Re: licensing issues

2001-01-14 Thread Russ Allbery

David Grove [EMAIL PROTECTED] writes:

 However, maybe you can find out something for us. Specifically, why
 isn't Perl 5.6 a part of "official" Debian in this latest release, and
 5.005_03 still is? Is Debian slow at getting this out, or is there a
 more obvious reason from the Perl end? (I'm being provocative, not
 insulting.)

Because tons of the most critical stuff in Debian is written in Perl,
including key pieces of its packaging system, making upgrading Perl
without breaking the entire system an... interesting problem.
Particularly given that Perl was previously handled by the alternatives
system and they're trying to switch away from that.

5.6 is in unstable, I believe, and they're busily sorting out the
dependency issues and upgrade issues and it will be out when it's out.

Upgrading Perl on a Debian system is like upgrading libc, and this
particular upgrade, due to internal structural changes, is even more
complicated than that.

-- 
Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/



Re: licensing issues

2001-01-14 Thread Simon Cozens

On Sun, Jan 14, 2001 at 03:27:56AM +, David Grove wrote:
 However, maybe you can find out something for us. Specifically, why isn't
 Perl 5.6 a part of "official" Debian in this latest release, and 5.005_03
 still is?

simon@pembro26 ~/fonts % apt-cache show perl-5.6
Package: perl-5.6
Priority: important
Section: interpreters
Installed-Size: 11708
Maintainer: Darren Stalder [EMAIL PROTECTED]
Architecture: i386
Version: 5.6.0-6.2
Replaces: data-dumper, perl, perl-5.6-base
Provides: data-dumper, perl5
Depends: perl-5.6-base (=5.6.0-6.2), libc6 (= 2.1.97), libdb2, libgdbmg1
Pre-Depends: perl-base (=5.004.04-2) | perl5-base
Suggests: perl-5.6-suid, perl-5.6-debug, perl-5.6-doc (=5.6.0-6.2)
Conflicts: data-dumper (= 2.09-1)
Filename: pool/main/p/perl-5.6/perl-5.6_5.6.0-6.2_i386.deb
Size: 2767736

-- 
Also note that i knew  _far_ more about the people that call address
mungers names like 'lusers', 'egoists' or try to make luser giraffes.
-- Megahal (trained on asr), 1998-11-06



Re: licensing issues

2001-01-14 Thread Ben Tilly

Dave Rolsky [EMAIL PROTECTED] wrote:

On Sun, 14 Jan 2001, David Grove wrote:

  Ladies and gentlemen, maybe licensing isn't the method of choice of
  preventing the abuses that are harming this community, but it seems to 
be
  the appropriate place to affect at least one of the two:

What abuses?  What the heck are you talking about?

Why are people suddenly freaking out about licenses?  Perl's dual license
seems to have served it well, since AFAIK there has never been a major
issue where someone has been in violation of the license (intentional or
otherwise).

Actually there was a serious issue before the OnePerl
effort.  ActiveState was shipping binaries based on
5.003 except modified in ways that not only were not
public, but could not be made public because they
involved intellectual property owned by Microsoft.
Sarathy had his 5.004 (_02 IIRC) based binaries that
he ported and compiled.  Where you could use them, they
were preferable.  But they could not be used with a lot
of ActiveState modules.

At the time there was considerable illwill towards
ActiveState, and disagreement about whether or not
ActiveState was in violation of the Artistic license.
They claim to this day that they were not.  Reading the
license I think that they were in violation of section 3.
(They come closest to satisfying section 3c, but did not
ship unmodified executables.)

That situation definitely had ActiveState violating the
spirit of the Artistic License, whether or not they were
violating the letter.  And resolving that situation took
a lot of work, including a substantial investment by
O'Reilly in ActiveState.

I know that after that I would have been far slower to
ever trust ActiveState were it not for the vocal support
of people like Sarathy.  Even so the bad taste of having
been a Windows user back then (this was around the same
time that I started with Perl) has left me willing to
believe the worst of ActiveState.

What problem are you aiming to fix?

  2) The existing policies (or lack thereof) or lack of attention or 
concern
  allow a(ny) company to purchase strong control in the development and
  direction of the Perl language for proprietary goals (which is why I 
asked
  the question about 5.005_03... linux distros are outright rejecting it
  everywhere, as is FreeBSD).

Uh-oh, here comes the ActiveState rant.  AFAICT, your ActiveState
hostility mostly stems from your assertion that they were responsible for
the release schedule of 5.6.0.  This has two big assumptions:

What David has said to me in the past indicates that the
majority of his hostility towards them results from his
having worked there and seen things that got him really
upset.

