GNAT GPL 2005: Too clever by half?

GNAT GPL 2005: Too clever by half?

[Marc A. Criley]

Maybe AdaCore is smarter than we've realized, intentionally or perhaps 
accidentally?

Consider this scenario:

Consider an Ada enthusiast or one-man-shop (OMS) developing a particular 
kind of product in Ada.  The "particular kind", or nature, of this 
product is such that Ada's role is only as its implementation language. 
  To a user it's just a black-box piece of software.  Unless it actually 
says "Ada" or they see the source code, they would never know Ada was 
involved.

Products such as container libraries, math libraries, etc., are not what 
I'm talking about, since you either have to interact with them as Ada, 
or be linking in the Ada runtime (and have to conform to any of its 
licensing constraints).

So, accounting software, CAD programs, word processors, that sort of 
stand-alone black-box application is what I'm going to address.  I'll 
use DTraq as an example, since it's the one I'm most familiar with :-)

DTraq is a data logging, monitoring, and playback application usable for 
development and test.  The current version (0.986a) does Ada code 
generation and requires compiling and linking that code, but the next 
release (scheduled for the end of this year) eliminates all that and 
will appear as a black box application to the user.  DTraq only works 
with Ada applications right now, but will be enhanced to support C++ and 
Java in the coming year.

DTraq has been licensed under the GMGPL, as has most other GNAT aided 
software developed by Ada enthusiasts.  If I were to use the new GNAT 
GPL 2005, I would pretty much be driven to changing that to the full GPL.

And then it struck me.

So what?

If I were a greedy capitalist, that might even be a good thing, in a 
bizarre, twisted way.

If I sell DTraq to some customer, I'll have to provide them the source 
code.  If they're not an Ada shop, if no one there knows Ada, if they 
couldn't care less about Ada, then I don't care that they have the 
source.  If they want support and maintenance, they're going to have to 
have a support contract with McKae--the fact that they have the entire 
source code base for the product is irrelevant to our business 
relationship, since they're not going to do anything with it.

If the customer does themselves develop non-free Ada or even GMGPL 
software, they can't reuse any of my software because it's under the GPL 
(unless they want to starting GPLing their own stuff).  If they do, then 
it's the usual GPL license violation, and you call down upon them the 
wrath of Stallman.

So here's the net result:

If you're hard over to free software, GPL GNAT 2005 is just fine and 
dandy, since it'll force people to make their software free (libre), 
which is the goal of the FSF.

On the other hand, if you're a greedy proprietary capitalist developing 
a product whose external functionality and use is language independent, 
GPL GNAT 2005 is also just fine, since by the requisite GPLing of your 
application, it will keep the vast majority of your customers from 
taking your source code and doing anything other than reading it.  All 
they'll be able to do with your product then is _use_ it (and perhaps 
modify it on their own, _if_ they have the expertise to do so).

I'm mulling this over...

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

"Marc A. Criley" <mcNOSPAM@mckae.com> writes:
> If you're hard over to free software, GPL GNAT 2005 is just fine and 
> dandy, since it'll force people to make their software free (libre), 
> which is the goal of the FSF.
>
> On the other hand, if you're a greedy proprietary capitalist developing 
> a product whose external functionality and use is language independent, 
> GPL GNAT 2005 is also just fine, since by the requisite GPLing of your 
> application, it will keep the vast majority of your customers from 
> taking your source code and doing anything other than reading it.  All 
> they'll be able to do with your product then is _use_ it (and perhaps 
> modify it on their own, _if_ they have the expertise to do so).
>
> I'm mulling this over...

Ah, the beauty of free software will never cease to amaze me.

But I can think of two flaws in your reasoning.  One, your customer
could hire anybody to do the maintenance; they don't have to hire you.
AdaCore solves this by hiring all the experts who are good enough to
tamper with the compiler, but depending on your product this may or
may not be possible.  The second flaw is if a greedy capitalist wants
to sell Ada development tools - e.g. code coverage, static analysis or
whatnot.  Then the customer can be expected to be quite literate with
Ada.

Me personally, I refuse to purchase or recommend any software if I
can't see and compile its sources for myself :) Call me a control
freak.

I'll repeat once more: I personally have no objection to libgnat being
GPLed.  And now today's idea: if AdaCore came out of the closet and
said "from now on we do dual licensing like TrollTech, MySQL or
Sleepycat", I'd have no problems with that, and I'd find it difficult
to sympathise with "greedy proprietary capitalists".

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Samuel Tardieu]

>>>>> "Marc" == Marc A Criley <mcNOSPAM@mckae.com> writes:

Marc> DTraq has been licensed under the GMGPL, as has most other GNAT
Marc> aided software developed by Ada enthusiasts.  If I were to use
Marc> the new GNAT GPL 2005, I would pretty much be driven to changing
Marc> that to the full GPL.

Marc> And then it struck me.

Marc> So what?

So I hope DTraq doesn't use OpenSSL for example. Linking and
distributing GMGPL code with OpenSSL is fine, linking GPL code with
OpenSSL is not. You would not be able to provide your customers with a
binary, they would have to do the link on site.

Sure, you could add an exception in DTraq code so that linking with
OpenSSL is excluded from the GPL requirements, as has been done with
OpenVPN (http://tinyurl.com/97xcq). However, you cannot add this
exception to the GNAT runtime (you're not the copyright owner and the
GPL doesn't allow you to do so). So you're still stuck if you use the
GNAT GPL 2005 compiler, while you're ok if you use a GMGPL runtime.

And this is only one example...

So what?

  Sam
-- 
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/sam

Re: GNAT GPL 2005: Too clever by half?

[Marc A. Criley]

Ludovic Brenta wrote:
> 
> Ah, the beauty of free software will never cease to amaze me.

Which is probably why it took a week before this aspect hit me :-)

> But I can think of two flaws in your reasoning.  One, your customer
> could hire anybody to do the maintenance; they don't have to hire you.
> AdaCore solves this by hiring all the experts who are good enough to
> tamper with the compiler, but depending on your product this may or
> may not be possible.  The second flaw is if a greedy capitalist wants
> to sell Ada development tools - e.g. code coverage, static analysis or
> whatnot.  Then the customer can be expected to be quite literate with
> Ada.

I wouldn't call them flaws.

While yes, anybody could be hired, the application's developer is most 
likely going to be high on the list of prospective maintainers, since 
it's unlikely anybody else knows the product as well as they do, and can 
turn around an improvement or fix as fast.

And second, the industry isn't exactly bustin' at the seams with 
individuals that are "quite literate with Ada".  Nor is the market for 
Ada development tools.

So they're not so much "flaws" as "nits" :-)

If Ada is irrelevant to a customer, and irrelevant to an application's 
functionality, then in the end the licensing of GNAT GPL 2005 is mostly 
irrelevant, and by and large the product can be treated *just like any 
other commercial* product.

Marc

Re: GNAT GPL 2005: Too clever by half?

[Keith Thompson]

"Marc A. Criley" <mcNOSPAM@mckae.com> writes:
> Maybe AdaCore is smarter than we've realized, intentionally or perhaps
> accidentally?
>
> Consider this scenario:
[...]
> DTraq has been licensed under the GMGPL, as has most other GNAT aided
> software developed by Ada enthusiasts.  If I were to use the new GNAT
> GPL 2005, I would pretty much be driven to changing that to the full
> GPL.
>
> And then it struck me.
>
> So what?
[...]
> So here's the net result:
[...]
> On the other hand, if you're a greedy proprietary capitalist
> developing a product whose external functionality and use is language
> independent, GPL GNAT 2005 is also just fine, since by the requisite
> GPLing of your application, it will keep the vast majority of your
> customers from taking your source code and doing anything other than
> reading it.  All they'll be able to do with your product then is _use_
> it (and perhaps modify it on their own, _if_ they have the expertise
> to do so).
>
> I'm mulling this over...

Couldn't your customer also redistribute your sources to someone else,
who might otherwise have paid you for a copy of your product?

-- 
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center             <*>  <http://users.sdsc.edu/~kst>
We must do something.  This is something.  Therefore, we must do this.

Re: GNAT GPL 2005: Too clever by half?

