What would people think of binary-only software on Linux?

35 views
Skip to first unread message

Tim Smith

unread,
Feb 26, 1993, 5:13:19 AM2/26/93
to
If someone developed applications or drivers for Linux and sold them
in binary-only form, what would the reaction be?

--Tim Smith

Jaakko Hyvatti

unread,
Feb 26, 1993, 9:46:03 AM2/26/93
to
t...@carson.u.washington.edu (Tim Smith) writes:
>If someone developed applications or drivers for Linux and sold them
>in binary-only form, what would the reaction be?

If the applications are compiled with gcc or linked with GNU
libraries you should get a copy of 'GNU Library General Public License',
found in any GNU distributing site in a file called COPYING.LIB and
study it. I cannot see any other restrictions provided that there is no
other copyrighted code involved.

Of course someone here would complain, but he doesn't have to buy your
software. This is a free world, isn't it?

Another point: the idea behind GPL is that there should be no
restrictions on use of any piece of software. If we did not accept
the use of non-GPL software with Linux, we would restrict possible
applications of Linux.

Jaakko
--
main(){puts("This is a most clever signature program hack, isn't it?");}

The Entropic one

unread,
Feb 26, 1993, 1:49:24 PM2/26/93
to
In article <1mkqfv...@shelley.u.washington.edu> t...@carson.u.washington.edu (Tim Smith) writes:
>If someone developed applications or drivers for Linux and sold them
>in binary-only form, what would the reaction be?

Better than not selling it at all :)
I'm still uncertain how the library license applies to jump tables,
so we will be putting out anything commercial as source or a link
kit. (personally I prefer source)
--
------ Linux, Religion, and electronic shopping all in one place!
\ / Repent of Bean eating, MS-DOS, OS/2, and windows NT, no sales man
\ / will hassle you, no priest will sacrifice you. Free 9600 baud download.
\/ Call the Virtual World Information Systems BBS at 508-793-9568

william E Davidsen

unread,
Feb 26, 1993, 3:57:57 PM2/26/93
to
In article <1mkqfv...@shelley.u.washington.edu>, t...@carson.u.washington.edu (Tim Smith) writes:
| If someone developed applications or drivers for Linux and sold them
| in binary-only form, what would the reaction be?

I wouldn't plan on retiring on the profits. At the moment the people
using Linux tend to be pretty rabid about using only free software
(witness the people who refuse to allow support of Diamond video cards)
and/or semi-destitute. So I doubt that you would get a lot of sales, and
someone would probably write a program better than yours and give it
away just to spoil your day.

In short, I don't advise it.

--
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
Windows NT is a *great* program!
It's everything CP/M should have been all along.

Thomas McWilliams

unread,
Feb 26, 1993, 5:30:40 PM2/26/93
to

No, no, no. This is completely against the goals of the Free Software
movement. Binary only distributions take away our freedom. There is
no life in a binary. Life and growth spring from the source code.
There would be no Linux without the hundreds who have contributed
source to utilities, compilers, libraries, device drivers. Please
take your binaries elsewhere.

If you feel like contributing to Linux but you are legally unable
to release your source, do the next best thing. Provide a detailed
textual description of what the device driver must do. Provide
documentation and specifications which will allow someone else to
write a truly free driver. The hardest part about writing a device
driver is gathering the documentation and understanding what must
be done. The coding itself then becomes almost trivial. Yes, for
you to do this would be real contribution. Don't even think about
binaries unless you provide full source code under the GPL.

Below you will find the words of Richard Stallman which appeared
in gnu.misc.discuss on Jan. 19, 1993. He discusses the meaning
of free software. Muse on this.

From: r...@gnu.ai.mit.edu (Richard Stallman) Subject: Copyleft

Copyleft vs Public Domain

Some people ask why the FSF uses copyleft (the General Public
License or GPL) to specify conditions for copying GNU software.
Why not just put the software in the public domain?

The purpose of the GNU project is to give users in general the
freedom to use software in many ways. We want them to be free
to study and change programs, and to cooperate with each other
by sharing programs. This is what we mean by "free software".
The GPL achieves this better than the public domain, because
(1) it prevents the freedom from being stripped off when the
program is distributed, and (2) it takes away the incentive to
be uncooperative by refusing to share an improvement.

Copyleft does prevent certain people from doing what they would
like to do. Those who would like to take GNU software, make
some changes, and call the result their property are not free
to do so. We think this is a good thing.

To understand why, first note that it is not possible for
society to permit "all possible freedom," because some freedoms
are incompatible with others. This is often stated as, "Your
freedom to swing your fist ends where my face begins."

We always resolve the conflicts between freedoms by
prioritizing them. For example, the quotation above implicitly
assumes that the freedom not to be punched is more important
than the freedom to swing a fist.

There is more than one way to apply a concept such as "free" to
the area of software, because there are different choices of
priority. The question is not, which is the true meaning of
"free software", but rather, which of the valid meanings is
best.

The GNU project is based on the idea that the freedom to decide
your own actions with the programs you use--for example,
whether to copy them or change them--is more important than
occasional power over other people's actions.

Making a program proprietary means interfering with the
important freedoms--other people's freedom to study, share and
change the program. This is the software analogue of swinging
the fist through a user's face. Preventing this may bother
those who want to swing the fist. But don't sympathize too
much; you might be one of the users who would get it in the
face.

If not for the GPL, most users of our software would not have
the freedom to redistribute and change it. That is not just
speculation; the examples of X Windows, TeX, and Berkeley's
Unix extensions show that most users of these programs have
only proprietary versions and do not have the freedom to share
or change them. The first authors of these programs did not
themselves take away those freedoms, but did not defend them
either. Where that path leads was clear when the GNU project
was started, and therefore we chose another path.

The GPL also encourages companies which make improved versions
to return their improvements for inclusion in the standard
version. If not for this, GCC and Emacs would not be nearly as
good as they are.

But is this enough justification? That is a fundamental
philosophical question. Some people believe it wrong to place
any restrictions on anyone, ever--even restrictions against
making any other restrictions. Those readers who believe in
pacifism and condemn use of force even to protect innocent
victims would naturally disagree with our approach.

That is not the philosophy of the GNU project, however. We are
not pacifists, and being passive and never saying "No" to
anyone is not our goal. Our aim is positive--to give the users
the freedom to cooperate, which is distinguished from the
freedom to obstruct. That has been the goal ever since the
beginning.