1.  That it was released too early.  I happen to agree but that's largely
irrelevant.

I think it needed to be released for other reasons.

But as I have said before, I have no problems with 5.6.0
having been released when it was.

2.  That AS somehow had a vested interest in this early release and
knowingly forced a buggy 5.6.0 on the community.

FWIW David was working at ActiveState when the 5.6.0
release happened.  He claims inside information.

Even agreeing with #1, I have yet to see any evidence on #2.

David's claim is that for a company trying to establish
itself as the One True Source of Perl, it is good when
their binaries are objectively better than building from
source.  In my eyes this is plausible, but the claim is
not proven.

And mind you, I am very anti-corporate and skeptical of all that
corporations do.  Yet I still think you have no case.

I am perhaps a little more sympathetic because I was a
Windows user during previous issues with ActiveState.

Note, that I am sympathetic but not totally convinced.
Were I to only take into account what David has said in
private to me, I would be convinced.  But I hate ever
making my mind up after only hearing one side of the
story.  Clearly while working there something happened
that got him very upset.  But right now there is
something of the feel of a "he said, she said" fight
after a marriage breaks up...

Cheers,
Ben

_
Get your FREE download of MSN Explorer at http://explorer.msn.com




Re: licensing issues

2001-01-14 Thread Chris Nandor

At 09.19 -0500 01.14.2001, Ben Tilly wrote:
That situation definitely had ActiveState violating the
spirit of the Artistic License, whether or not they were
violating the letter.

They violated neither the spirit nor the letter.

-- 
Chris Nandor  [EMAIL PROTECTED]http://pudge.net/
Open Source Development Network[EMAIL PROTECTED] http://osdn.com/



Re: licensing issues

2001-01-14 Thread Simon Cozens

On Sun, Jan 14, 2001 at 09:27:28AM -0500, Chris Nandor wrote:
 At 09.19 -0500 01.14.2001, Ben Tilly wrote:
 That situation definitely had ActiveState violating the
 spirit of the Artistic License, whether or not they were
 violating the letter.
 
 They violated neither the spirit nor the letter.

Incorrect. Indeed, the entire point of the OnePerl thingy was to
resolve the violation!

If what you say is true, the whole discussion can be cleaned up by you telling
us which part of clause three they actually did fulfil:

a) made their modifications freely available.
b) used package internally only.
c) provided *both* modified and unmodified versions with separate names
   and separate documentation.
d) make other arrangements with Larry.

They certainly didn't do a, b, or c. So that leaves d.

-- 
"IT support will, from 1 October 2000, be provided by college and
departmental card locks." - J-P Stacey



Re: licensing issues

2001-01-14 Thread Ben Tilly

Chris Nandor [EMAIL PROTECTED] wrote:

At 09.19 -0500 01.14.2001, Ben Tilly wrote:
 That situation definitely had ActiveState violating the
 spirit of the Artistic License, whether or not they were
 violating the letter.

They violated neither the spirit nor the letter.

They were shipping something that they marketed as Perl,
which behaved differently than Perl, had been integrated
into other projects, and for which Larry Wall had little
or no input.  If you do not see that as violating Larry's
artistic control, then our understanding of artistic
control is so different that we might as well not be
speaking the same language.

Had ActiveState done everything that they did except
called their product "winperl", I would have had little
or no objection to their actions.

Regards,
Ben
_
Get your FREE download of MSN Explorer at http://explorer.msn.com




Re: licensing issues

2001-01-14 Thread Chris Nandor

At 15.27 + 01.14.2001, Simon Cozens wrote:
On Sun, Jan 14, 2001 at 09:27:28AM -0500, Chris Nandor wrote:
 At 09.19 -0500 01.14.2001, Ben Tilly wrote:
 That situation definitely had ActiveState violating the
 spirit of the Artistic License, whether or not they were
 violating the letter.

 They violated neither the spirit nor the letter.

Incorrect. Indeed, the entire point of the OnePerl thingy was to
resolve the violation!

No.  It was to have Windows support built-in to the standard distribution.

-- 
Chris Nandor  [EMAIL PROTECTED]http://pudge.net/
Open Source Development Network[EMAIL PROTECTED] http://osdn.com/



Re: licensing issues

2001-01-14 Thread Simon Cozens

On Sun, Jan 14, 2001 at 10:43:36AM -0500, Chris Nandor wrote:
 No.  It was to have Windows support built-in to the standard distribution.

I see.

I notice that you still haven't told me which part of clause three they
actually kept.

-- 
In this talk, I would like to speculate a little, on ... the development
of intelligent life. I shall take this to include the human race, even
though much of its behaviour throughout history has been pretty
stupid... - Stephen Hawking



Re: licensing issues

2001-01-14 Thread John van V

Ben Tilly Wrote:
 But as I have said before, I have no problems with 5.6.0
 having been released when it was.

I work in a 16 trillion dollar settlement environment.  5.5.4/5.6 has 
broken a lot of administrative tools.

You do the math.





Re: licensing issues

2001-01-14 Thread Dave Rolsky

On Sun, 14 Jan 2001, David Grove wrote:

 1. What if a company, ANY company, whether through collusion or by any
 other means, historically has had, currently has, or in the future will
 have, the ability to disregard the perl license mechanism as it stands
 because of questionable "grammar", or the spirit of the licenses because
 of unstated "spirit"? (Forget for a moment that it's now been discussed as
 historical fact, and keep it in the abstract.)

That does seem to be a good argument for tightening up the AL, which has
been discussed and submitted to Larry as an RFC, I believe.

 2. What if a company, ANY company, hires key members of whatever governing
 Perl body exists, for the specific purpose of affecting public opinion
 about that company and controlling the development of the Perl language;
 and that company can affect public opinion concerning itself and its
 actions due to control of "public" media; and that company can affect
 elite (not elitist) opinions due to misguided devotion to those key
 members?

This seems to me to be a problem of the community, rather than a license.
If members of the community working on the core of Perl allow themselves
to be bought and sold, _AND_ nobody else in the community complains, then
we've gotten what we deserve.  I would suggest that the proper way to
handle this is for the community to be self-policing.  If someone in a
position of influence in the community is obviously acting in the best
interests of their employer (without taking into account the community's
interests) then they should be asked to leave the community.

I just don't see how this particular problem could be solved through
licensing.


-dave

/*==
www.urth.org
We await the New Sun
==*/




Re: licensing issues

2001-01-14 Thread David Grove


Dave Rolsky [EMAIL PROTECTED] wrote:

  On Sun, 14 Jan 2001, David Grove wrote:
 
   1. What if a company, ANY company, whether through collusion or by
any
   other means, historically has had, currently has, or in the future
will
   have, the ability to disregard the perl license mechanism as it
stands
   because of questionable "grammar", or the spirit of the licenses
because
   of unstated "spirit"? (Forget for a moment that it's now been
discussed
  as
   historical fact, and keep it in the abstract.)
 
  That does seem to be a good argument for tightening up the AL, which
has
  been discussed and submitted to Larry as an RFC, I believe.
 
   2. What if a company, ANY company, hires key members of whatever
  governing
   Perl body exists, for the specific purpose of affecting public
opinion
   about that company and controlling the development of the Perl
language;
   and that company can affect public opinion concerning itself and its
   actions due to control of "public" media; and that company can affect
   elite (not elitist) opinions due to misguided devotion to those key
   members?
 
  This seems to me to be a problem of the community, rather than a
license.
  If members of the community working on the core of Perl allow
themselves
  to be bought and sold, _AND_ nobody else in the community complains,
then
  we've gotten what we deserve.  I would suggest that the proper way to
  handle this is for the community to be self-policing.  If someone in a
  position of influence in the community is obviously acting in the best
  interests of their employer (without taking into account the
community's
  interests) then they should be asked to leave the community.
 
  I just don't see how this particular problem could be solved through
  licensing.

I'm suggesting licensing only as a necessary first step. It's a document
where we put on paper (or in bits and bytes) what the nature of our
"spirit" is. Without this as a groundwork, there's very little to base
further action and policy on.

All law in my country (the United States) is, in one way or another, based
upon a single document, our Constitution. However, that document is based
upon a previous document which is equally important, in that it expresses
the nature of our "spirit", our Declaration of Independence from England.
(Recent events have forced us to turn to other early writings, however, to
further discover the nature of this "spirit".) Many if not most countries
have similar foundations written in paper (or stone or clay tablets or
whatever). These documents speak, in few words, the entire nature of the
cultures whom they represent.

Licensing could effectively express the nature, desires, and spirit of the
Perl community, not only to grant rights, but to say once and for all that
we cannot tolerate abuse of those desires and our kind and generous
spirit.

It's a first step of many that must be taken in order for Perl 6 to be a
cultural phenomenon, or a "Perl for the People".

p





Re: licensing issues

2001-01-14 Thread Ben Tilly

"John van V" [EMAIL PROTECTED] wrote:

Ben Tilly Wrote:
  But as I have said before, I have no problems with 5.6.0
  having been released when it was.

I work in a 16 trillion dollar settlement environment.  5.5.4/5.6 has
broken a lot of administrative tools.

Did you blindly roll it out?

There comes a point where a project has done all of the
internal testing it is likely to do and needs to have
it tested in wider release.  IMO perl 5.6.0 was not
out of line as such a release.  However that release
should be regarded by all wise people as potentially
broken and timely bug fixes are required.

However as Perl goes through the spectrum from testing
to release significant bugs should result in a few point
releases until there is an acceptable release out there.
5.6.0 is not acceptable for that second release.

Release early, release often, and admit to your
mistakes.

You do the math.

I looked at the situation and where I work no Perl 5.6
stuff has been installed.  Upon my advice.  Nor will it
be until there is a release out there which has sat for
at least a month without any bugs that I consider
critical.

Speaking personally the Perl 5.6.0 disaster (and I
consider it no less) has made me a lot more cynical
about Perl and willing to look at switching languages.
I do not currently know whether I will make the Perl 5
to Perl 6 transition...

Cheers,
Ben
_
Get your FREE download of MSN Explorer at http://explorer.msn.com




Making sure Perl means Perl (was Re: licensing issues)

2001-01-14 Thread Bradley M. Kuhn

Ben Tilly [EMAIL PROTECTED] wrote:

 They were shipping something that they marketed as Perl, which behaved
 differently than Perl, had been integrated into other projects, and for
 which Larry Wall had little or no input.

Controling this sort of behavior with a copyright license is very difficult,
as has been discussed here.  I have tried to do my best in the Artistic-2.0
to mitigate this problem as much as it can be mitigated via copyright law,
but we'll never have a perfect solution.

The better solution is to have a trademark on the word "Perl", in Larry's
name, and have the trademark license require that if they call it "Perl", it
really is the canonical Perl implementation.  (I believe I wrote an RFC that
proposed this; it's presumably currently under Larry's advisement).

-- 
Bradley M. Kuhn  -  http://www.ebb.org/bkuhn

 PGP signature


Re: Making sure Perl means Perl (was Re: licensing issues)

2001-01-14 Thread Ben Tilly

"Bradley M. Kuhn" [EMAIL PROTECTED] wrote:

Ben Tilly [EMAIL PROTECTED] wrote:

  They were shipping something that they marketed as Perl, which behaved
  differently than Perl, had been integrated into other projects, and for
  which Larry Wall had little or no input.

Controling this sort of behavior with a copyright license is very 
difficult,
as has been discussed here.  I have tried to do my best in the Artistic-2.0
to mitigate this problem as much as it can be mitigated via copyright law,
but we'll never have a perfect solution.

I still think a copyright that offers a contract (ie the
same structure as the GPL) can do it.  But the result is
complex enough that people didn't like it.

The better solution is to have a trademark on the word "Perl", in Larry's
name, and have the trademark license require that if they call it "Perl", 
it
really is the canonical Perl implementation.  (I believe I wrote an RFC 
that
proposed this; it's presumably currently under Larry's advisement).

This, of course, presupposes that the legal system is
actually capable of providing a solution.  My
impression (not knowing Larry directly) is that he would
not be by personality inclined to seek legal redress
even if it were clearly within his rights to do so.

As long as that is the case, what is written on paper
won't be worth all that much.  (This is not to mention
the rather large gap between how the legal system should
work in theory and how it works in practice.)

Cheers,
Ben
_
Get your FREE download of MSN Explorer at http://explorer.msn.com




Re: no one is asking for Perl to be GPL-only (was Re: licensing issues)

2001-01-14 Thread Russ Allbery

Ben Tilly [EMAIL PROTECTED] writes:
 "Bradley M. Kuhn" [EMAIL PROTECTED] wrote:

 The FSF surely wants Perl to be under a GPL compatible license (and,
 (GPL|SOMETHING) is always GPL-compatible, by default).  I don't think
 the FSF has ever expressed a desire that Perl be GPL-only.  In fact,
 the FSF has a policy of encouraging everyone to always dual-licensing
 (GPL|Artistic) for Perl modules, to encourage uniformity, and avoid
 licensing confusion for those who use lots of Perl modules.

 Could you point me at this policy?

http://www.gnu.org/philosophy/license-list.html#PerlLicense

-- 
Russ Allbery ([EMAIL PROTECTED]) http://www.eyrie.org/~eagle/