[Brian May]

>>>>> "Marc" == Marc A Criley <mcNOSPAM@mckae.com> writes:

    Marc> If you're hard over to free software, GPL GNAT 2005 is just
    Marc> fine and dandy, since it'll force people to make their
    Marc> software free (libre), which is the goal of the FSF.

The problem, as Samuel Tardieu has already mentioned, is that not all
"free software" uses the GPL.

There are heaps of other licenses out there, from Apache license,
openssl license, BSD license, libstdc++ license, LGPL, just to mention
the first ones that come to the top of my head.

It is simply not possible to convert all of these projects to use the
GPL (or GPL compatible[1]) licenses, no matter how much Richard
Stallman would like this.

Note:

[1] For example, the python license is GPL compatible; Does this mean
a GPL application compiled with GNAT GPL 2005 could also depend on
python?

Or another (dodgy) example: suppose I wanted to enhance python by
adding new Ada code to it. Would I be allowed to compile the Ada parts
with GNAT GPL 2005?

Trying to think these examples through is making me dizzy...
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

Brian May a écrit :
> [1] For example, the python license is GPL compatible; Does this mean
> a GPL application compiled with GNAT GPL 2005 could also depend on
> python?

Interesting; the GNAT Programming Studio (gnat-gps in Debian) links
with Python and uses it as a scripting language.  I'd better work out
the legal consequences when I do GPS 3.0.0.

> Or another (dodgy) example: suppose I wanted to enhance python by
> adding new Ada code to it. Would I be allowed to compile the Ada parts
> with GNAT GPL 2005?
>
> Trying to think these examples through is making me dizzy...

Nothing a good hard look to the licenses involved wouldn't solve.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Maybe I missed something? I thought ACT had licensed Gnat under terms 
that allowed you to develop "proprietary" software with it? Did that 
somehow change? Is it now such that anything compiled by Gnat has to 
fall under their licensing terms?

That would be Double-Plus-Ungood.

MDC

Marc A. Criley wrote:
> 
> DTraq has been licensed under the GMGPL, as has most other GNAT aided 
> software developed by Ada enthusiasts.  If I were to use the new GNAT 
> GPL 2005, I would pretty much be driven to changing that to the full GPL.


-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Preben Randhol]

In article Marin David Condic wrote:
>Maybe I missed something? I thought ACT had licensed Gnat under terms 
>that allowed you to develop "proprietary" software with it? Did that 
>somehow change? Is it now such that anything compiled by Gnat has to 
>fall under their licensing terms?

Yes. Unless you buy support. The GNAT GPL 2005 is pure GPL.

I mean you can of course put whatever license on you uncompiled code,
but if you want to compile it with GNAT GPL 2005 you must change to GPL
as far as I understand.

Preben

Re: GNAT GPL 2005: Too clever by half?

[Marc A. Criley]

Brian May wrote:
>>>>>>"Marc" == Marc A Criley <mcNOSPAM@mckae.com> writes:
> 
> 
>     Marc> If you're hard over to free software, GPL GNAT 2005 is just
>     Marc> fine and dandy, since it'll force people to make their
>     Marc> software free (libre), which is the goal of the FSF.
> 
> The problem, as Samuel Tardieu has already mentioned, is that not all
> "free software" uses the GPL.
> 
> There are heaps of other licenses out there, from Apache license,
> openssl license, BSD license, libstdc++ license, LGPL, just to mention
> the first ones that come to the top of my head.
> 
> It is simply not possible to convert all of these projects to use the
> GPL (or GPL compatible[1]) licenses, no matter how much Richard
> Stallman would like this.

But this isn't a problem for a developer that is an adherent of the free 
(libre) software development philosophy of Stallman, et.al.  If you want 
to use that developer's GPLed software, you have to have a compatible 
license.  If you can't or won't employ such a license, then it's your 
loss, you can't use the software.

Conversely, if the developer would like to freely exploit GPL software, 
but doesn't buy into FOSS theology, by actually making their code GPL 
they seriously limit what developers using non-GPL compatible licenses 
can do with it.  Almost like proprietary software.

I'm not passing judgement on these approaches to the use or misuse of 
the GPL, I'm just finding that there may be different utilizations of it 
that arise from a developer's motivation.

-- Marc A. Criley
-- McKae Technologies
-- www.mckae.com
-- DTraq - XPath In Ada - XML EZ Out

Re: GNAT GPL 2005: Too clever by half?

[Marc A. Criley]

Marin David Condic wrote:
> Maybe I missed something? I thought ACT had licensed Gnat under terms 
> that allowed you to develop "proprietary" software with it? Did that 
> somehow change? Is it now such that anything compiled by Gnat has to 
> fall under their licensing terms?
> 
> That would be Double-Plus-Ungood.

Yes, you must've been hiding under a rock for the last couple weeks :-) 
  The licensing change, though, only applies to the just released GNAT 
GPL 2005 public release.

If you got half a day or so you can catch up by reading the "GNAT GPL 
2005 Edition is now available" thread.

Much consternation and gnashing of teeth.

Marc

Re: GNAT GPL 2005: Too clever by half?

[Marc A. Criley]

Preben Randhol wrote:
> I mean you can of course put whatever license on you uncompiled code,
> but if you want to compile it with GNAT GPL 2005 you must change to GPL
> as far as I understand.

Just a minor point:  It's not the compiling, it's linking with the 
runtime library, or using any of the vendor (g-*) packages; and the 
instantiation exception to the GPL (such as for Integer_IO) has been 
removed.

Marc

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Without trying to start the great Free/Proprietary debate, I'll say 
this: I don't mind whatever ACT decides they want to do with their 
compiler. If they've determined that it is in their own best interest to 
drop the GMGPL in favor of the GPL, that is 100% within their rights. 
However, for a publicly released compiler, I think it would only be fair 
that they have it put up big, visible announcements whenever it is 
invoked to the effect that "The Software You Are About To Compile MUST 
Be Distributed Under Our Terms And Conditions - Or None At All!!!!" The 
reason being that the unsuspecting user might think it is perfectly OK 
and within their rights to distribute THEIR OWN software under any T's & 
C's they like - and not realize they have been infected by the Stallman 
Virus.

MDC


Preben Randhol wrote:
> 
> Yes. Unless you buy support. The GNAT GPL 2005 is pure GPL.
> 
> I mean you can of course put whatever license on you uncompiled code,
> but if you want to compile it with GNAT GPL 2005 you must change to GPL
> as far as I understand.
> 


-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Its a point that makes little difference. Its pretty hard to write 
anything that is non-trivial or at all useful without linking to the 
runtime library or instantiating a compiler-supplied generic.

I don't mind them doing it. I mind them doing it without a real clear 
notice that their compiler is now unacceptable for a large range of 
development. Someone could easily be in violation of their T's and C's 
and not even know it because it isn't intuitively obvious to the casual 
observer that when you write your own "Hello World" app and give it to 
someone else without source and a GPL license that you just opened 
yourself up to a lawsuit.

Maybe ACT is planning an "Apple Computer Strategy"? Make money by sueing 
everyone who violates their intellectual property rights?

MDC


Marc A. Criley wrote:

> Just a minor point:  It's not the compiling, it's linking with the 
> runtime library, or using any of the vendor (g-*) packages; and the 
> instantiation exception to the GPL (such as for Integer_IO) has been 
> removed.
> 
> Marc

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

I suppose it is an issue of me hiding under a rock. With a relatively 
new son crawling on the floor and some increased responsibilities at 
work, I don't get to read this newsgroup as much anymore. Dropping by 
occasionally and trying to catch up is a little tough. (Like a child 
wandering in in the middle of a movie...) ;-)

MDC


Marc A. Criley wrote:

> 
> Yes, you must've been hiding under a rock for the last couple weeks :-) 
>  The licensing change, though, only applies to the just released GNAT 
> GPL 2005 public release.

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Marin David Condic wrote:

> the unsuspecting user might think it is perfectly OK 
> and within their rights to distribute THEIR OWN software under any T's & 
> C's they like - and not realize they have been infected by the Stallman 
> Virus.

What is the reason that people think they can distribute
software under their own terms and conditions when (a) the
software is linked with software they don't own, and (b)
the terms and conditions of that software talk about
redistribution permissions?