If we put our software in the public domain, then we would have
a great excuse to make. We could say, "Don't blame us if you
have no freedom to share and change this program--it was that
other guy who redistributed it with a nondisclosure license and
no source." But we want to succeed in giving users that
freedom, not prepare excuses for failure. We use the GPL
because it succeeds.
----Richard Stallman

David Fenyes,Neurobiol/Anatomy,5705,7901935

unread,
Feb 26, 1993, 6:12:21 PM2/26/93
to
In article <1993Feb26....@crd.ge.com> davi...@crd.ge.com (bill davidsen) writes:
>In article <1mkqfv...@shelley.u.washington.edu>, t...@carson.u.washington.edu (Tim Smith) writes:
>| If someone developed applications or drivers for Linux and sold them
>| in binary-only form, what would the reaction be?
>
> I wouldn't plan on retiring on the profits. At the moment the people
>using Linux tend to be pretty rabid about using only free software
>(witness the people who refuse to allow support of Diamond video cards)
>and/or semi-destitute. So I doubt that you would get a lot of sales, and
>someone would probably write a program better than yours and give it
>away just to spoil your day.

That's silly. Some may use the best thing available for the price. Linux
falls in that category. Wordperfect for $200 would fit that category for
a WP devotee, as would SigmaPlot for $500. I was perfectly willing to
pay $75 for a binary only distribution of 4-DOS when I used DOS, and
I'd pay reasonable prices for a binary of WordPerfect, SigmaPlot, etc.
for Linux, as long as I knew I could run with future shared libs. (for
example, if the binary were provided with a link kit (wordperfect.a) that
I could link (cc -s wordperfect.a -llibs -o wp) to reproduce it.

If some application vendor wants more stability in the OS, that vendor
can Adopt a version of Linux, bring it to their specifications, put
together and support a standard distribution to send out with their
application, and make source code freely available.

David.
--
David Fenyes dfe...@thesis1.med.uth.tmc.edu
University of Texas Medical School Houston, Texas

Rick Miller, Linux Device Registrar

unread,
Feb 27, 1993, 7:06:06 AM2/27/93
to

What you might do instead is develop your application/driver and sell it to
companies that desire such a product, simply saying "This runs under Linux,
which I'll throw in at cost". Thus, the operating system comes with the
product. How do you think Sun[tm] became so popular?

Sell it as a "Value Added" system. So, your package might look like this:

WAN-Watch network management software ............ $150.00 per site
Linux 1.0.2 operating system ..................... $1.00 service fee
IBM-compatible PC w/ net board (we re-sell) ...... (market value)

So... Go develop your application under Linux. Make it a really *good* one.
(It might even mimic the functionality of a Sun[tm] product, oh, like maybe
SunNet Manager[tm] <hint>.) Then your sales pitch could go something like:

"My network management package can do everything SunNet
Manager[tm] 2.0 can do, AND MORE! And get this...
You don't have to fork out the big bucks for a Sun[tm]!

Oh sure, I've got a version that'll run just fine on a
Sun[tm], but what I'd recommend is the *Linux* version.

The Linux OS will turn a PC that you already *have* into a
powerful, UNIX[tm]-like workstation which will not only be
adequate for running this network management software, but
will also be useful for just about anything you'd use a
generalized UNIX[tm] workstation for, just like a Sun[tm].

So, you can get this obviously superior WAN Management tool
*and* the power of a UNIX[tm]-compatible OS all for a very
reasonable price. Installation of the entire package is
included, of course." :-)

You do that, *and* throw in an IBM Token Ring driver so it can watch
Token Rings too, and you'll have some *VERY* interested clients in
(er, "on") your hands. I guarantee it. Call me. +1 414 221 3403.

(Do you folks have any idea what companies are shelling out for SunNet
Manager[tm] 2.0?!? Just think how many companies would be *CLAMBORING*
to get a mini-network-manager so inexpensive that they could afford to
put one on *EVERY* *SEGMENT* of their WAN!)

Rick Miller <ri...@ee.uwm.edu> | <ri...@discus.mil.wi.us> Ricxjo Muelisto
Occupation: Husband, Father, WEPCo. WAN Mgr., Discus Sys0p, and Linux fan

Thomas McWilliams

unread,
Feb 27, 1993, 5:05:57 PM2/27/93
to

dfe...@thesis1.med.uth.tmc.edu (David Fenyes) writes:

> That's silly. Some may use the best thing available for the
> price. Linux falls in that category. Wordperfect for $200
> would fit that category for a WP devotee, as would SigmaPlot
> for $500. I was perfectly willing to pay $75 for a binary
> only distribution of 4-DOS when I used DOS, and I'd pay
> reasonable prices for a binary of WordPerfect, SigmaPlot, etc.

YOU are using the WRONG operating system. Please run to
ftp-os2.nmsu.edu and FTP your free version of OS/2. Yes it is
free and legal. Then you can happily run SmegmaPlot and all of
your MS_DOS apps. Though if your checkbook is so open, why not
just pop the 90 bucks for the retail version? Hundreds of
vendors will happily sell you applications without source code.
You can even get an OS/2 version of your 4-DOS. Won't that be dandy?
After you get your free version of OS/2, please take the time
to unsubscribe to comp.os.linux. Be happy, and enjoy life.

Thomas

Thomas McWilliams

unread,
Feb 27, 1993, 7:08:30 PM2/27/93
to

ri...@ee.uwm.edu (Rick Miller) writes:

> (Do you folks have any idea what companies are shelling out
> for SunNet Manager[tm] 2.0?!? Just think how many companies
> would be *CLAMBORING* to get a mini-network-manager so
> inexpensive that they could afford to put one on *EVERY*
> *SEGMENT* of their WAN!)

Binary only software distributions do not advance the art.
Secreted source code does not benefit the Linux community at
large. Were it not for the openness of the source code, Linux
would be a mere a toy, if indeed it existed at all. Software
hoarding can not be considered a "contribution". The GPL and
the open source code have made Linux the success that it is.
Cygnus and other commercial interests are quite comfortable
with this open paradigm, and in fact prosper. One need only
pull the source code to GCC and read the list of many
commercial contributors to realize this.

