Re: "proprietary", was Re: ada on linux

[Robert Dewar <dewar@gnat.com>]

In article <anhY4.1121$M72.360470@news.pacbell.net>,
  tmoran@bix.com wrote:
> >The license does NOT permit any redistribution.  Microsoft
> I think you mean "the kind of license typically given to an
end
> user allows him to run, but not redistribute, the program".

> And
> other licenses, typically given to retailers, allow
> redistribution

That's just completely wrong. You are inventing, and you
are inventing wrong. Don't guess and publish it as fact :-)
A retailer buys a copy of a program like Power Point. Now
under 117, the retailer owns a copy of the progrma (not the
program itself but a copy). The retailer is free under the
copyright law to sell this copy to someone else (if you
buy it, you inherit this right, and can then sell your
unopened copy to someone else.) This right comes from the
copyright law, not from any license. In fact the vendor
never executes any license agreement with regard to this
particular copy of Power Point (there may be a separate
sales contract, but that's different). If you sell your
unopened copy, you also never execute a license agreement
(the shrink wrap license comes into effect only when the
shring wrap is opened).

> and still others given to system vendors may
> allow a variety of different things.

Yes, that's right, see the Microsoft trial for many details
about such licenses.

>  The point is, your
> statement
> sounds like a general one about licenses (or at least
"proprietary
> software" licenses) but in fact it's a specific statement
about
> just one kind of license, and in that way misleading.

I am talking about end user licenses, and have always made
that clear. yes, system vendors do have other licenses.
Retailers do not, that's just a Moran invention :-)

>   If you receive a cake, are you free to redistribute it only
in
> its entirety,

For sure in its entirety, which is in practice the critical
issue, and is what distinguishes it from proprietary products
like CLAW. I realize that you are trying to minimize the
difference here Tom, but it is really a VERY important
difference, and trying to obfuscate this fact does not
change the fact that there is for example a very big difference
between CLAW and GtkAda here.

> Does it allow you to examine the software, decide it has
> a very low probability of bugs, and then to redistribute it
> backed with a warranty?

Yes, of course. Why are you asking questions like this whose
answer should be obvious if you read the license. These
questions seem designed to obfuscate.

> Your "absolutely free" is in fact a freedom
> with very definite

Noone said you are absolutely free to do anything you want.
That's an invention on your part.

> and sometimes onerous, restrictions, and in
> that way misleading.

Onerous in one case in particular, which is the case where
a proprietary software vendor wants to redistribute a
derived work in proprietary form. For example, I quite
see that the CLAW folks might like to use GtkAda to
make a proprietary version of CLAW for non-NT targets.
Well they definitely cannot do this.


>   Is it really practical to offer GPLed software with source
> code
> and support like "We offer no-cost support for this product,
> as
> long as you have made no changes.  If you have changed
anything,
> it makes it harder for us to help you, so you must pay $xxx
for
> support.

Yes, it is practical, and basically that is the situation at
ACT, we support GNAT as we distribute it, and will fix any
problems that are found. If you want us to guarantee to fix
problems in versions you have modified, we can do this but
we definitely write specialized (more expensive) support
contracts in this place.

What on earth would make you think it is impractical? It is
the only approach that makes sense.

> If you say you've made no changes, and we spend time
> helping you, and it turns out you have in fact made changes,
> then
> you must $yyy for the time we've spent."  I suspect that in
> practice, ie, "effectively", you will find that you *must*
charge for
> support, or you'll lose a lot of time and money giving free
> support to some people who create problems and then lie about
> the source of those problems.

You know nothing about distributing GPL'ed software and
supporting it, because you have never done it. I suggest
people take the actual experience of those who DO distribute
GPL'ed software over the suspicions of someone with no
experience at all.

I quite understand the kind of arguments people like Tom (CLAW)
use to justify choosing to go the proprietary route, but these
arguments are often misinformed, as in this case. This does not
mean that it is wrong to choose the proprietary route, but
certainly uninformed suspicions should not be the basis for
the decision.

> >Certainly I understand that you (or rather the company you
> >work for) have decided to use restrictive licenses, and that
>   Calling other people's licenses "restrictive" and your own
> "absolutely free", is not, IMHO, a fair statement of the case.

I understand how those who have chosen a proprietary route
would like to minimize the differences, since they have a
commercial interest in doing so, but in trying to make this
argument, you should not exaggerate. Only *you* have called
the GPL absolutely free. All I said was that one characteristic
of the GPL is that you are absolutely free to redistribute what
you received. By this I mean redistributing the entire product,
I am not talking about deriviative works.

It is definitely the case that the GPL is like the Microsoft
License in that, for example, both ACT and MS are selling
copies of copyright license with a license that places
restrictions on the user. For more details on this, see
the transcript of the GEAC vs GRACE trial in Newark Federal
District Court, where I testified in detail on this point
(the attorney for the other side being confused like Tom :-)
The difference is in the qualitative nature of the restrictions.
The reason that lawyers find the GPL odd as a license agreement
is that it reads as being more interested in ensuring what the
user CAN do than what they CANNOT do.

If you can't see that CLAW has a much more restrictive license
than GtkAda (I am choosing vaguely competing products here),
then you are deluding yourself, and you will not serve your
customers well by this delusion.

As I said earlier, companies choose licenses to best meet their
interests. No one ever said that the GPL is the right license
for everything. In fact as you should know well, Ada Core
Technologies considers the GPL to be unsuitable for a
considerable chunk of the technology (including the Ada runtime
and ASIS for example).

Customers choose products on many factors, including the
license. Some license provisions are important some are not.
Actually for most of our customers, the redistribution right
is not particularly relevant, big companies are not in the
business of doing extra work to distribute copies of software
to their potential competitors. For them, what is important
about the GPL is the accessibility of the sources, and the
open source nature of the technology in terms of development.

To repeat, I have absolutely no criticism of the decision to
use a proprietary license, by which I mean a license that is
considerably more restrictive than the GPL, and in fact more
akin to the Microsoft license, for CLAW. Many, in fact most
of our tool partners (including the CLAW folks) choose to
license their software in this manner.

There are definitely arguments that can be made that this
helps increase revenue to the company involved. Whether these
are correct arguments or not depends on the circumstances.
Certainly the automatic assumption that a proprietary license
increases revenues is suspect.

However, it is really not possible to construct an argument that
a more restrictive license is beneficial to your users (other
than perhaps arguing that it decreases costs). In terms of
user benefits, a more permissive license is a clear win in
all terms.

Robert Dewar
Ada Core Technologies


Sent via Deja.com http://www.deja.com/
Before you buy.