No formal difference between GPL, EULA, Apple, or whatever
license when it states something about redistribution permissions.

Re: GNAT GPL 2005: Too clever by half?

[Dr. Adrian Wrigley]

On Fri, 23 Sep 2005 18:29:12 +0200, Georg Bauhaus wrote:

> What is the reason that people think they can distribute
> software under their own terms and conditions
...

because that is the normal thing for code in C, C++ and
many other languages compiled with GPL compilers and
only using the built-in capabilities of the language
and standard libraries.

As others have said, it'd be an easy mistake to make.
Are there any other GPL compilers which generate
encumbered binaries from unencumbered sources and not
linked against any special-purpose libraries?
-- 
Adrian

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

<quote>
 I'd find it difficult to sympathise with "greedy proprietary
capitalists
<\quote>

What's the difference between greedy proprietary capitalists and greedy
ideologists?

I've watched Stallman get angry with a crowd of two hundred academics
when someone ask him about the viral nature of the GPL.  He said that
calling  the GPL "viral" is flat out wrong.   He demonstrated that he
wants us to reform our minds around his so that the GPL is immune to
criticism.

The greedy proprietary capitalist wants our money.  The greedy
ideologist wants our soul.

I've watched Stallman speak.  He is either a liar or delusional about
the nature of the GPL.  I use to be a strong supporter of the GPL, but
when I heard Stallman speak, I thought, "This man is nuts!".  He wants
us to give up the right to decide how we are to be compensated for our
labor with respect to software.

The GMGPL isn't about being greedy.  It is about having the freedom to
choose your method of compensation as oppose to having Stallman dictate
your rights to you.

-CRM

P.S.  In general, I support free software; however, I disagree with
trying to force people to release free software.  This in my opinion is
not "free software".  If we the users are not allow the copyright
rights over works that we created by using the so called "free
software" to write and compile, the users aren't really free are they?

Re: GNAT GPL 2005: Too clever by half?

[Hyman Rosen]

Chad  R. Meiners wrote:
> He wants us to give up the right to decide how we are to be compensated
> for our labor with respect to software.

Not unlike the way Andrew Wiles labored for over seven years,
and yet mathematicians everywhere will greedily use the
Taniyama-Shimura theorem in their own proofs without paying
him one cent (or Taniyama and Shimura, for that matter).

And you're wrong, by the way. You can arrange to be compensated
for your labor in any way you like. Stallman would simply deny
you the ability to prevent others from using the product of your
labor.

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Dr. Adrian Wrigley wrote:
> On Fri, 23 Sep 2005 18:29:12 +0200, Georg Bauhaus wrote:
> 
> 
>>What is the reason that people think they can distribute
>>software under their own terms and conditions
> 
 [>> ...(b)
  >>the terms and conditions of that software talk about
  >>redistribution permissions? ]

> because that is the normal thing for code in C, C++ and
> many other languages compiled with GPL compilers and
> only using the built-in capabilities of the language
> and standard libraries.

 It is still the normal thing for code in C, C++
and many other languages, including Ada, compiled with GPL
compilers and only using the built-in capabilities of
the language and standard libraries, to distribute your
software under your own terms and conditions.
You just can't use the aptly named GNAT GPL Edition.
So what? Use the FSF Edition, or get the Pro Edition,
you have the freedom to do so.

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Chad R. Meiners wrote:

> I've watched Stallman speak.  He is either a liar or delusional about
> the nature of the GPL.  I use to be a strong supporter of the GPL, but
> when I heard Stallman speak, I thought, "This man is nuts!".  He wants
> us to give up the right to decide how we are to be compensated for our
> labor with respect to software.

Did he say so?

> The GMGPL isn't about being greedy.  It is about having the freedom to
> choose your method of compensation as oppose to having Stallman dictate
> your rights to you.


   "This General Public License does not permit incorporating your
program into proprietary programs.  If your program is a subroutine
library, you may consider it more useful to permit linking proprietary
applications with the library.  If this is what you want to do, use the
GNU Library General Public License instead of this License."

The viral effect is not created by the GPL. As with every licence,
the viral effect is created by those who place a license on their
software, provided the software is used.
Whether the virus is good or bad is a different question.
In the case of the GPL, those who use it state the rules of giving
and taking. If you don't want to give, then you can't take.


> P.S.  In general, I support free software; however, I disagree with
> trying to force people to release free software.  This in my opinion is
> not "free software".  If we the users are not allow the copyright
> rights over works that we created by using the so called "free
> software" to write and compile, the users aren't really free are they?

What, in your view, is a derived work? And what is it worth?

Re: GNAT GPL 2005: Too clever by half?

[Michael Bode]

Ludovic Brenta <ludovic@ludovic-brenta.org> writes:

> I'll repeat once more: I personally have no objection to libgnat being
> GPLed.  And now today's idea: if AdaCore came out of the closet and
> said "from now on we do dual licensing like TrollTech, MySQL or
> Sleepycat", I'd have no problems with that, and I'd find it difficult
> to sympathise with "greedy proprietary capitalists".

But that's just what happend now. The only problem is that AdaCore
seems to say "Gnat is for 2 types of developers: GPL developers and
the ones like Airbus and Boeing. If you are a One-Man-Show(*)
closed-source developer, go away, use ObjectAda or forget about Ada
altogether." Which is still fine, but could someone please explain to
me what's their profit from this strategy? 

They already have the contracts from Airbus and Boeing and they sure
won't get much contracts from the One-Man-Shows at 14000$/year when
one can get Visual C++/C# between 119ￜ and 899ￜ.

(*) this also includes larger businesses where software is only a
small part of the product.

What should a "greedy proprietary capitalist" do if he simply happens
to need a third party closed-source library in his Ada app? Something
like a DLL to access an device like a spectrometer or a PLC. I can't
make an app GPL if one part of it is a CSS DLL which has to pe payed
per copy and for which I don't have the source. Am I evil?

My consequence is to look for alternatives to Gnat be it other Ada
compilers in an affordable price range (0 .. 2000ￜ) or other
languages.

BTW: does GtkAda (as long as it is GMGPL) work with ObjectAda?

-- 
Michael Bode

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Michael Bode wrote:

> If you are a One-Man-Show(*)
> closed-source developer, go away, use ObjectAda or forget about Ada
> altogether."

Or use FSF's GNAT.

Re: GNAT GPL 2005: Too clever by half?

[Michael Bode]

Georg Bauhaus <bauhaus@futureapps.de> writes:

>> If you are a One-Man-Show(*)
>> closed-source developer, go away, use ObjectAda or forget about Ada
>> altogether."
>
> Or use FSF's GNAT.

I've just browsed through
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/. Is there a file
stating that libgnat is GMGPL or does one have to check every source
file if it is part of libgnat and has the exception to the GPL? Is
that what AdaCore means by:

  There are other places where you can obtain partial versions of
  GNAT, typically consisting of the compiler. For instance the GNAT
  compiler can be obtained in source form from the Free Software
  Foundation (FSF). Other versions of the GNAT compiler exist (e.g. in
  GNU Linux distributions). AdaCore has no control over these and
  cannot guarantee their quality or suitability for a particular
  purpose. Most important you should ascertain the license and IPR
  (Intellectual Property Rights) guarantees from its provider.

I mean, I want to develop applications, not analyse compilers.

-- 
Michael Bode

Re: GNAT GPL 2005: Too clever by half?

[Brian May]

>>>>> "Chad" == Chad R Meiners <chad.rmeiners@gmail.com> writes:

    Chad> I've watched Stallman speak.  He is either a liar or
    Chad> delusional about the nature of the GPL.  I use to be a
    Chad> strong supporter of the GPL, but when I heard Stallman
    Chad> speak, I thought, "This man is nuts!".  He wants us to give
    Chad> up the right to decide how we are to be compensated for our
    Chad> labor with respect to software.

Just because you dislike Stallman isn't really a good reason to
dislike the GPL IMHO.

I believe as the copyright owner you have the right to decide how your
software will be used. Hence I don't think the GPL is any more "viral"
then other licenses that exist, e.g. "not available for commercial
use". In fact, the GPL is considerably more flexible then such a
license.

