's produced (sharing
bitstreams is different from sharing apples) does not take away from
anyone else. Scarcity, the most fundamental rule of economics, is
violated when it comes to modern bitstreams, unless you artificially
attempt to enforce scarcity via legal means, which is one thing
copyright law attempts to do.

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote:
> And if I might add, this sort of attitude makes the commercial
> vendors even more unhappy, and less likely to be cooperative with
> open source projects.

I have never seen a free software project endorse copyright
infringement. If proprietary software developers aren't collaborating
with free software, it must be for other reasons.

Note that I'm not endorsing copyright infringement either, for reasons
I will detail below. If I say that most people do not believe
copyright infringement to be immoral, I am simply attempting to state
the facts.

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote:
> Jordi Gutierrez Hermoso wrote:
> > Nobody will check the status of your Matlab license agreement if you
> > publish a paper in which you acknowledge your use of Matlab.
>
> But what if they do?

Why would the Mathworks do such a thing? It has to be the Mathworks
who take action, because it's almost always up to the copyright
holders to claim enforcement of their copyright. That would be
terrible PR for the Mathworks. If they did it, it would be easy to
take them to the court of public opinion and make villians out of
them, as has happened with the RIAA and their efforts to enforce
copyright on individuals ranging from little girls to elderly
grandmothers. Even a media celebrity like Weird Al has parodied this
with "Don't Download This Song":

    http://www.youtube.com/watch?v=0C93_0L2Z9c

The difference is that while the RIAA can afford for the time a bad
reputation because it has so many customers, the Mathworks on the
contrary has a relatively small and tightly-knit group of customers,
being engineers, scientists, industrialists, and academics, who could
much more easily band together and make a strong statement against
them. The resignation of the editorial board of the Topology journal
last summer comes to mind as an example of this tightly-knit group
taking action against abuses of copyright law.

What I think most likely thing the Mathworks could do is settle
privately with whoever published an article using an unlicensed copy
of Matlab. They would either offer to give them a license at a
discount price or a gratis license altogether, since the person who
published the article has already done them the service of providing
gratis marketing for their product.

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote:
> Jordi Gutierrez Hermoso wrote:
> > Indeed, if proprietary software really got paid as much as its
> > license agreements say it needs to be paid, or used as restrictively
> > as the same licenses say, nobody would use it. We all know it's
> > overpriced.
>
> Some of us do pay those prices (and we don't like being reminded of
> it).

Sorry. I personally have never seen a Matlab price tag, even when I
did use it. At first, because it was covered under a site-wide
license, or so I was led to believe. After that, because my school's
sysadmin gave me a copy to install on my personal computer. If it's
any consolation to you, I erased all traces of Matlab from my personal
computer two years ago.

I believe my experience as a student who used Matlab for several years
but never paid for it is typical. Such students are then prepared to
deal with the industry after learning its standard and can go on to
work in companies like Motorola that employ Matlab. It's the software
equivalent of giving out gratis cigarette samples to young adults in
their late teens. Giving out site-wide licenses for students or
company employees isn't charity: it's good business and an insidious
marketing ploy. I'm glad that I found my nicotine patch with GNU
Octave.

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote:
> Jordi Gutierrez Hermoso wrote:
> > Few of us have read the licenses enough to know how ridiculous their
> > terms are.
>
> Again, you are speaking for yourself only.

No, I'm not speaking for myself, since I have read some of the
licenses. I can't recall particulars about the Matlab license, but I
believe, and correct me if I'm wrong, that there are certain
restrictions on the code you produce with Matlab and how you may share
your own code.

I'm speaking about the majority of Matlab's users and indeed of
proprietary software users. Have you read in detail the Windows
license, the Matlab license, the Acrobat Reader license, and whatever
other piece of proprietary software that you have used, if any? If you
do indeed read in detail all the clickthrough agreements you may have
clicked if you use proprietary software, I can guarantee you that
you're part of a very small minority.

On 13/06/07, Tom Holroyd (NIH/NIMH) wrote
> Jordi Gutierrez Hermoso wrote:
> > Unlicensed copies of Matlab and the Mathworks mostly turning a blind
> > eye towards copyright infringements is a big reason for why Matlab
> > turned into an industry standard. The same goes for many other
> > pieces of proprietary software as industry standards, like
> > Photoshop.
>
> That's ridiculous.

