gmane.comp.lang.caml.inria

Re: Licenses - Confusion

Stefano Zacchiroli — 2013-05-23T22:11:02

Here we're getting *largely* OT for this list, but I bite (SCNR).

The topic you raise here is fairly complex, so I'd be weary of
"certainty" if I were you :) But the problem of the un-readability of
GPLs by random developers, and in particular of GPLs 3.x, is real.
There are interesting experiments in the area to try to improve things.
The most prominent one I'm aware of is copyleft-next; you can read a
recent license release here:

  https://gitorious.org/copyleft-next/copyleft-next/blobs/master/Releases/copyleft-next-0.3.0

Be sure to also read README and DISCLAIMER in the repository before
trying to do anything with the license (even though the license has a
GPL way out, so risks of screwing up very badly are quite low).

A more introductory article on the underlying legal-ish issues is
available at

  https://lwn.net/Articles/537559/

Cheers.

Re: Licenses - Confusion

Chet Murthy — 2013-05-23T21:45:55


-Well-.  Suppose you wrote a really, really useful piece of software.
Big Software Company X (BSCX) starts using it, and they develop some
really nice stuff on top of it, filling a really important gap.  But
let's be honest: your stuff was the key contribution, and they're just
adding a very useful thing to it.  But they patent that useful
addition, and -then- open-source it.

So now, any company that whats to use your software, and get the value
of that nifty addition, has to license the patent from BSCX.  But
that's not at all what you wanted, is it?

I guess what I'm saying is, you might -think- that legalese is there
to employ lawyers, whereas the -lack- of legalese can often leave
loopholes thru which corporate lawyers and their clients will slip.

I'm not saying you should use the GPL.  Just that there's a -good-
reason most corps don't like the GPL: it restricts their ability to
suck value out of open-source without giving back.

--chet--

Re: Licenses - Confusion

Ernesto Posse — 2013-05-23T21:39:19

On Thu, May 23, 2013 at 4:41 AM, Daniel Bünzli
<daniel.buenzli< at >erratique.ch>wrote:

I'm afraid this is a little bit too naive. I am not a lawyer, and I agree
that the GPL is quite unreadable, but the choice of licence does matter,
and has real consequences. A BSD or MIT licence is very permissive, but one
could argue that they is too permissive, with no attribution required for
example. If what you want is "anyone can use it in any way and I don't
care", then BSD-like is fine. GPL, imposes restrictions on derivative
works, which could be argued as limiting freedom. My point is, not all
"copyleft" licences are the same, and the choice depends on your needs and
how you want people to use your software. Personally I like Apache. I find
it much more readable than GPL and without the same derivative works
restrictions.

Re: Licenses - Confusion

Stéphane Glondu — 2013-05-23T21:36:18

Le 23/05/2013 10:41, Daniel Bünzli a écrit :

Humans don't care about licenses. They do whatever is technically
possible, guided by their conscience.

License are there because of bad practices of the software industry
(hidden sources, patents, etc.). Mathematics (theorems and proofs) have
been developped freely of any licences since Antiquity.


Cheers,

Re: Licenses - Confusion

Marek Kubica — 2013-05-23T21:07:35

On Thu, 23 May 2013 09:41:09 +0100
Daniel Bünzli <daniel.buenzli< at >erratique.ch> wrote:


But then there is the other legal framework that makes some
assumptions. That's why the Apache license has some clauses on patents.
I read the GPL2 some time ago and found it not that difficult to
understand, actually.

If you want a BSD-like license for humans, there's the WTFPL. I think
I'd agree with Stéphane about the classification.

regards,
Marek

TPNC 2013: 2nd call for papers

GRLMC - URV — 2013-05-23T19:55:46

*To be removed from our mailing list, please respond to this message with UNSUBSCRIBE in the subject line*

*************************************************************************
2nd INTERNATIONAL CONFERENCE ON THE THEORY AND PRACTICE OF NATURAL COMPUTING

TPNC 2013

Cáceres, Spain

December 3-5, 2013

Organized by:

Computer Architecture and Logic Design Group (ARCO)
University of Extremadura

Research Group on Mathematical Linguistics (GRLMC)
Rovira i Virgili University

http://grammars.grlmc.com/tpnc2013/
*********************************************************************

AIMS:

TPNC is a conference series intending to cover the wide spectrum of computational principles, models and techniques inspired by information processing in nature. TPNC 2013 will reserve significant room for young scholars at the beginning of their career. It aims at attracting contributions to nature-inspired models of computation, synthesizing nature by means of computation, nature-inspired materials, and information proces

Re: Licenses - Confusion

Daniel Bünzli — 2013-05-23T08:41:09

Le jeudi, 23 mai 2013 à 07:36, Stéphane Glondu a écrit :


Well as everyone seems to know and speak for the others, here's my take.  

Regardless of copyleft -- which I'm certain could be expressed in a BSD-like concise way -- I would personally say that BSD-style licenses are for humans and GPL-style for lawyers.

If you need a lawyer to understand your freedom you are not free anymore. Stop feeding the lawyers and the bureaucrats.