Thomas

Tim Smith

unread,
Feb 27, 1993, 7:46:42 PM2/27/93
to
tmc...@ukelele.GCR.COM (Thomas McWilliams) writes:
>you to do this would be real contribution. Don't even think about
>binaries unless you provide full source code under the GPL.

If I were to provide source code, I would do so by making it public
domain. I either keep it myself (i.e., binary only), or I make it
available to anyone to do anything that they want (i.e., public
domain).

--Tim Smith

Ross Biro

unread,
Feb 27, 1993, 9:17:53 PM2/27/93
to

In article <C32uJ...@ukelele.GCR.COM> tmc...@ukelele.GCR.COM (Thomas McWilliams) writes:
>
> t...@carson.u.washington.edu (Tim Smith) writes:
>>
>> If someone developed applications or drivers for Linux and sold them
>> in binary-only form, what would the reaction be?
>>
>> --Tim Smith
>>
>
>No, no, no. This is completely against the goals of the Free Software
>movement. Binary only distributions take away our freedom. There is
>no life in a binary. Life and growth spring from the source code.
>There would be no Linux without the hundreds who have contributed
>source to utilities, compilers, libraries, device drivers. Please
>take your binaries elsewhere.


A GPL'ed kernel says nothing about the applications used under
it. I for one would like to see commercial binary-only (actually link
kits due to the GLPL) applications. Device drivers are another
matter. Since they need to be linked to the kernel the GPL does come
into effect and source must be provided as well as no limit on
redistribution. Since there is so much confusion I'll try to
summarize the restrictions, however I'm sure I will be simplifying things
to the point that you should read the license yourself.

1) The Linux Kernel is covered by the GPL. That means that
anything distributed as part of the kernel, or to be linked into the
kernel must include an offer of source and no limit on the
redistribution. Also anything that includes code copied out of kernel
source or include files (not those currently in /usr/include/linux) but
those in places like linux/net/tcp and linux/net and ... must also be
distributed with source and with no limits on redistribution.

2) Many of the utilities/applications under Linux are covered
by the GPL so the same restrictions apply to them as apply to the kernel.

3) GCC is covered by the GPL, so the restrictions that apply
to the kernel apply to it. However the GPL says NOTHING about the
code produced by GCC (as long as none of the source was covered by the
GPL) and in fact you may do with it as you please (again within
limitations placed on you by the license for the source you compiled.)
To illustrate this: Lotus 123 for SYSV/386 is rumored to have been
compiled with gcc. The Next compiler is gcc derived hence most
applications on the Next were compiled with gcc. And in fact to
illustrate that Next cared about the GPL, the objective-C that is part
of gcc 2.x came from Next because of the GPL.

4) The Linux C Library and most of the header files are
covered by the GLPL which basically means if you distribute a program
which includes one of the linux header files or is linked to the linux
C library or is linked to crt0.o (basically all programs compiled on
linux even with shared libraries.) you must provide a link kit (or a
means to link in new libraries/new crt0.o) and source for the library.
You also cannot limit redistribution of the library. You can however
put whatever limits you want on the redistribution of the link kit,
it is after all your code.

5) There is no restriction on how you USE the GPL or GLPL covered
applications there is just no warranty, and limits on what you can do
if you redistribute them.


I hope this helps clear up the confusion.

Ross Biro bi...@leland.stanford.edu
Member League for Programming Freedom (LPF)
mail l...@uunet.uu.net to protect your Freedom

Russell Nelson

unread,
Feb 27, 1993, 11:23:34 PM2/27/93
to

If someone developed applications or drivers for Linux and sold them
in binary-only form, what would the reaction be?

Well, they would have to compete with source-code-available versions
of the same things. Plus, drivers that get linked into the kernel
must use the GPL and so must have source code.

-russ <nel...@crynwr.com> What canst *thou* say?
Crynwr Software Crynwr Software sells packet driver support.
11 Grant St. 315-268-1925 Voice | LPF member - ask me about
Potsdam, NY 13676 315-268-9201 FAX | the harm software patents do.

Tim Smith

unread,
Feb 28, 1993, 12:10:38 AM2/28/93
to
bi...@leland.Stanford.EDU (Ross Biro) writes:
> 1) The Linux Kernel is covered by the GPL. That means that
>anything distributed as part of the kernel, or to be linked into the
>kernel must include an offer of source and no limit on the
>redistribution.

Not quite. If I were to make an object file, using tools not covered by
GPL, and I made that object file compatible with whatever object format
is needed for the Linux kernel, I don't have to distribute source
unless I'm actually distributing the kernel. If all I'm doing is selling
or giving out a floppy with my object files, GPL has nothing to do with
it.

--Tim Smith

Ken Wallace

unread,
Feb 28, 1993, 11:10:00 PM2/28/93
to
(Tim Smith) writes:
| If someone developed applications or drivers for Linux and sold them
| in binary-only form, what would the reaction be?


With regard to binary-only applications....

I have developed a bbs/mailer system which will be released in binary form
either as shareware or a demonstration version for DOS, OS/2, SCO & Coherent,
regardless of the method there will be a modest charge to get the full
version/manuals etc but the source code will not be made available... I also
have a Linux version - if there is the slightest risk that I or anyone else
would be subject to a law suit or forced into publishing the sources then I
would not consider distributing it.


This would be a bit ironic since any money raised from it would go towards
financing the bbs, a major cost of which has been collecting Linux to make
it available to those who don't have Internet access.


Do I release the Linux version or not?


Ken

Jeff Kopmanis

unread,
Feb 28, 1993, 11:23:54 PM2/28/93
to
I think the key word is "reasonable" when it comes to price. Paying UP TO
$200 for an application is about anyone's limit. I would like to see
single copy prices for WP, 123 and others go towards the $100 range to put
them in the reach of more people. I think this contributes to the software
piracy problem when prices are too high.

I would buy binary-only software if it was worth the price. For example,
there doesn't seem to be any Appletalk support for Linux. I would *pay*
for that. But for a device driver, the upper price limit is going to be
lower than an application. Say $50-$75 tops.

Again, "reasonable" is the key.
--
---------------------------------------------------------------------------
Jeff Kopmanis je...@cyberspace.org
(313) 769-6911 (home) (313) 393-4784 (work)

John F Carr

unread,
Feb 28, 1993, 6:13:53 AM2/28/93
to

> 1) The Linux Kernel is covered by the GPL. That means that
>anything distributed as part of the kernel, or to be linked into the
>kernel must include an offer of source and no limit on the
>redistribution.

That is not true. I can produce a .o file to be linked into the kernel and
not provide source, if the source for my .o is not derived from copyrighted
source distributed under the GPL. A kernel containing my .o file is covered
by the GPL, since the rest of the kernel contains code covered by the GPL.
Complete source is not available, so the kernel can not be redistributed,
but that doesn't prevent me from distributing the object file.


--
John Carr (j...@athena.mit.edu)

Ross Biro

unread,
Feb 28, 1993, 1:35:32 PM2/28/93
to

For those in gnu.misc.discuss: The question is weather or not it is ok
to distribute proprietary .o files to be linked into the Linux kernel
(a GPL covered work.) I am arguing that it is not.

In article <1mqu7c...@senator-bedfellow.MIT.EDU> j...@athena.mit.edu (John F Carr) writes:
>In article <1993Feb28.1...@leland.Stanford.EDU>
> bi...@leland.Stanford.EDU (Ross Biro) writes:
>
>> The GPL has never been tested in a court of law, so legally it
>>is an unknown. I do know the intent of the GPL is to include object
>>modules whose primary purpose is to be linked into a GPL'ed package.
>
>Authors of GPL software can claim that intent, and I'm sure RMS would
>appreciate it if people acted that way, but there is no way that my
>software, not derived from other copyrighted source, can be held to the
>terms of the GPL, the legal authority of which derives from copyright.

Here is the relavent sections of the GPL

These requirements apply to the modified work as a whole. If
identifiable sections of that work are not derived from the Program,
and can be reasonably considered independent and separate works in
themselves, then this License, and its terms, do not apply to those
sections when you distribute them as separate works. But when you
distribute the same sections as part of a whole which is a work based
on the Program, the distribution of the whole must be on the terms of
this License, whose permissions for other licensees extend to the
entire whole, and thus to each and every part regardless of who wrote it.

Thus, it is not the intent of this section to claim rights or contest
your rights to work written entirely by you; rather, the intent is to
exercise the right to control the distribution of derivative or
collective works based on the Program.

In addition, mere aggregation of another work not based on the Program
with the Program (or with a work based on the Program) on a volume of
a storage or distribution medium does not bring the other work under
the scope of this License.
...
3. You may copy and distribute the Program (or a work based on it,
under Section 2) in object code or executable form under the terms of
Sections 1 and 2 above provided that you also do one of the following:

a) Accompany it with the complete corresponding machine-readable
source code, which must be distributed under the terms of Sections
1 and 2 above on a medium customarily used for software interchange; or,

b) Accompany it with a written offer, valid for at least three
years, to give any third party, for a charge no more than your
cost of physically performing source distribution, a complete
machine-readable copy of the corresponding source code, to be
distributed under the terms of Sections 1 and 2 above on a medium
customarily used for software interchange; or,

c) Accompany it with the information you received as to the offer
to distribute corresponding source code. (This alternative is
allowed only for noncommercial distribution and only if you
received the program in object code or executable form with such
an offer, in accord with Subsection b above.)

...
4. You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License. Any attempt
otherwise to copy, modify, sublicense or distribute the Program is
void, and will automatically terminate your rights under this License.
However, parties who have received copies, or rights, from you under
this License will not have their licenses terminated so long as such
parties remain in full compliance.
....

7. If, as a consequence of a court judgment or allegation of patent
infringement or for any other reason (not limited to patent issues),
conditions are imposed on you (whether by court order, agreement or
otherwise) that contradict the conditions of this License, they do not
excuse you from the conditions of this License. If you cannot
distribute so as to satisfy simultaneously your obligations under this
License and any other pertinent obligations, then as a consequence you
may not distribute the Program at all. For example, if a patent
license would not permit royalty-free redistribution of the Program by
all those who receive copies directly or indirectly through you, then
the only way you could satisfy both it and this License would be to
refrain entirely from distribution of the Program.


End of GPL quotes

So it looks as if distributing proprietary .o files is acceptable.
However linking them in (being a form of copying) is questionable at
best. I would say that since it makes a copy of GPL covered code
which cannot be redistributed is a violation of the GPL under section
3, and hence makes the .o distributions worthless.

As for the Linux kernel, I may change the licensing agreement on
all modules which I've contributed to explicitly disallow linking to
proprietary modules. This would not affect modules that have been
distributed in the past, but it probably would affect ones in the
future.

Charles Hannum

unread,
Feb 28, 1993, 12:06:31 PM2/28/93
to
> If someone developed applications or drivers for Linux and sold them
> in binary-only form, what would the reaction be?

In practice, you would likely have very few customers, as Linux is
still (and will probably always be) a small niche, often used by people
who are not willing or able to pay for a commercial product.

(This is the story of BSD-derived Unix, too.)

--
\ / Charles Hannum, myc...@ai.mit.edu
/\ \ PGP public key available on request. MIME, AMS, NextMail accepted.
Scheme White heterosexual atheist male (WHAM) pride!

Ross Biro

unread,
Feb 28, 1993, 11:46:06 AM2/28/93
to

The GPL has never been tested in a court of law, so legally it


is an unknown. I do know the intent of the GPL is to include object
modules whose primary purpose is to be linked into a GPL'ed package.

Otherwise I could make any GPL software proprietary by adding a .o
file which contained something like

void proprietary
{
printf ("this software is now proprietary.\n");
}

And putting calls to it all throught the code. I could then limit
redistribution of a GPL covered binary, and the GPL would effectively
be useless.

Message has been deleted

Eric Youngdale

unread,
Feb 28, 1993, 7:37:36 PM2/28/93
to
In article <C32uJ...@ukelele.GCR.COM> tmc...@ukelele.GCR.COM (Thomas McWilliams) writes:
>
> t...@carson.u.washington.edu (Tim Smith) writes:
>>
>> If someone developed applications or drivers for Linux and sold them
>> in binary-only form, what would the reaction be?
>>
>> --Tim Smith
>>
>
>No, no, no. This is completely against the goals of the Free Software
>movement. Binary only distributions take away our freedom. There is
>no life in a binary. Life and growth spring from the source code.
>There would be no Linux without the hundreds who have contributed
>source to utilities, compilers, libraries, device drivers. Please
>take your binaries elsewhere.

I take more of a free market view of things. If someone wants to
try and sell a binary version of something, I say go ahead and try. It is up
to each and every person who uses linux to decide whether they want to buy what
you have or whether they want to use something else which may be free. Now my
personal *preferance* is that the source code be provided for all applications,
so I would be unlikely to ever buy anything unless it is obviously superior
to the free alternatives (and priced reasonably enough that it is not worth
anyones time to improve the free version).

Actually, as I recall there was someone selling something (BBS
software???) for linux already.

-Eric
--

Drew Eckhardt

unread,
Mar 1, 1993, 1:16:04 AM3/1/93
to
>If someone developed applications or drivers for Linux and sold them
>in binary-only form, what would the reaction be?

Applications : IMHO, I'd rather have applications in source
form (not shrouded 'C') so that I can fix any bugs. However,
if it's a choice between no source and no application, I'll
choose the former.

Device drivers : Chances are that you won't have much of a
market for device drivers - there are plenty of hackers
who'll write free drivers for popular hardware, and Linux
isn't seeing wide use in niches where data aquisition,
etc would require special hardware and device drivers.

The legal end of it could be tricky because of the GPL. With the
current system, Linux device drivers are an integral part of the
kernel. This means that the kernel + device drivers is a
derived work. As a derived work, source must be available for
the device drivers.

--
Boycott USL/Novell for their absurd anti-BSDI lawsuit. | Drew Eckhardt
Condemn Colorado for Amendment Two. | dr...@cs.Colorado.EDU
Use Linux, the fast, flexible, and free 386 unix |

Steve Chapin

unread,
Mar 1, 1993, 1:52:06 AM3/1/93
to
}} In article <1993Mar1....@colorado.edu> dr...@juliet.cs.colorado.edu (Drew Eckhardt) writes:
}}
}} The legal end of it could be tricky because of the GPL. With the
}} current system, Linux device drivers are an integral part of the
}} kernel. This means that the kernel + device drivers is a
}} derived work. As a derived work, source must be available for
}} the device drivers.

Question: What in the GPL prevents me from linking the kernel with
the supplied device drivers, and not redistributing it?

Yes, *if* I redistribute the mutated kernel, I have to supply source
to not violate the GPL. And because I only have the binaries, I can't
give the kernel away. That much is obvious. But I think I see people
claiming that I couldn't link a non-GPL'ed driver with the Linux
kernel, *just for my own use*.

Are people actually claiming this to be true? It's too ridiculous...

s...@cs.purdue.edu Steve Chapin ...!purdue!sjc
"If you loose your arrow, you're likely to lose it in the weeds,"
was often heard in days of yore.

Beef: Real food for a dead planet.

Steve Chapin

unread,
Mar 1, 1993, 1:58:33 AM3/1/93
to
}} In article <1msbqm...@bredbeddle.cs.purdue.edu> s...@cs.purdue.edu (Steve Chapin) writes:
}}
}} Are people actually claiming this to be true? It's too ridiculous...

And before I inadvertantly raise hackles, I meant that it would be
ridiculous of the GPL to say this. I'm not making derogatory comments
about any netters...

Ross Biro

unread,
Mar 1, 1993, 2:42:51 AM3/1/93
to
>>For those in gnu.misc.discuss: The question is weather or not it is ok
>>to distribute proprietary .o files to be linked into the Linux kernel
>>(a GPL covered work.) I am arguing that it is not.
>
>Your mistake is an implicit assumption that you can put me under a
>license by using my software.
>
>Assume I do the following:
>
> 1. Write a device driver in C for Linux. I use my own header
> files rather than any Linux header files.
>
> 2. Compile this with Metaware C under DOS.
>
> 3. Run a program of my own that converts object files from
> Metaware's format to Linux's format.
>
>I'm not using any GPL'ed software, and so have not had to agree to GPL
>for anything I've done.
>
>Assume you do the following:
>
> 1. Obtain Linux.
>
> 2. Use it.
>
>You have agreed to GPL and you are bound by it.
>
>Now, I post the binary of my object file to the net. You grab it, and link
>it into Linux.
>
>*You* are still bound by GPL. *You* cannot distribute your resulting version
>of Linux. If you do, and someone asks for source, and you can't give it
>to them, it's you who will get sued[1], not me. The worst that will happen
>to me, legally, is I might be called as a witness.
>
>If the owners of the Linux copyrights wanted to prevent the above
>scenerio, they would have to do something like claim a copyright
>on the driver interface itself.

I maintain that the act of linking the proprietary .o file into
the GPL'ed kernel would be a violation of the GPL. Otherwise I could get
around the GPL by simply distributing my derived work in three parts.

1) gpl.o all the gpl'ed parts ( I would include source for this.)
2) proprietary.o (the proprietary part which you are not
aloud to re-distribute.)
3) an install script to link them together.

This is clearly a violation of the GPL as RMS has mentioned.
Furthermore since I would be providing materials whose sole use and
purpose is to help people to violate a contract, if the copyright
holders on the GPL'ed software didn't sue me, my customers would.
I think one of the best arguments that this is the case is that Next,
Lucid and all the other companies which have modified GPL'ed software
have not tried it. Furthermore this is exactly the case that RMS
talked about in the message that was posted earlier.

Basically the benefits to be gained by attempting to get around the
GPL in such a fasion are not worth the risk.

Tim Smith

unread,
Mar 1, 1993, 1:15:55 AM3/1/93
to
bi...@leland.Stanford.EDU (Ross Biro) writes:
>For those in gnu.misc.discuss: The question is weather or not it is ok
>to distribute proprietary .o files to be linked into the Linux kernel
>(a GPL covered work.) I am arguing that it is not.

Your mistake is an implicit assumption that you can put me under a


license by using my software.

Assume I do the following:

1. Write a device driver in C for Linux. I use my own header
files rather than any Linux header files.

2. Compile this with Metaware C under DOS.

3. Run a program of my own that converts object files from
Metaware's format to Linux's format.

I'm not using any GPL'ed software, and so have not had to agree to GPL
for anything I've done.

Assume you do the following:

1. Obtain Linux.

2. Use it.

You have agreed to GPL and you are bound by it.

Now, I post the binary of my object file to the net. You grab it, and link
it into Linux.

*You* are still bound by GPL. *You* cannot distribute your resulting version
of Linux. If you do, and someone asks for source, and you can't give it
to them, it's you who will get sued[1], not me. The worst that will happen
to me, legally, is I might be called as a witness.

If the owners of the Linux copyrights wanted to prevent the above
scenerio, they would have to do something like claim a copyright
on the driver interface itself.

[1] OK, maybe they would sue me. You can sue anyone for anything, pretty
much. It wouldn't get to trial, though. It would be thrown out after
the first round of interrogatories.

--Tim Smith

Tim Smith

unread,
Mar 1, 1993, 12:59:56 AM3/1/93
to
myc...@hal.gnu.ai.mit.edu (Charles Hannum) writes:
>> If someone developed applications or drivers for Linux and sold them
>> in binary-only form, what would the reaction be?
>
>In practice, you would likely have very few customers, as Linux is
>still (and will probably always be) a small niche, often used by people
>who are not willing or able to pay for a commercial product.
>
>(This is the story of BSD-derived Unix, too.)

Actually, I wasn't thinking of charging. The idea was to take some
commercial stuff for other systems (where source is not usually
distributed) and port it to Linux and give away the binaries.

The hope is that the kind of people who want to use Linux on their own
machines are probably knowledgable Unix hackers who have a lot of
influence in their departments/schools/companies when it comes to
buying software. If they like our stuff on Linux, maybe when someone
wants to know what to use on SCO, Univel, OS/2, Windows, DOS, etc., our
version for SCO, Univel, OS/2, Windows, DOS, etc., might get recommended.
Those we charge for.

--Tim Smith

Eli Boaz

unread,
Mar 1, 1993, 12:57:02 PM3/1/93
to
In article <1msbqm...@bredbeddle.cs.purdue.edu> s...@cs.purdue.edu (Steve Chapin) writes:
>
>Question: What in the GPL prevents me from linking the kernel with
>the supplied device drivers, and not redistributing it?

Nothing. I've read the the GPL backwards and forwards, and it doesn't
say that you can't link a proprietary object file with a GPL product.
You can't do much >after< this point other than run the resulting
program (ie. cannot redistribute it).

>Yes, *if* I redistribute the mutated kernel, I have to supply source
>to not violate the GPL. And because I only have the binaries, I can't
>give the kernel away. That much is obvious. But I think I see people
>claiming that I couldn't link a non-GPL'ed driver with the Linux
>kernel, *just for my own use*.
>
>Are people actually claiming this to be true? It's too ridiculous...

If they are claiming that you can't link the [proprietary, non-source,
commercial, whatever] driver object file with something like the Linux
kernel, then whoever is claiming this needs to re-read the GPL.

And yes, it is ridiculous. . .

--
Reality is but an illusion; he who creates illusions creates realities.
"...sleep my friend and you will see, the dream is my reality..." -- Metallica
------------------------------------------------------------------------------
E. Eli Boaz | rais...@wixer.cactus.org | main(){for(;;) printf("party!!\n");}

Philip Rhoades

unread,
Mar 1, 1993, 4:49:01 PM3/1/93
to
bi...@leland.Stanford.EDU (Ross Biro) writes:

This is how I understood the situation to be - in fact we are in the
process of moving our scientific/multi-lingual WP to linux, coherent,
SCO etc and we will be distributing a freely usable and copyable base
set of files but will charge for extra files that allow PRINTING of
of the fancy character sets. I am not in a position to make available
source code but all the source is our own and will be compiled with
gcc.

... stuff deleted ...

> 4) The Linux C Library and most of the header files are
>covered by the GLPL which basically means if you distribute a program
>which includes one of the linux header files or is linked to the linux
>C library or is linked to crt0.o (basically all programs compiled on
>linux even with shared libraries.) you must provide a link kit (or a
>means to link in new libraries/new crt0.o) and source for the library.
>You also cannot limit redistribution of the library. You can however
>put whatever limits you want on the redistribution of the link kit,
>it is after all your code.

> 5) There is no restriction on how you USE the GPL or GLPL covered
>applications there is just no warranty, and limits on what you can do
>if you redistribute them.


> I hope this helps clear up the confusion.

>Ross Biro bi...@leland.stanford.edu
>Member League for Programming Freedom (LPF)
>mail l...@uunet.uu.net to protect your Freedom

--
Philip Rhoades
First Year Biology
University of Sydney
ph...@biox.bio.su.OZ.AU or phi...@extro.ucc.su.OZ.AU

Kelly E Murray

unread,
Mar 1, 1993, 5:34:46 PM3/1/93
to
In article <1993Feb26.1...@klaava.Helsinki.FI>, hyv...@klaava.Helsinki.FI (Jaakko Hyvatti) writes:
|> t...@carson.u.washington.edu (Tim Smith) writes:
|> >If someone developed applications or drivers for Linux and sold them
|> >in binary-only form, what would the reaction be?
|>
|> If the applications are compiled with gcc or linked with GNU
|> libraries you should get a copy of 'GNU Library General Public License',
|> found in any GNU distributing site in a file called COPYING.LIB and
|> study it. I cannot see any other restrictions provided that there is no
|> other copyrighted code involved.
|>
|> Of course someone here would complain, but he doesn't have to buy your
|> software. This is a free world, isn't it?
|>
|> Another point: the idea behind GPL is that there should be no
|> restrictions on use of any piece of software. If we did not accept
|> the use of non-GPL software with Linux, we would restrict possible
|> applications of Linux.
|>
|> Jaakko

If Linux won't support binary-only, proprietary commercial applications,
then it will never really become a substitute for DOS, SCO-Unix, SUNs, etc.
Remember, an operating system is nothing more than a platform for applications.
If those applications are restricted to GNU-ified, or public-domain
software, it will dramatically limit the potential user-base.

With all the no-cost stuff available, A commerical application would have
to provide significant value to justify its expense and binary-only status.
With the rapidly growing Linux user-base, you may actually see some
commercial software ported to Linux. I would think this would be a good thing.

-Kelly Murray (k...@cis.ufl.edu)
Systems Programmer, University of Florida

Tim Smith

unread,
Mar 2, 1993, 7:51:18 AM3/2/93
to
bi...@leland.Stanford.EDU (Ross Biro) writes:
> I maintain that the act of linking the proprietary .o file into
>the GPL'ed kernel would be a violation of the GPL. Otherwise I could get

Under my scenerio, *I'm* not linking, so even under your interpretation,
*I'm* not violating the GPL. (By the way, we don't even have to look
at GPL to see this. If I'm not using GPL'ed code, then it has no
power over me whatsoever. Person A and person B can't make a contract
or license that tells person C what to do unless they already have some
legal power over person C.)

--Tim Smith

Doug DeJulio

unread,
Mar 1, 1993, 9:07:55 PM3/1/93
to
In article <38...@uflorida.cis.ufl.edu> k...@reef.cis.ufl.edu (Kelly E Murray) writes:
> If Linux won't support binary-only, proprietary commercial
> applications, then it will never really become a substitute for DOS,
> SCO-Unix, SUNs, etc. Remember, an operating system is nothing more
> than a platform for applications. If those applications are
> restricted to GNU-ified, or public-domain software, it will
> dramatically limit the potential user-base.

This is true. It is not universally agreed upon as to whether this is
good or bad, however. I haven't made up my mind myself. One would
hope that by preventing commercial software from being used on Linux,
someone will eventually create a free replacement for it.
--
Doug DeJulio
dd...@cmu.edu

jcb...@gatsibm.larc.nasa.gov

unread,
Mar 2, 1993, 9:20:31 AM3/2/93
to

Thomas,
Chill out...your flame was *totally* uncalled for...try thinking
a little bit first. Linux is an excellent product, but whats the
purpose? Is it supposed to be just a toy that a bunch of hackers can
play with? I don't think so. Many people are turning to Linux and
386BSD because they are tired of paying for the limitations of DOS,
or tired of the outrageous prices charged for 386/486 UNIX. Many of
these people actually *use* their Linux machines to get some work done
and need certain tools to do this...unfortunately some of these tools
are not available in the public domain, such as FrameMaker, IDL,
Mathematica to name a few. It would be *nice* if they were available
via P.D. but until then many would be willing to settle for a
shrink-wrapped version that works under Linux...If this were
to occur, Linux *could* start pulling people away from the commercial
version of UNIX, and therefore have a great impact on the pricing
structure of the commercial UNIX distributions, to the benifit of
*all* UNIX programmers...

Another alternative is to move Linux toward binary compatibility with
one of the commercial versions of UNIX, so that products available
for that platform would work on Linux...

Think...

John
jcb...@gatsibm.larc.nasa.gov

Mike Jagdis

unread,
Mar 2, 1993, 6:00:00 PM3/2/93
to
* In message <C38oL8...@cs.cmu.edu>, Doug DeJulio said:

DD> One would hope that by preventing commercial software from being used
DD> on Linux, someone will eventually create a free replacement for it.

How long will it take a dozen people to write a FrameMaker clone in their
spare time I wonder? :-)

Mike

Thomas Koenig

unread,
Mar 3, 1993, 11:43:37 AM3/3/93
to
tmc...@ukelele.GCR.COM (Thomas McWilliams) writes:


> ri...@ee.uwm.edu (Rick Miller) writes:

>Binary only software distributions do not advance the art.
>Secreted source code does not benefit the Linux community at
>large. Were it not for the openness of the source code, Linux
>would be a mere a toy, if indeed it existed at all. Software
>hoarding can not be considered a "contribution". The GPL and
>the open source code have made Linux the success that it is.
>Cygnus and other commercial interests are quite comfortable
>with this open paradigm, and in fact prosper. One need only
>pull the source code to GCC and read the list of many
>commercial contributors to realize this.

Let's, for example, take Maple V, which, as far as I know, is in the
process of being ported to Linux. I'm using it under MS-DOS and HP-UX
at the moment (under a campus - wide license), and I am very interested
to run it on my work PC under Linux.

So, this happens to advance my "art", in this case chemical engineering,
quite considerably.

If Linux is going to be more than a hacker's toy, it needs commercial
software, or else we are going to run Windows 3.1 or NT for the rest of
our lives :-(
--
Thomas Koenig, ig...@rz.uni-karlsruhe.de, ig...@dkauni2.bitnet
The joy of engineering is to find a straight line on a double
logarithmic diagram.

william E Davidsen

unread,
Mar 4, 1993, 9:15:08 AM3/4/93
to

Just get two dozen lawyers lined up to keep you out of jail after you
do it!

--
bill davidsen, GE Corp. R&D Center; Box 8; Schenectady NY 12345
Windows NT is a *great* program!
It's everything CP/M should have been all along.

John Steele

unread,
Mar 4, 1993, 3:10:39 PM3/4/93
to

>If Linux won't support binary-only, proprietary commercial applications,
>then it will never really become a substitute for DOS, SCO-Unix, SUNs, etc.
>Remember, an operating system is nothing more than a platform for applications.
>If those applications are restricted to GNU-ified, or public-domain
>software, it will dramatically limit the potential user-base.

>With all the no-cost stuff available, A commerical application would have
>to provide significant value to justify its expense and binary-only status.
>With the rapidly growing Linux user-base, you may actually see some
>commercial software ported to Linux. I would think this would be a good thing.

> -Kelly Murray (k...@cis.ufl.edu)
> Systems Programmer, University of Florida

I would have to agree. IMHO, this is *exaclty* what *must* happen, if we are
to see Linux spread far and wide. Some of the first newbie questions I hear
are:
Can I run Lotus 1-2-3?
Will it run MS Word?
etc.

Without commercial vendor support Linux will never reach critcal mass outside
of the student/hacker world. I will not speculate on the much argued
good/bad discussion :>

My .02 you know,
John
--
#include <disclaimer.h> /* route all flames to /dev/null */
/**************************************************************************/
/* Systems Analyst | John Steele email: jst...@netcom.com */
/* Video Business Systems | */
/* 1-800-255-3088 | Most people spell COBOLSUX incorrectly... */
/**************************************************************************/

John Steele

unread,
Mar 4, 1993, 3:19:58 PM3/4/93
to
In <1mvqff...@rave.larc.nasa.gov> jcb...@gatsibm.larc.nasa.gov () writes:

[flame deleted :>]

>Mathematica to name a few. It would be *nice* if they were available
>via P.D. but until then many would be willing to settle for a
>shrink-wrapped version that works under Linux...If this were
>to occur, Linux *could* start pulling people away from the commercial
>version of UNIX, and therefore have a great impact on the pricing
>structure of the commercial UNIX distributions, to the benifit of
>*all* UNIX programmers...

>Another alternative is to move Linux toward binary compatibility with
>one of the commercial versions of UNIX, so that products available
>for that platform would work on Linux...

I would *love* to be able to run some of my *very expensive* Xenix software,
especially after removing Xenix to make space for Linux. This is one thing
that Coherent has going for it. Lets face it, there *is* a large amount
of commercial *nix software already out there on the shelves. I don't
intend to reinvent the wheel myself, even if I had the time (which I don't).

Reinventing just 'tastes' bad to some programmers...

>Think...

Well said (former IBM Almaden Research Center programmer :)

>John
>jcb...@gatsibm.larc.nasa.gov

Nice name too,

Jacques Gelinas

unread,
Mar 6, 1993, 9:08:46 AM3/6/93
to
This is the market game. If I do a commercial package for LINUX
(binary only, restricted license), then there will be pression
to get something either cheaper, or more flexible, developped
by the net. A good exemple of this is pkzip from pkware. They
showed the way, made some money out of it, and were immitated
later (portable zip ...).

If the price is outraging, then the pression will be high.
This is one rease GCC exist on UNIX, while it is less important
on DOS (a free compiler), where compiler are much more affordable.

So this is a good thing. If you like it, you buy it.
One kind of commercial product I think should prosper on
Linux, is shareware. They are generally inexpensive, and
you can try them before paying.

Is there an archive site already accepting shareware for Linux ?


--

--------------------------------------------------------
Jacques Gelinas (jac...@solucorp.qc.ca)
Maintainer of US4BINR jac...@us4binr.login.qc.ca

Message has been deleted

pmki...@flash.lakeheadu.ca

unread,
Mar 7, 1993, 3:39:00 AM3/7/93
to
I was wondering if anybody had ported one of the fido<-> usenet type of
gatways? I got Binkey mostly working, but now I need something to shuffle
the messages back and forth between mail and fido. I looked at rfmail and
fido-use-gw, but they look like a lot of work to port. If anybody had ported one of thes or has an alternative, I would love to hear about it. Note my cnews
wouldn't even have to have to get along with anything else..I am only one ended
(no usenet connection)

Paul K.
pmki...@flash.lakeheadu.ca
1:229/518
-

Louis Lagendijk

unread,
Mar 8, 1993, 5:05:56 AM3/8/93
to
Hi Paul,
I have Rfnews (0.43) working on my own box. I am working on a port of Rfmail 0.5
right now. The only problem I have with Rfmail is that it can only generate
FTSC-0001 packets, while one of my bosses expects packet types 2+. I am working
on that too. I can release my copy of Rfmail 0.43 if people would like to test
it. Be prepared to read the source to get it working... I can include my config
files if that helps.
I have not made any shell scripts to pack/unpack the mail: I call everything
manually.
I am also working on making Rfmail working in a multi-address environment (I am
a point but have two bosses with 2 addresses). With Rfmail 0.43 I solved that with
2 setups, but that is not so desirable.
I am planning to let Rfmail generate Binkley compatible flow files etc., but that
will definately take som more time.

B.t.w. Which Binkley version are you using? The port of the ST version that I did
based on the Next version by Ben Stuyts is about to be released in a consolidated
Next/Linux version. I hope that Ben could test the combined version over the weekend
so that I can release it this week.
If there are more people who are interested in the Rfmail 0.43, please let me know.
If there are people interested in testing the patched 0.5 I would be interested to
know that too.

Cheers, Louis
2:285/311.58 or 2:281/202.5

Dennis Robinson

unread,
Mar 8, 1993, 11:05:21 PM3/8/93
to
I have been thinking about the Linux/ and commercial software issue for
a while myself. Initially I thought it would be cool if commercial soft
ware would be ported to linux. But since I have changed my mind.
One of the problem's with *nix, is there is not enough software support.
Even though there is 10x more than OS2 there is probably 10x less than
dos. It is up to us to develop software, games that will make *nix the
desktop OS of choice for the whole world. So by having Linux stay
free, this will benefit us hackers/programmers that could develop software
nearly for free. Commercial vendors could issue shareware (with source)
so the hackers could do things as improve performance, or get ideas
for another package. Linux is a great thing. Lets make something happen.

Christoph Adomeit

unread,
Mar 9, 1993, 7:00:15 PM3/9/93
to
nel...@crynwr.com (Russell Nelson) writes:

>In article <1mkqfv...@shelley.u.washington.edu> t...@carson.u.washington.edu writes:

>Well, they would have to compete with source-code-available versions
>of the same things. Plus, drivers that get linked into the kernel
>must use the GPL and so must have source code.

a) I would like all those programs like WordPerfect,Vpix, Informix, CorelDraw
on Linux and I'am willing to pay for binary-only versions.
Also, DOS has come so far only because more applications exist than in any
other OS. And they are in general BINARY Versions.
With GPL's Restrictions much less people will use Linux and the effect is
less good software, that causes less users, that again causes less good
applications and so forth..

H. Peter Anvin N9ITP

unread,
Mar 12, 1993, 8:30:26 PM3/12/93
to
In article <C3nC0...@bigcomm.gun.de> of comp.os.linux,

a...@bigcomm.gun.de (Christoph Adomeit) writes:
>
> a) I would like all those programs like WordPerfect,Vpix, Informix, CorelDraw
> on Linux and I'am willing to pay for binary-only versions.
> Also, DOS has come so far only because more applications exist than in any
> other OS. And they are in general BINARY Versions.
> With GPL's Restrictions much less people will use Linux and the effect is
> less good software, that causes less users, that again causes less good
> applications and so forth..

As far as I know, there are no GPL restrictions on how to use Linux,
nor how to produce programs for it (including files compiled with GCC
and linked with the Linux libraries).

The only restriction, I believe, is in distributing Linux itself or
the programs that come with it-- you have to provide source code.
Here obviously a CD-ROM release of Linux with both binaries and a
complete source tree would be *extremely* beneficial.

/hpa

--
INTERNET: h...@nwu.edu TALK: h...@merle.acns.nwu.edu
BITNET: HPA@NUACC IBMNET: 16331@IBMX400
HAM RADIO: N9ITP SM4TKN NeXTMAIL: h...@speedy.acns.nwu.edu
Most inappropriatly named startup command, winner: Microsoft Windows

Reply all
Reply to author
Forward
0 new messages