Not at all. Above all, what software needs in order to survive is
users. In fact, don't take it from me. Take it from Bill Gates, in
reference to copyright violations in China: "As long as they are going
to steal it, we want them to steal ours. They'll get sort of addicted,
and then we'll somehow figure out how to collect sometime in the next
decade."

    http://news.com.com/2100-1023-212942.html

Once software has users, then there's general interest to promote
further use, to promote its formats or in the case of Matlab its
coding language. There's interest to develop it, and some of the
users, the more dedicated ones, may become developers themselves if
the software's development model allows it. The thousands upon
thousands of lines of Matlab code out there written by Matlab users
are testament to this. As far as many proprietary software users go,
once the software has enough users, a number of those users, the ones
who for whatever reason feel like they should, they will pay and the
software will have money to fund further development. This is the
shareware model that was initially popularised by computer games like
the amazing smash hit Doom, which twelve years after its release and
initial boom still has a loyal following and developers working for
it, no doubt helped by the fact that it has since been released as
free software. Matlab works almost de facto on the same shareware
model, except it prefers to label the freeloaders as "pirates".

Users really will not put up with all the clauses of license
agreements and copyright law, and the reason this has not been too
important despite modern means of reproducing data is because most of
those clauses have never been enforced. We are now seeing an attempt
to enforce those clauses with DRM, and I have yet to see any user who
seems happy with DRM once you tell them that it's designed to not let
them play their music or videos except for the one very specific way
in which the media vendor intended. This seems like a fair
representation of DRM, and the users who know it are understandably
annoyed.

The unlicensed copies around the world of Microsoft Word have helped
Microsoft make the .doc format a de facto standard worldwide. It's
difficult to estimate the actual number of Microsoft Word users for
the same reason that it's difficult to esimate the number of free
software users: many, perhaps most of the users will never report
their usage of the software, and they may get it through other means
than directly acquiring it from the software distributor. In the case
of proprietary software, they may not even report by paying for it,
since they'll use it without paying.

Something similar has happened for the Mathworks, albeit on a much
smaller scale. People want to run Matlab code that's out there in the
'net or in the textbooks, manuals or journal articles, and although
some of them will use unlicensed copies of Matlab to do it, enough of
them will pay for their copy of Matlab, especially in the richer
countries. The Mathworks may whine about "software piracy" in their
webpage, but they know that as long as people are using their
software, they can make up for it at the industrial level by making
sure the big companies are also using it. I'm about to say a few more
things about this.

On 13/06/07 David Bateman wrote:
> My company does pay the full price for all the matlab products it
> uses, and having a pirate copy on your work computer would be subject
> to discipline. We also respect all license terms,

Yes, of course. Motorola is a huge target. A copyright violation could
be very damaging to the company. The Mathworks could get Motorola into
a lot of hot water if it found out that copyright violations are
taking place there. If indeed someone at Motorola had an unlicensed
copy of Matlab, all it takes is one disgruntled employee to report it
to the Mathworks and Motorola would suffer a strong blow.

The same goes for the other big companies that use products from the
Mathworks. I have heard it said that the three big M's, Maple, Matlab,
and Mathematica get most of their revenue not from the students and
universities who are using their software, but from the big companies
who do. I'm not sure this is true, but it certainly sounds plausible.

Lawsuits 101: never, ever sue poor people or small targets if you're a
giant yourself (a recommendation, by the way, that the RIAA has broken
and is paying for it). It isn't worth the Mathworks' time to sue
smalltime students copying Matlab to their home computers from the
school network. If you're going to sue anyone, you're going to go for
the big fish. Since you know you can sue the big fish, you're also
going to make the big fish pay much higher license fees than the
smalltime student who might not even pay you anything at all anyways.

Motorola knows this too, which is why it makes sure its employees
follow each and every clause of Mathworks' license agreement, even the
stupid clauses.

Incidentally, here at CIMAT (http://www.cimat.mx), we have very little
if any software properly licensed. Yet, this hasn't stopped any of my
colleagues from heavily employing the likes of Matlab, Maple,
Mathematica, cracked copies of WinEDT, and Borland's C/C++ compiler in
their research. A lot of what I have said in this email is in response
to those who say to me, "why would I want to use free software when I
can already get a gratis copy of whatever proprietary software I want,
which is of better quality anyways?" I am writing this response also
for those people who get their gratis copies of proprietary software
not through copyright violation, but through site-wide licenses as I
once did myself and many other students and company employees are.

On 13/06/07, John W. Eaton wrote:
> Jordi, from your previous posts you seem to be a free software and
> GPL advocate, which I think is great.

Yes, indeed I am. Despite all of my ranting above and stating how
rampant copyright violations are, let me reiterate: I believe free
software is the way to go.

On 13/06/07, John W. Eaton wrote:
> But I don't think it helps to encourage people to disregard software
> licenses that don't happen to suit them, especially if you want
> those same people to respect the terms of the GPL or other free
> software licenses.

Apologies. I did not pretend to encourage people to violate the terms
of Matlab's license. Indeed, you are correct that I do not want them
to violate the terms of the GPL, although this would be difficult to
do except for the copyleft clauses which are relevant only for the
relatively small minority of Octave users who may wish to modify and
redistribute Octave for their own purposes. For the record, what I
mean to say here is that the GPL isn't a license agreement like the
license agreements in proprietary software; it's simply a license that
you accept once you exercise any of the rights that it grants (this is
clause 5 of GPLv2). I have seen some Windows programs attempt to
portray the GPL as if it were another clickthrough license agreement,
which is a bit misleading, since the GPL does not require such
clickthrough. If you reject the GPL, don't use the software, don't
attempt to study its source, and don't attempt to redistribute it. If
you accept the GPL, just exercise any of the rights that ordinary
copyright law would otherwise deny you.

Back to my intent with saying how easy it is to obtain a gratis copy
of Matlab, what I was trying to say is not that we should do it, but
rather that this perceived benefit of Octave over Matlab, that Octave
is almost always gratis, is the smallest of the benefits of Octave
over Matlab. If perhaps it's difficult to obtain this benefit for
Matlab while working in Motorola, it's not difficult to later head
home to one's own personal space and download an unlicensed copy of
Matlab online after a few minutes of dedication and a little help from
anonymous online friends. On this price tag aspect, Octave holds a
very small margin of benefit over Matlab, although few people seem to
realise this and can only think of Octave as "a free [gratis] clone of
Matlab", when in fact it's so much more than that.

The most important benefits of Octave, and I think I'm about to echo some
thoughts that jwe has previously voiced on the Octave mailing lists,
are that

    1) Absolutely ALL of Octave's source is available for
       a) study,
        b) modification,
        c) further use in other kinds of code.

    2) Direct and frequent contact with Octave's developers is
       available. Its developers are also its most frequent users and
       its harshest critics.

    3) It's available now and if computers can hold its source code
       twenty years from now and the terms of the GPL are still valid
       under the laws of the future, it will be available twenty
       years from now. As free software, it is as perennial as a good
       mathematical theorem, free from whatever shackles a particular
       company may choose to apply to it further down the line for
       whatever profit motive they may have.

As engineers, scientists, and mathematicians, this is what should
matter to us most about Octave: it conforms perfectly with the
principles under which an academic and scientific community should
collaborate. Although much of Matlab's source is visible, much of it
isn't, and its reuse in industrial code is largely forbidden while
Octave's isn't. Any copy of Matlab, even a gratis copy, will not
fully conform to the ethics of our community as long as the Mathworks
does not change the terms under which it distributes Matlab.

The best way to send a clear message to the Mathworks that their
industry standard is harmful to our community is to not use their
software altogether, not to use it without paying for it. As I have
written above, software needs users, and without users it fizzles
away.

Matlab may be "Accelerating the pace of engineering and science" by
its own slogan, but at what price? Money may be part of the price, but
it's not the most important part. Matlab is a harmful addiction of our
community, and Octave is the cure or at least is on the right path to
becoming the cure. It certainly has already become the cure for me.

Dixi, ;-)
- Jordi G. H.