    Chad> P.S.  In general, I support free software; however, I
    Chad> disagree with trying to force people to release free
    Chad> software.  This in my opinion is not "free software".  If we
    Chad> the users are not allow the copyright rights over works that
    Chad> we created by using the so called "free software" to write
    Chad> and compile, the users aren't really free are they?

I think it is inappropriate having a run-time library required by a
compiler impose license restrictions on the code you compile.

The code I write is *my* code, not Adacore's code, and I think the
decision on how I should license should be up to me. It should also be
entirely up to me what other libraries I link with - using openssl
should be acceptable.

Having said that, I would be considerably happier if the restriction
was "software must adhere with the DFSG" as opposed to "software must
be GPL or GPL compatible".
-- 
Brian May <bam@snoopy.apana.org.au>

Re: GNAT GPL 2005: Too clever by half?

[Frank]

Hi

What oportunities does an open licensing give to the big customers of 
AdaCore?
What if a big company choses to keep all their development on the 3.15p 
version --- because they have 25 developers
this becomes cheap. However they need the quality from the AdaCore service- 
so they buy one licence of "5-seats"-
on which they compile all other code regulary (especially before testing) 
and from which they base their support requests. Does the licensing prevent 
this? If this is allowed it must surely be bad for AdaCore business.

Frank

Re: GNAT GPL 2005: Too clever by half?

[Larry Kilgallen]

In article <43350a13$1@news.broadpark.no>, "Frank" <myworstenemy@somewhere.com> writes:

> What if a big company choses to keep all their development on the 3.15p 
> version --- because they have 25 developers
> this becomes cheap. However they need the quality from the AdaCore service- 
> so they buy one licence of "5-seats"-
> on which they compile all other code regulary (especially before testing) 
> and from which they base their support requests. Does the licensing prevent 
> this?

Certainly AdaCore would be dumb if their license terms did not consider this.
Too dumb to maintain an Ada compiler.

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

Georg Bauhaus <bauhaus@futureapps.de> writes:

> Dr. Adrian Wrigley wrote:
>> On Fri, 23 Sep 2005 18:29:12 +0200, Georg Bauhaus wrote:
>>
>>>What is the reason that people think they can distribute
>>>software under their own terms and conditions
>>
> [>> ...(b)
>  >>the terms and conditions of that software talk about
>  >>redistribution permissions? ]
>
>> because that is the normal thing for code in C, C++ and many other
>> languages compiled with GPL compilers and only using the built-in
>> capabilities of the language and standard libraries.
>
> It is still the normal thing for code in C, C++ and many other
> languages, including Ada, compiled with GPL compilers and only using
> the built-in capabilities of the language and standard libraries, to
> distribute your software under your own terms and conditions.  You
> just can't use the aptly named GNAT GPL Edition.  So what? Use the
> FSF Edition, or get the Pro Edition, you have the freedom to do so.

I'd like to add that the FSF (gcc.gnu.org) does *not* distribute
binaries of GCC.  See http://gcc.gnu.org/install/binaries.html

If you want to use GCC, you must either compile it yourself or obtain
if from someone else who compiled it for you.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

Michael Bode <m.g.bode@web.de> writes:
> Georg Bauhaus <bauhaus@futureapps.de> writes:
>
>>> If you are a One-Man-Show(*)
>>> closed-source developer, go away, use ObjectAda or forget about Ada
>>> altogether."
>>
>> Or use FSF's GNAT.
>
> I've just browsed through
> http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/. Is there a file
> stating that libgnat is GMGPL or does one have to check every source
> file if it is part of libgnat and has the exception to the GPL? Is
> that what AdaCore means by:
>
>   There are other places where you can obtain partial versions of
>   GNAT, typically consisting of the compiler. For instance the GNAT
>   compiler can be obtained in source form from the Free Software
>   Foundation (FSF). Other versions of the GNAT compiler exist (e.g. in
>   GNU Linux distributions). AdaCore has no control over these and
>   cannot guarantee their quality or suitability for a particular
>   purpose. Most important you should ascertain the license and IPR
>   (Intellectual Property Rights) guarantees from its provider.
>
> I mean, I want to develop applications, not analyse compilers.

You have to look at every source file, because every source file is
licensed individually.  You can start with

$ grep -rL "As a special exception" gcc/gcc/ada

but this will list the many files that make up the compiler itself.
There is a plan to move the files that make up libgnat to gcc/libada,
but this plan has not been executed yet.  When this is done, things
will be much more simple.

Like most people, you probably won't have to analyse the compiler for
yourself though; you can look to distribution maintainers to do that
for you.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

"Frank" <myworstenemy@somewhere.com> writes:
> Hi
>
> What oportunities does an open licensing give to the big customers of 
> AdaCore?
> What if a big company choses to keep all their development on the 3.15p 
> version --- because they have 25 developers
> this becomes cheap. However they need the quality from the AdaCore service- 
> so they buy one licence of "5-seats"-
> on which they compile all other code regulary (especially before testing) 
> and from which they base their support requests. Does the licensing prevent 
> this? If this is allowed it must surely be bad for AdaCore business.

A quote from a recent paper by Jamie Ayre and Franco Gasperoni [1] says:

"How do we check that users do not exceed the subscription level they
purchased?  We don't, we trust our customers.  Trust builds trust."

And later in the same paper:

"AdaCore has seen steady growth over the 10 years since launch and
received several awards for strong financial performance.  From the
core staff of 3 at the time the company was created, AdaCore now has
over 40 permanent staff as well as several consultants and interns."

[1] http://oss2005.case.unibz.it/Papers/10.pdf

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Brian May wrote:
> 
> The code I write is *my* code, not Adacore's code, and I think the
> decision on how I should license should be up to me. It should also be
> entirely up to me what other libraries I link with - using openssl
> should be acceptable.
> 
That is the real issue - and it goes more towards the relative value of 
the compiler to someone rather than an issue of "rights". An average 
developer might not give it two seconds of thought and just assume that 
the compiler used to generate machine code from his code doesn't impose 
some sort of ownership rights or restrictions on his code. ACT can 
impose any restrictions they like, but I think that makes their compiler 
significantly less useful - and they should make it CLEAR to everyone 
running the compiler that by using it, they just got infected.

In my job, I have been able to get a handful of companies using Ada for 
various purposes by saying "Here's the Gnat compiler and its free and 
you can use it for anything you want without restriction." They've gone 
on to use it for a variety of purposes - usually some kind of internal 
development. Some of them have subsequently coughed up some money to ACT 
to become supported users. My own company may end up doing this at my 
instigation. But now the scenario changes.

Since the free Gnat compiler will contain these GPL restrictions on the 
generated code, I can't tell my customers they can use it for anything 
they want. They will almost certainly REFUSE to accept any limitations 
on their intellectual property. In their position, I would and will and 
do. Whatever I pay to develop, I want 100% of the rights to or (and this 
is the critical point:) I WILL FIND ANOTHER WAY.

My customers (and my company) will be looking at that situation and here 
are their alternatives:

1) Go buy a support agreement from ACT - something that is significantly 
more expensive than, say, buying a copy of Microsoft Visual C++.

2) Go buy a licensed copy of someone else's Ada compiler - perhaps 
paying about as much (or maybe more) than one would for MS-VC++.

3) Go use whatever native compilers they already have for other 
languages - such as the C compiler native on their Sun workstations. And 
after all, C is such a ubiquitous language, with just about every 
developer already knowing it and thousands of tools available to support 
it and huge libraries of stuff already available to link to and that's 
what the OS is written in so its just plain easier to get to the OS 
services, and all their tools like Matlab, Simulink, etc. all generate C 
code anyway and etc. etc. etc.

Which way do you guess they'll go? Hint: It was an uphill battle to get 
them to use Ada at all - even with a free compiler with no restrictions 
on their code.

Unless my customers are doing *embedded* development, they really don't 
need much in the way of support from a compiler vendor. What they need 
is s shrink-wrap copy of a compiler with an unrestrictive license on 
their end-product-code and some level of acceptable reliability and 
documentation. They probably won't go to ACT to get that since the bill 
would be excessive for the minimal amount of support they'd need. Since 
Microsoft is willing to sell them that for something less than $1000, 
guess what way they'll go now?

Now if ACT wanted to package up a shrink-wrap developer's kit and sell 
it for a few hundred bucks, (imposing no restrictions on the developer's 
code) then I'd bet a few of the people I've got using Gnat might 
actually spring for it.

Perhaps there's a good capitalist business opportunity here?

MDC

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Dr. Adrian Wrigley wrote:
> 
> because that is the normal thing for code in C, C++ and
> many other languages compiled with GPL compilers and
> only using the built-in capabilities of the language
> and standard libraries.

Dittos. And note that the moment it gets noticed that with *Ada* there 
is a licensing issue but with *C* or *C++* or some other language there 
is NO licensing issue asnd what do you imagine the pressure will become? 
"Why do you insist on using that antiquated, has-been, unimportant, 
niche language with LICENSING RESTRICTIONS when you could just use what 
everyone else uses and get all those compatibility advantages with other 
tools and libraries and NO LICENSING RESTRICTIONS..."

How do I make an argument for Ada if business managers see nothing but 
negatives to using it?

MDC

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Ludovic Brenta wrote:

> 
> 
> I'd like to add that the FSF (gcc.gnu.org) does *not* distribute
> binaries of GCC.  See http://gcc.gnu.org/install/binaries.html
> 
> If you want to use GCC, you must either compile it yourself or obtain
> if from someone else who compiled it for you.
> 
The harder it is to use Ada, the less it will get used. ACT, et alia, 
don't have to like that fact - they just have to live with that fact. I 
won't say they have no right to do whatever they want with the licensing 
of their IP - its their right. I just think that if it is at all 
difficult to get a working Ada05 compiler for most of the "casual" users 
(those who don't need some kind of industrial strength support) that Ada 
  (at least Ada05) starts becoming a less attractive alternative - and 
do be aware that there are very few people who HAVE TO use Ada.

There are enough pressures today against Ada. Lack of an easily 
accessible, freely available, 
unencumbered-with-restrictive-licenses-on-YOUR-code compiler is just one 
more pressure to switch to something else.

MDC

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Ludovic Brenta]

Marin David Condic <nobody@noplace.com> writes:
> Brian May wrote:
> In my job, I have been able to get a handful of companies using Ada for 
> various purposes by saying "Here's the Gnat compiler and its free and 
> you can use it for anything you want without restriction." They've gone 
> on to use it for a variety of purposes - usually some kind of internal 
> development. Some of them have subsequently coughed up some money to ACT 
> to become supported users. My own company may end up doing this at my 
> instigation. But now the scenario changes.

Not for internal development; the GPL only kicks in if you want to
distribute your program.  For internal development, GNAT GPL 2005
Edition is just fine.

But this is all the "marketing argument".

[...]
> My customers (and my company) will be looking at that situation and here 
> are their alternatives:
>
> 1) Go buy a support agreement from ACT - something that is significantly 
> more expensive than, say, buying a copy of Microsoft Visual C++.
>
> 2) Go buy a licensed copy of someone else's Ada compiler - perhaps 
> paying about as much (or maybe more) than one would for MS-VC++.
>
> 3) Go use whatever native compilers they already have for other 
> languages - such as the C compiler native on their Sun workstations. And 
> after all, C is such a ubiquitous language, with just about every 
> developer already knowing it and thousands of tools available to support 
> it and huge libraries of stuff already available to link to and that's 
> what the OS is written in so its just plain easier to get to the OS 
> services, and all their tools like Matlab, Simulink, etc. all generate C 
> code anyway and etc. etc. etc.

4) Use Debian, AIDE, Red Hat, MinGW, or any of the other binary
   distributions that ship GNAT or GCC with Ada support under GMGPL.

5) Compile GCC and all required libraries yourself from the sources in
   the CVS repositories.  These are under GMGPL.

6) Use GNAT GPL 2005 Edition unless and until they plan to distribute
   their program in binary-only form.

> Which way do you guess they'll go? Hint: It was an uphill battle to get 
> them to use Ada at all - even with a free compiler with no restrictions 
> on their code.
> 
> Unless my customers are doing *embedded* development, they really don't 
> need much in the way of support from a compiler vendor. What they need 
> is s shrink-wrap copy of a compiler with an unrestrictive license on 
> their end-product-code and some level of acceptable reliability and 
> documentation. They probably won't go to ACT to get that since the bill 
> would be excessive for the minimal amount of support they'd need. Since 
> Microsoft is willing to sell them that for something less than $1000, 
> guess what way they'll go now?
>
> Now if ACT wanted to package up a shrink-wrap developer's kit and sell 
> it for a few hundred bucks, (imposing no restrictions on the developer's 
> code) then I'd bet a few of the people I've got using Gnat might 
> actually spring for it.

Agreed.

> Perhaps there's a good capitalist business opportunity here?

Yes, probably; I've said that before in this thread.  There appears to
be a small market of people and companies willing to pay for minimal
support.

-- 
Ludovic Brenta.

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Ludovic Brenta wrote:
> 
> 
> Not for internal development; the GPL only kicks in if you want to
> distribute your program.  For internal development, GNAT GPL 2005
> Edition is just fine.
> 
Doesn't matter. Even if it is strictly internal development, a big 
corporation won't want to accept any restrictions on what they develop 
as their own IP. They don't know how they might want to use it tomorrow. 
If it has license restrictions on their IP, this might preclude them 
from selling it to someone on their own terms at a later date. Like I 
said, as a businessman, I would not accept that for anything I pay to 
develop.

Arguments that they could use another compiler (or buy support from ACT) 
won't really help. The path of least resistance will be to go with some 
alternative - and that alternative will likely not be Ada.


> But this is all the "marketing argument".
> 
Perhaps - but I think it is a losing marketing argument.


> 
> 
> 4) Use Debian, AIDE, Red Hat, MinGW, or any of the other binary
>    distributions that ship GNAT or GCC with Ada support under GMGPL.
> 
They tend to use Windows for most of this. They don't want to spend time 
trying to figure out how to get a compiler environment built for windows 
- they want to install it and get on with developing. Availability of a 
binary that is unencumbered with restrictions on their software may help 
- but for the tenuously held market, I'd really hate to see it be even a 
little bit harder to get something that would work.

Most companies don't want to spend money on hiring a bunch of geeks to 
get something to work. They want a disk full of software that just plugs 
in and they can go to town. They want it on their existing Windows 
platforms. They want it to install with the ease they are used to for 
other Windows software. If they can only get it with Linux, it becomes a 
dead issue. If they have to dork around with getting it to work on 
Windows, its a dead issue. They'll go somewhere else - there are 
alternatives.


> 5) Compile GCC and all required libraries yourself from the sources in
>    the CVS repositories.  These are under GMGPL.
> 
Unless they have some critical use, they won't spend time and energy 
building it for themselves. Nor would I.


> 6) Use GNAT GPL 2005 Edition unless and until they plan to distribute
>    their program in binary-only form.
> 
See the problem with a company's IP I identified above. Even if it is 
strictly internal, they won't want to be encumbered. They'll find 
alternatives.

> 
>>Perhaps there's a good capitalist business opportunity here?
> 
> 
> Yes, probably; I've said that before in this thread.  There appears to
> be a small market of people and companies willing to pay for minimal
> support.
> 
Most of the people I've got using Ada are using Gnat 3.15p (including 
myself) and they may not feel a need to switch (I don't). The issue may 
come up WRT Ada05. Unless there is some freely (or inexpensively) 
available Ada05 compiler that doesn't impose restrictions, my belief is 
that a large number of users won't look at Ada05 and will go down 
alternate routes. (Stick with what you've got and/or start migrating to 
C/C++ like "everyone else".)

Ada does not have some big, captive market and is thus not in a good 
position to start dictating T's and C's on people. The worst thing that 
can happen is that a perception may get started that "You can't use Ada 
for proprietary development like you can with other languages..."

Remember, the bulk of the users out there are not that sophisticated in 
distinguishing between a language and an implementation - and in a world 
with only a small number of implementations, there *isn't* much 
distinction. If they see that they can develop proprietary code with a 
gcc C compiler but can't develop proprietary code with a Gnat Ada 
compiler - you'll start hearing about how "You can develop proprietary 
code with C but you can't do it with Ada..."

MDC


-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Marin David Condic wrote:

> An average 
> developer might not give it two seconds of thought and just assume that 
> the compiler used to generate machine code from his code doesn't impose 
> some sort of ownership rights or restrictions on his code. ACT can 
> impose any restrictions they like, but I think that makes their compiler 
> significantly less useful - and they should make it CLEAR to everyone 
> running the compiler that by using it, they just got infected.

The most widespread free C++ compilers by Microsoft and Borland
do impose restrictions on what you are allowed to do with
the compiled code, and what you have to do.

In the case of GPL, and only GPL, you speak about licensing as
if it were a desease, when in fact every license just regulates.

For contrast, consider this extract from 
END-USER LICENSE AGREEMENT FOR MICROSOFT SOFTWARE
MICROSOFT VISUAL C++ TOOLKIT 2003

"...
 (ii) that the Redistributables only operate in conjunction with
Microsoft Windows platforms;"

Technically, I could say this is no less viral, in particular since
.NET assemblies are supposed to be runnable not just on MS Windows.


"1.1 General License Grant. Microsoft grants to you as an individual,
a personal, nonexclusive license to make and use copies of the Software
(i) for your internal use; (ii) for designing, developing,
testing and demonstrating your software product(s); and
(iii) for evaluation of the Software."

It is not clear to me how I can conclude form this that I am free
to use Microsofts free compilers in production use.
And who would expect that I am free to use Microsoft products for
free, and without restrictions?
 (After all, someone must have payed them some money, given their
wealth.)

If you can't use the GPL, as you mention, there are other supported
Ada compilers, and FSF GNAT.  For trying out compilers, what is
wrong with choosing FSF GNAT or taking an evaluation offer?


> Perhaps there's a good capitalist business opportunity here?

Always has been.

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Marin David Condic wrote:

> They want a disk full of software that just plugs 
> in and they can go to town.

In other words, they fall victim to the illusion of a DWIM
environment.

> They want it on their existing Windows 
> platforms. They want it to install with the ease they are used to for 
> other Windows software.

This is the issue. Impression management. I'm serious, and I think
the failure is not so much a licensing issue. All the more when
licenses are evaluated by comparing the impressions you get, either
from the wording, or from the people that some see associated with the
license. Chad Meiners has told us an example. He dislikes RMS,
and this reflects in part on the GPL.
(I don't know RMS, but I can read the GPL.)

Have you ever tried to do a non-trivial thing with Visual Studio?
If so then you know that Microsoft development is as complicated
as anything, if not more so, given the big complex, backwards
compatible collection of software they support. The learning curve
is not linear.

Yet MS marketing manages to make people believe that it is easy to
write Windows software. They let you make some easy steps in the
beginning. After that you are on your own, but you have already bought
the thing.

Re: GNAT GPL 2005: Too clever by half?

[Robert A. Matthews]

On Sat, 2005-09-24 at 13:40 +0200, Ludovic Brenta wrote:

> I'd like to add that the FSF (gcc.gnu.org) does *not* distribute
> binaries of GCC.  See http://gcc.gnu.org/install/binaries.html
> 
> If you want to use GCC, you must either compile it yourself or obtain
> if from someone else who compiled it for you.
> 

On the subject of FSF's GNAT, this seems to 
progress in a rather opaque manner. 
Is there a roadmap for this work?
An occasional status update would be appreciated.

Robert A. Matthews

(to reply replace "ignored" with "cla")

Re: GNAT GPL 2005: Too clever by half?

[Simon Wright]

Marin David Condic <nobody@noplace.com> writes:

> Doesn't matter. Even if it is strictly internal development, a big
> corporation won't want to accept any restrictions on what they
> develop as their own IP. They don't know how they might want to use
> it tomorrow. If it has license restrictions on their IP, this might
> preclude them from selling it to someone on their own terms at a
> later date. Like I said, as a businessman, I would not accept that
> for anything I pay to develop.

But the only restriction is that if you compile with the GPL compiler
and release *that* binary you are affected; if you recompile with any
other GNAT you are not.

Re: GNAT GPL 2005: Too clever by half?

[Simon Wright]

Georg Bauhaus <bauhaus@futureapps.de> writes:

> If you can't use the GPL, as you mention, there are other supported
> Ada compilers, and FSF GNAT.  For trying out compilers, what is
> wrong with choosing FSF GNAT or taking an evaluation offer?

It certainly seems to be true that, no matter how daft you or I might
think it, management would prefer to use Windows to develop on.

I wonder how difficult it is to build GNAT on Windows? I'm pretty sure
it has to be under MinGW. I did manage after many hours to get PolyORB
to build under MinGW, but it crashed dismally when used .. in contrast
to the Linux build, which worked out of the box. Perhaps it would be
worth a try .. on the other hand, I would like a little bit of a life!

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Simon Wright wrote:

> It certainly seems to be true that, no matter how daft you or I might
> think it, management would prefer to use Windows to develop on.
> 
> I wonder how difficult it is to build GNAT on Windows? I'm pretty sure
> it has to be under MinGW. I did manage after many hours to get PolyORB
> to build under MinGW, but it crashed dismally when used .. in contrast
> to the Linux build, which worked out of the box. Perhaps it would be
> worth a try .. on the other hand, I would like a little bit of a life!

Maybe "Solaris" is the magic word? The new workstations seem comparatively
affordable...

Re: GNAT GPL 2005: Too clever by half?

[Marin David Condic]

Simon Wright wrote:
> 
> 
> But the only restriction is that if you compile with the GPL compiler
> and release *that* binary you are affected; if you recompile with any
> other GNAT you are not.

Fair enough, but will there be a GMGPL version of Ada05? One that will 
install-and-go rather than one my customers will have to build from 
sources? I can keep using Gnat 3.15p forever (as can my customers) but 
will that line of compilers ever result in an Ada05?

I understand that the Ada code *could* be compiled with a different 
compiler but a) Ada is not as portable as it wishes it were and b) going 
into the project, I've got to presume I won't have another compiler. (If 
I had another one, I'd probably use that one if this is the ultimate 
destination.)

I'm not going to say "ACT - You owe me a no cost binary release of all 
your compilers..." I hope I'm being really clear on that. I'm merely 
pointing out two issues I see: 1) Making a public Ada compiler release 
that creates images that may not be released except under the GPL should 
be very visibly noted lest the uninformed start making assumptions about 
what they can legally do or not. (especially since you can do that with 
gcc and other non-Ada compilers rather readily) 2) Lack of a no-or-low 
cost Ada compiler for Ada05 (without restrictions) could hurt adoption 
of Ada.

ACT can do as they want with all my blessings behind them and I hope 
they make lots of money and stay in business a long time. I'd also thank 
them for being wonderful guys and making public releases of Gnat in the 
past with no restrictions on what I can do with my own code. I hope that 
with Ada05, we get something similar lest it discourage Ada usage.


MDC


-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: No.Mcondic.Spam@Del.Mindspring.Com
(Remove the "No.", ".Spam" and "Del." for the real address.)

     "The Christian ideal has not been tried and found wanting. It has
      been found difficult; and left untried."

         --  G. K. Chesterton
======================================================================

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

>From Brian May
<quote>
Just because you dislike Stallman isn't really a good reason to
dislike the GPL IMHO.
<\quote>

I never said that I dislike Stallman.  I watched and listened to him
during a talk, and I watched him get really irate towards anyone that
called the GPL viral.  He made it quite clear clear during his talk
that he wants us to

1.)  never call the GPL viral.

2.) never ever sell software.

He also thinks that a public trust fund is sufficient to support all
software development.  This in my view qualified him for the "nuts"
category; I don't dislike him; I think some of his ideas are wrong, and
since he refuses to discuss his views in a reasonable manner, he is
nuts.  The GPL has its uses.  I don't have an axe to grind with
Stallman or the GPL.

My comment was to the "greedy capitalists" statement.  "Greedy
capitalists" want money.  However, Stallman is an example of someone
that wants to set public policy to favor his own ideals as oppose to
the public good.  Forcing your ideals upon others is also greedy.
Stallman is just as greedy as those greedy capitalist.  That's my
point.

That being said the GPL can be used in non-greedy ways.  Just as a
capitalistic business can be run in a non-greedy manner.

-CRM

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

From: Hyman Rosen
<quote>
Not unlike the way Andrew Wiles labored for over seven years,
and yet mathematicians everywhere will greedily use the
Taniyama-Shimura theorem in their own proofs without paying
him one cent (or Taniyama and Shimura, for that matter).
<\quote>

When you invoke a property established by a theorem, are you making a
copy of that theorem.  No, you are not.  This is beside the point.

From: Hyman Rosen
<quote>
And you're wrong, by the way. You can arrange to be compensated
for your labor in any way you like. Stallman would simply deny
you the ability to prevent others from using the product of your
labor.
<\quote>

Your statement is self-contradictory.  Furthermore, the quote
"Stallman would simply deny
you the ability to prevent others from using the product of your
labor."

reads as "Stallman prevents you from protecting yourself from the theft
of your labor", which is a pretty serious violation of our rights.

-CRM

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

From: Georg Bauhaus

<quote>
Did he say so?
<\quote>

It was the implication of everything that he proposed during the talk.
He made it clear that he wants everyone to release under the GPL.  He
made it clear that he thought that people that "sold" software were
immoral.  He made it clear that he was very angry at those that sell
software.  In my opinion following what Stallman proposes would require
us to give up our rights.

With regards to the viral nature of the GPL, I don't care if it is
viral.  I do care if Stallman claims with a straight face that it isn't
viral, and I do care if Stallman gets angry with people that disagree
with him about the viral nature of the GPL.  In my opinion Stallman
isn't being honest.

<quote>
What, in your view, is a derived work? And what is it worth?
<\quote>

I am currently happy with the legal definition.  The worth is
multi-faceted. 

-CRM

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Chad R. Meiners wrote:
> 
> reads as "Stallman prevents you from protecting yourself from the theft
> of your labor", which is a pretty serious violation of our rights.

This statement would be incomplete because it fails to mention
that your labor is stolen from others'. ;-)

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

From: Georg Bauhaus

<quote>
Chad Meiners has told us an example. He dislikes RMS,
and this reflects in part on the GPL.
<\quote>

I don't dislike RMS!  I stated that RMS is greedy just like those
greedy capitalists.  I draw my opinion from personal experience with
the man, and from reading his writings.

-CRM

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

<quote>
This statement would be incomplete because it fails to mention
that your labor is stolen from others'. ;-)
<\quote>

My labor i.e. labor that I do myself.

How can the labor that I do myself be stolen from someone else?

I reply to words that people speak or write.  Not words that they
intended to write, or the words that I think that they meant to write.
If Hyman didn't mean what he wrote, he can always reply and correct his
statement.

-CRM

Re: GNAT GPL 2005: Too clever by half?

[Martin Krischik]

Frank wrote:

> Hi
> 
> What oportunities does an open licensing give to the big customers of
> AdaCore?
> What if a big company choses to keep all their development on the 3.15p
> version --- because they have 25 developers
> this becomes cheap. However they need the quality from the AdaCore
> service- so they buy one licence of "5-seats"-
> on which they compile all other code regulary (especially before testing)
> and from which they base their support requests. Does the licensing
> prevent this? If this is allowed it must surely be bad for AdaCore
> business.

Only programmers which have GNAT Tracker account may issue a support request
or download GNAT/Pro. Support Requests by Proxy are disallowed. How AdaCore
checks agaings proxy support requests I don't know.

But one thing you should know: It's "Support Requests" not "Bug Reports".
Programmers with GNAT Tracker account can also ask questions like:

I could implement A by using technique X or technique Y - which would be
better considing B - and why?

or

Why is "Hello World!" is soooo big? - ;-)

Not that I belive there are many such request.

Martin

-- 
mailto://krischik@users.sourceforge.net
Ada programming at: http://ada.krischik.com

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Chad R. Meiners wrote:
> <quote>
> This statement would be incomplete because it fails to mention
> that your labor is stolen from others'. ;-)
> <\quote>
> 
> My labor i.e. labor that I do myself.
> 
> How can the labor that I do myself be stolen from someone else?

That's contained in the smiley.
People tend to sell others' work as their own. I'm not saying you
do. Some licenses permit this, some licenses ask to not mention,
others require that you mention others' work.
In case people complain that they cannot "overwrite" the terms
and conditions of the GPLicense with their own, it is only fair
to ask whether they intend to sell others' work as their own.
I did use the word "ask" knowing that I didn't really
ask in the sense of asking, asking you not to question my use
of asking on the ground of questionable monosemy.

> I reply to words that people speak or write.

Natural language not only invites interpretation, the fact that
it cannot be formalized, let alone be coverd by {T, F}, is central
to some of the most precious pieces of human heritage.

Re: GNAT GPL 2005: Too clever by half?

[Georg Bauhaus]

Chad R. Meiners wrote:
> From: Georg Bauhaus
> 
> <quote>
> Chad Meiners has told us an example. He dislikes RMS,
> and this reflects in part on the GPL.
> <\quote>
> 
> I don't dislike RMS!

OK, sorry. I was driven to this quick conclusion from your statement that RMS
must be nuts, and that you seem to disagree with some of his ideas.

Re: GNAT GPL 2005: Too clever by half?

[Chad R. Meiners]

From: Georg Bauhaus
<quote>
That's contained in the smiley.
People tend to sell others' work as their own. I'm not saying you
do. Some licenses permit this, some licenses ask to not mention,
others require that you mention others' work.
In case people complain that they cannot "overwrite" the terms
and conditions of the GPLicense with their own, it is only fair
to ask whether they intend to sell others' work as their own.
I did use the word "ask" knowing that I didn't really
ask in the sense of asking, asking you not to question my use
of asking on the ground of questionable monosemy.
<\quote>

Oh come on.  I haven't even complained about not being able to
overwrite the GPL.  You are jumping the gun and assuming that I have
all these problems with the GPL.  Just because I am critical of RMS
doesn't mean that I am against various projects that he works on.  I am
an avid user of emacs for example.

However, the GPL doesn't prevent people from stealing your labor.  It
might prevent someone from taking you labor and hiding it under a
bucket, but it does not prevent thief.

<quote>
Natural language not only invites interpretation, the fact that
it cannot be formalized, let alone be coverd by {T, F}, is central
to some of the most precious pieces of human heritage.
<\quote>

Just because we can't formalize natural language does not mean that we
do not need to be careful when interpreting it.  If anything it means
that we need to be more precise, and that we should ask for
clarification more often.

This is why I am critical of statements like "The GPL simply prevents
others from stealing your labor."  Such statements are so imprecise
that they are useless.  Sure it sounds like great propaganda, but the
statement isn't actually true.

If you want to discuss this more, let's do it over email.

-CRM

Re: GNAT GPL 2005: Too clever by half?

[rleif]

Those of you attending SIGAda 2005 in Atlanta
(http://www.acm.org/sigada/conf/sigada2005/)can continue this
discussion at the Commercializing Ada workshop on Wednesday, November
16.
A presentation on the Progresssive Software approach will be given by
Robert Leif with material created by Marius Alves.
Bob Leif

Chad  R. Meiners wrote:
> From: Georg Bauhaus
> <quote>
> That's contained in the smiley.
> People tend to sell others' work as their own. I'm not saying you
> do. Some licenses permit this, some licenses ask to not mention,
> others require that you mention others' work.
> In case people complain that they cannot "overwrite" the terms
> and conditions of the GPLicense with their own, it is only fair
> to ask whether they intend to sell others' work as their own.
> I did use the word "ask" knowing that I didn't really
> ask in the sense of asking, asking you not to question my use
> of asking on the ground of questionable monosemy.
> <\quote>
>
> Oh come on.  I haven't even complained about not being able to
> overwrite the GPL.  You are jumping the gun and assuming that I have
> all these problems with the GPL.  Just because I am critical of RMS
> doesn't mean that I am against various projects that he works on.  I am
> an avid user of emacs for example.
>
> However, the GPL doesn't prevent people from stealing your labor.  It
> might prevent someone from taking you labor and hiding it under a
> bucket, but it does not prevent thief.
>
> <quote>
> Natural language not only invites interpretation, the fact that
> it cannot be formalized, let alone be coverd by {T, F}, is central
> to some of the most precious pieces of human heritage.
> <\quote>
>
> Just because we can't formalize natural language does not mean that we
> do not need to be careful when interpreting it.  If anything it means
> that we need to be more precise, and that we should ask for
> clarification more often.
>
> This is why I am critical of statements like "The GPL simply prevents
> others from stealing your labor."  Such statements are so imprecise
> that they are useless.  Sure it sounds like great propaganda, but the
> statement isn't actually true.
> 
> If you want to discuss this more, let's do it over email.
> 
> -CRM

Re: GNAT GPL 2005: Too clever by half?

[rleif]

Those of you attending SIGAda 2005 in Atlanta
(http://www.acm.org/sigada/conf/sigada2005/)can continue this
discussion at the Commercializing Ada workshop on Wednesday, November
16.
A presentation on the Progresssive Software approach will be given by
Robert Leif with material created by Marius Alves.
Bob Leif

Chad  R. Meiners wrote:
> From: Georg Bauhaus
> <quote>
> That's contained in the smiley.
> People tend to sell others' work as their own. I'm not saying you
> do. Some licenses permit this, some licenses ask to not mention,
> others require that you mention others' work.
> In case people complain that they cannot "overwrite" the terms
> and conditions of the GPLicense with their own, it is only fair
> to ask whether they intend to sell others' work as their own.
> I did use the word "ask" knowing that I didn't really
> ask in the sense of asking, asking you not to question my use
> of asking on the ground of questionable monosemy.
> <\quote>
>
> Oh come on.  I haven't even complained about not being able to
> overwrite the GPL.  You are jumping the gun and assuming that I have
> all these problems with the GPL.  Just because I am critical of RMS
> doesn't mean that I am against various projects that he works on.  I am
> an avid user of emacs for example.
>
> However, the GPL doesn't prevent people from stealing your labor.  It
> might prevent someone from taking you labor and hiding it under a
> bucket, but it does not prevent thief.
>
> <quote>
> Natural language not only invites interpretation, the fact that
> it cannot be formalized, let alone be coverd by {T, F}, is central
> to some of the most precious pieces of human heritage.
> <\quote>
>
> Just because we can't formalize natural language does not mean that we
> do not need to be careful when interpreting it.  If anything it means
> that we need to be more precise, and that we should ask for
> clarification more often.
>
> This is why I am critical of statements like "The GPL simply prevents
> others from stealing your labor."  Such statements are so imprecise
> that they are useless.  Sure it sounds like great propaganda, but the
> statement isn't actually true.
> 
> If you want to discuss this more, let's do it over email.
> 
> -CRM

Re: GNAT GPL 2005: Too clever by half?

[Hyman Rosen]

Chad  R. Meiners wrote:
> When you invoke a property established by a theorem, are you making a
> copy of that theorem.  No, you are not.  This is beside the point.

Yes you are. Whether the proof includes existing results by
reference rather than by value is irrelevant, just as it is
(mostly) in your favorite programming language. The fact is
that without the intense labor of the people whose proofs
are being used, further work based upon them could not be
done. And yet those people remain uncompensated by the users.

> Your statement is self-contradictory.

No. For example, I am compensated for my labor as a programmer.
But I work for a bank, so my work is for-hire. I do not control
what happens to my work, or who may use it. I will not receive
any further compensation should my work be incorporated into a
program which my employer might choose to sell. And yet, I do
not feel ill-used. Quite the contrary in fact.

> Furthermore, the quote "Stallman would simply deny you the
> ability to prevent others from using the product of your
> labor." reads as "Stallman prevents you from protecting
> yourself from the theft of your labor", which is a pretty
> serious violation of our rights.

But there is no theft involved. Stallman is advocating a
condition which he would like to see exist. Were it to be so,
programmers would know ahead of time that they could not control
how their work is used by others. Then they could freely choose
to participate or not. As it stands, I feel that it is theft
that the governments take my tax money and use it to prevent
people from copying protected works. Why should my money go to
assure someone else's profits?

Re: GNAT GPL 2005: Too clever by half?

[tmoran]

> people from copying protected works. Why should my money go to
> assure someone else's profits?
   Because their tax money goes to enforce other laws, which
protect the profits with which your employer pays you?
   Because that's the way it's done in this experiment called
America?

Re: GNAT GPL 2005: Too clever by half?

[Hyman Rosen]

tmo...@acm.org wrote:
>    Because their tax money goes to enforce other laws, which
> protect the profits with which your employer pays you?

In pretty much every other case, laws exist to facilitate the
business, but do not create it. It's only in the prevention of
copying at gunpoint where the business would not exist were it
not for the laws. As such, I find it impossible to feel any
sympathy for those who feel entitled to help themselves to the
labor of us all in order to harass, abuse, and jail those who
do not wish to tithe to them.


>    Because that's the way it's done in this experiment called
> America?

There is a great deal of evil in the world that continues to be
done as it's always been done. That is hardly a good argument
for its continuation.

Re: GNAT GPL 2005: Too clever by half?

[Ole-Hjalmar Kristensen]

>>>>> "GB" == Georg Bauhaus <bauhaus@futureapps.de> writes:

    GB> Simon Wright wrote:
    >> It certainly seems to be true that, no matter how daft you or I might
    >> think it, management would prefer to use Windows to develop on.
    >> I wonder  how difficult it is  to build GNAT on  Windows? I'm pretty
    >> sure

    >> it has to be under MinGW. I did manage after many hours to get PolyORB
    >> to build under MinGW, but it crashed dismally when used .. in contrast
    >> to the Linux build, which worked out of the box. Perhaps it would be
    >> worth a try .. on the other hand, I would like a little bit of a life!

    GB> Maybe "Solaris" is the magic word? The new workstations seem comparatively
    GB> affordable...

You can run Solaris on almost any Intel box as well....

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: GNAT GPL 2005: Too clever by half?

[Simon Wright]

Ole-Hjalmar Kristensen <ole-hjalmar.kristensen@substitute_employer_here.com> writes:

>     GB> Simon Wright wrote:
>     >> It certainly seems to be true that, no matter how daft you or I might
>     >> think it, management would prefer to use Windows to develop on.

> You can run Solaris on almost any Intel box as well....

It's the cost of the outsourced support (often more per year than the
cost of the hardware), and the management effort that would be
necessary to beat the outsource supplier down to a sensible price per
seat.

You and I know that Solaris is going to be less trouble to support,
but lack of familiarity breeds fear.

Last time I looked, Solaris was rather picky about the hardware it ran
on .. has that changed?

Re: GNAT GPL 2005: Too clever by half?

[Ole-Hjalmar Kristensen]

I think it is getting better, although on my office PC, the driver for
the on-board sound system does not work properly, for example. For
graphic cards, I think you can use any XFree86 driver. Also, the
latest version of Solaris promises to be compatible with Linux on a
system-call level. A list of supported hardware can be found at

http://www.sun.com/bigadmin/hcl

Also note that Solaris is now open source, but make sure to read the
license....


>>>>> "SW" == Simon Wright <simon@pushface.org> writes:

    SW> Ole-Hjalmar Kristensen <ole-hjalmar.kristensen@substitute_employer_here.com> writes:
    GB> Simon Wright wrote:
    >> >> It certainly seems to be true that, no matter how daft you or I might
    >> >> think it, management would prefer to use Windows to develop on.

    >> You can run Solaris on almost any Intel box as well....

    SW> It's the cost of the outsourced support (often more per year than the
    SW> cost of the hardware), and the management effort that would be
    SW> necessary to beat the outsource supplier down to a sensible price per
    SW> seat.

    SW> You and I know that Solaris is going to be less trouble to support,
    SW> but lack of familiarity breeds fear.

    SW> Last time I looked, Solaris was rather picky about the hardware it ran
    SW> on .. has that changed?

-- 
   C++: The power, elegance and simplicity of a hand grenade.