Best,

Daniel

Re: Licenses - Confusion

Stéphane Glondu — 2013-05-23T06:36:02

Le 22/05/2013 19:19, Adrien Nader a écrit :

For native code, checksums are also on the implementation approximations
(such as inlined functions, symbol names for direct access). Changing a
library to keep it binary compatible with that is impossible in practice
with the current tools.

But actually, someone linking a closed-source executable using a pure
LGPL library can just remove all *.cmx files of external libraries when
compiling so that the approximations are not used, and the binary
dependency will only be on interfaces. And, obviously, that person would
need also to distribute all the *.{cma,cmxa,a} files needed to link said
executable.


Keep in mind that programmers may be users but not copyright owners.

I would say BSD-style licenses cater for the needs of organizations, and
GPL-style ones cater for individuals.


Cheers,

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Daniel Bünzli — 2013-05-22T17:49:34


Le mercredi, 22 mai 2013 à 18:30, Florent Monnier a écrit :


Again, follow the instructions here, it will work perfectly:

https://github.com/OCamlPro/opam/issues/493#issuecomment-13950204

No need to touch the GPU memory. Set the swap size to something like 2G to compile ocaml and then set it back to the value it was initially (I'm sorry I don't have a system under hand to clearly state what variable but I think the variable to edit should something like CONF_SWAPSIZE=2048).  

Check the status with something like:  

 swapon -s  
 free -h  
  
Best,

Daniel

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Florent Monnier — 2013-05-22T17:30:49

2013/05/22, Daniel Bünzli:

In /boot/config I put:
gpu_mem_512=192
My swap is twice the RAM (512 * 2) which is what was adviced to my by
the other RPi users in my LUG.

Should I reduce the GPU memory?
Or remove the swap?

--

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Florent Monnier — 2013-05-22T17:15:04

2013/05/22, Alain Frisch:

Here is the error that I get now:

pi< at >raspberrypi ~ $ opam switch 4.01.0dev+trunk

+ ./boot/ocamlrun boot/myocamlbuild byte_stdlib_mixed_mode ocamlc
lex/ocamllex camlp4/Camlp4/Camlp4Ast.partial.ml
camlp4/boot/camlp4boot.byte camlp4/Camlp4.cmo camlp4/Camlp4Top.cmo
camlp4/camlp4prof.byte camlp4/mkcamlp4.byte camlp4/camlp4.byte
camlp4/camlp4fulllib.cma camlp4/camlp4boot.byte camlp4/camlp4boot.cma
camlp4/camlp4r.byte camlp4/camlp4r.cma camlp4/camlp4rf.byte
camlp4/camlp4rf.cma camlp4...[truncated]

+ ./boot/ocamlrun boot/myocamlbuild native_stdlib_mixed_mode ocamlopt
lex/ocamllex ocamlbuild/ocamlbuildlib.cmxa
ocamlbuild/ocamlbuildlightlib.cmxa ocamlbuild/ocamlbuild.native
ocamlbuild/ocamlbuildlight.native
+ STDLIB_MODULES=Pervasives,Arg,Array,Buffer,Char,Digest,Filename,Format,Hashtbl,Lazy,Lexing,List,Map,Printexc,Printf,Scanf,Set,String,Sys,Parsing,Int32,Int64,Nativeint,Obj,Queue,Sort,Stream,Stack

+ ./boot/ocamlrun boot/myocamlbuild -ignore
Pervasives,Arg,Array,Buffer,Char,Digest,File

Re: Licenses - Confusion

Adrien Nader — 2013-05-22T17:19:07

Hi,

I'll start with the usual IANAL (and I'm not the FSF either).

The thing to keep in mind is what the various license attempt to do,
roughly what their main goal is:
- GPL*: a program under the GPL* licenses allow people who execute the
  software to run it, study it, modify it and redistribute it under the
  same rules
- LGPL*: the same but for libraries and only for the libraries and not
  for the executable using the library
- AGPL*: same as GPL* but instead of people who execute it (i.e. on
  their _own_ machine), it's people who use it (remotely for instance)

If you chose the GPL, it's probably that you agreed with its goals so I
won't expand on that.

Now, the LGPL: it has been made mostly with C in mind. The modern way to
distribute software is dynamic libraries: libraries that are not part of
the main executable on disk, are loaded as needed into memory and are
shared among several programs.
Unfortunately, mostly for technical reasons, OCaml uses static and not
dynamic linking.

The LGPL require

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Daniel Bünzli — 2013-05-22T16:48:36

Le mercredi, 22 mai 2013 à 17:18, Marek Kubica a écrit :

I don't know if that changed recently but a few month ago there was no swap in the raspbian distributed from raspberrypi.org (to avoid wearing the sdcard).  

Besides if you actually read the segfault that was in Florent's initial message, it's exactly the one that occurs when the system gets out of memory during camlp4's compilation.  

Best,

Daniel

Re: Licenses - Confusion

Marek Kubica — 2013-05-22T16:40:49

On Wed, 22 May 2013 13:39:29 +0100
Anil Madhavapeddy <anil< at >recoil.org> wrote:


But it is also a completely different, non-copyleft license so it
wouldn't serve the same purpose.

regards,
Marek

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Marek Kubica — 2013-05-22T16:18:02

On Wed, 22 May 2013 15:57:35 +0100
Daniel Bünzli <daniel.buenzli< at >erratique.ch> wrote:


No need to increase swap, I built 4.00.1+rpi (the 4.00.1 with the ARM
patch for the RPi) on a 512 MB Raspberry Pi with the default 100 MB
swap that Raspbian creates.

regards,
Marek

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Daniel Bünzli — 2013-05-22T14:57:35

Le mercredi, 22 mai 2013 à 15:49, Alain Frisch a écrit :

Note Florent that this won't change anything to our original problem. To compile ocaml (camlp4 to be precise) on a raspberry pi you need to increase and activate the swap as suggested in my previous message.  

Best,

Daniel

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Alain Frisch — 2013-05-22T14:49:41

The trunk was temporarily broken.  Please try again!  (if OPAM gets a 
copy on the trunk only once in a while, you might need to wait a little 
longer)

Re: [rpi] opam switch to get ocamlopt on RPi with raspbian

Florent Monnier — 2013-05-22T14:39:39

2013/5/22, Anil Madhavapeddy:
[...]


So this is what I tryed, but got this error:

pi< at >raspberrypi ~ $ opam switch 4.01.0dev+trunk
[...]
boot/ocamlrun boot/ocamlc -nostdlib -I boot -strict-sequence -w
+33..39 -warn-error A -I utils -I parsing -I typing -I bytecomp -I
asmcomp -I driver -I toplevel -c asmcomp/closure.ml
boot/ocamlrun boot/ocamlc -nostdlib -I boot -strict-sequence -w
+33..39 -warn-error A -I utils -I parsing -I typing -I bytecomp -I
asmcomp -I driver -I toplevel -c asmcomp/cmmgen.mli
boot/ocamlrun boot/ocamlc -nostdlib -I boot -strict-sequence -w
+33..39 -warn-error A -I utils -I parsing -I typing -I bytecomp -I
asmcomp -I driver -I toplevel -c asmcomp/cmmgen.ml
make[2]: Leaving directory `/home/pi/.opam/4.01.0dev+trunk/build/ocaml'
make[1]: Leaving directory `/home/pi/.opam/4.01.0dev+trunk/build/ocaml'
### stderr ###
...[truncated]
be selected if the type becomes unknown.
File "asmcomp/cmmgen.ml", line 458, characters 23-35:
Warning 40: Record_float was selected from type Types.record_represen

CICLOPS 2013 - Last Call For Papers

Ricardo Rocha — 2013-05-22T14:07:33

                   CICLOPS 2013 - Call for Papers
                   ==============================

                  13th International Colloquium on
     Implementation of Constraint and LOgic Programming Systems

                 http://akira.ruc.dk/~cth/ciclops13

                Istanbul, Turkey, August 24/25, 2013
                      Co-located with ICLP 2013



Important Dates
===============

Abstract Submission: June 1, 2013
Paper Submission:    June 8, 2013
Notification:        July 1, 2013
Camera-ready:       July 13, 2013
Workshop:      August 24/25, 2013



Topics of Interest
==================

CICLOPS is a well established line of workshops. This will be the 13th
edition in a successful series of workshops which is traditionally
co-located with ICLP. The CICLOPS workshop aims at discussing and
exchanging experience on the design, implementation, and optimization
of constraint and logic programming systems, and other systems based
on logic as a means of expressing computations. Preference w

Re: Licenses - Confusion

Török Edwin — 2013-05-22T13:02:00

I think that only matters if you release a *library* under LGPL.

If you do not add the static linking exception then everyone using your library
would have to release their source code too. (also depending on the LGPL version
used it might be unclear whether static linking is a derived work or not).

The LGPL requires people to allow relinking their applications with newer
versions of your LGPL library. But OCaml has interface/implementation digests,
and as such relinking with a newer version would simply fail.
The only way to allow people to relink is to distribute the entire application's
source code to allow people to recompile and relink with your LGPL library.

So an OCaml library released under LGPL would be almost equivalent to a library
release under GPL. The way around this is with that linking exception.

If your library is under the GPL, or if you are writing an application as
opposed to a library then this doesn't matter as the executable will always be a
derived work and require people to rele

Re: Licenses - Confusion

Anil Madhavapeddy — 2013-05-22T12:39:29

The GPL, with all of its static linking verbiage, and OCaml are
not a good fit.  All of the Mirage libraries now prefer the ISC
license where possible, and the LGPL+linking exception is considered
deprecated for new code.  The last time I spoke to a lawyer about
the LGPL+exception (back in 2008 in Xensource), there were alarm
bells raised about the shoddy legal wording involved.  A BSD-style
license is significantly simpler, as long as it doesn't include the
advertising clause.

Copyright (c) Year(s), Company or Person's Name <E-mail address>

Permission to use, copy, modify, and/or distribute this software for any purpose with or without fee is hereby granted, provided that the above copyright notice and this permission notice appear in all copies.

THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAM