Peter Murray-Rust | 7 Jul 2010 15:16
Picon
Picon
Favicon
Gravatar

Concerns about restrictions imposed by LGPL

I am writing to ask guidance from anyone expert in LGPL and dual licensing.
Take hypothetically that I have written a piece of code PMRCODE which is contributed
to a large project BASE which uses the LGPL licence, the whole of which can be used
as a library LIB. As a result I decide to use the same licence (LGPL) for PMRBASE.

A company COMP wishes to use LIB in its products in a statically linked manner
and asserts that this is not possible with LGPL code. COMP therefore wishes LIB to be
relicensed under a less restrictive licence (e.g. BSD or MIT) or alternatively
to dual licence the code. As PMRCODE is part of LIB, this would require me, as author,
to relicence PMRCODE under the same changed licence strategy.

COMP's argument is that although the LGPL does
allow for inclusion into commercial closed source projects you either have to
distribute the LGPL'd portion as a dynamically linked library
which the end user could replace if desired or a API-based rebuild system. This may
apparently have undesirable commercial consequences for COMP.

I would appreciate clarification of these issues - I appreciate they are complex. I
assumed that the LGPL was a reasonable licence system to use and indeed it's widespread
in the BlueObelisk. However although I'm quite happy for companies to use and resell
my software - that's part of the OpenSource philosophy - I am not so sure I
should change my licence because it is a better business model for a downstream commercial
exploiter.

But I have an open mind and don't want to be obstructive but would like to know the arguments.

P.


--
Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
CB2 1EW, UK
+44-1223-763069
------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Blueobelisk-discuss mailing list
Blueobelisk-discuss@...
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss
Konstantin Tokarev | 7 Jul 2010 16:00
Picon
Favicon

Re: Concerns about restrictions imposed by LGPL

07.07.10, 17:16, "Peter Murray-Rust" <pm286@...>: 
> I am writing to ask guidance from anyone expert in LGPL and dual licensing. 
> Take hypothetically that I have written a piece of code PMRCODE which is contributed
> to a large project BASE which uses the LGPL licence, the whole of which can be used 
>  as a library LIB. As a result I decide to use the same licence (LGPL) for PMRBASE. 
> 
> A company COMP wishes to use LIB in its products in a statically linked manner 
> and asserts that this is not possible with LGPL code. COMP therefore wishes LIB to be 
>  relicensed under a less restrictive licence (e.g. BSD or MIT) or alternatively 
> to dual licence the code. As PMRCODE is part of LIB, this would require me, as author, 
> to relicence PMRCODE under the same changed licence strategy.
> 
> COMP's argument is that although the LGPL does
> allow for inclusion into commercial closed source projects you either have to
> distribute the LGPL'd portion as a dynamically linked library 
> which the end user could replace if desired or a API-based rebuild system. This may 
>  apparently have undesirable commercial consequences for COMP.
> 
> I would appreciate clarification of these issues - I appreciate they are complex. I 
> assumed that the LGPL was a reasonable licence system to use and indeed it's widespread
>  in the BlueObelisk. However although I'm quite happy for companies to use and resell
> my software - that's part of the OpenSource philosophy - I am not so sure I 
> should change my licence because it is a better business model for a downstream commercial
>  exploiter.
> 
> But I have an open mind and don't want to be obstructive but would like to know the arguments.
> 
> P.
> 
> 
> -- 
> Peter Murray-Rust
> Reader in Molecular Informatics
> Unilever Centre, Dep. Of Chemistry
>  University of Cambridge
> CB2 1EW, UK
> +44-1223-763069

Hi Peter,
I don't understand completely what are you asking about. There may be several kinds of question, e.g.
1) May you relicense PMRCODE? - yes, it's your copyright, you may do whatever you want
2) Is it possible to relicense LIB? - yes, if all contributors agree
3) Should you do any relicensing? - if dynamic linking is not an option (I'm in doubt why it could not be), you
need to do it. But you have more ways than just relicense to BSD or MIT or something similar. You may add
exception to LGPL to permit static linking. You may sell commercial license to COMP. Any of these actions
requires agreement from all contributors of relicensed code

Did I guess you problem right?

--

-- 
Regards,
Konstantin

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Paul Emsley | 7 Jul 2010 15:47
Picon
Picon
Favicon

Re: Concerns about restrictions imposed by LGPL

On 07/07/10 14:16, Peter Murray-Rust wrote:
> I am writing to ask guidance from anyone expert in LGPL and dual licensing.
> Take hypothetically that I have written a piece of code PMRCODE which is contributed
> to a large project BASE which uses the LGPL licence, the whole of which can be used
> as a library LIB. As a result I decide to use the same licence (LGPL) for PMRBASE.
>
> A company COMP wishes to use LIB in its products in a statically linked manner
> and asserts that this is not possible with LGPL code. COMP therefore wishes LIB to be
> relicensed under a less restrictive licence (e.g. BSD or MIT) or alternatively
> to dual licence the code. As PMRCODE is part of LIB, this would require me, as author,
> to relicence PMRCODE under the same changed licence strategy.
>
> COMP's argument is that  although the LGPL does
> allow for inclusion into commercial closed source projects you either have to
> distribute the LGPL'd portion as a dynamically linked library
> which the end user could replace if desired or a API-based rebuild system. This may
> apparently have undesirable commercial consequences for COMP.
> I would appreciate clarification of these issues -

I don't see anything to clarify - everything you assert is correct.

> I appreciate they are complex. I
> assumed that the LGPL was a reasonable licence system to use and indeed it's widespread
> in the BlueObelisk. However although I'm quite happy for companies to use and resell
> my software - that's part of the OpenSource philosophy - I am not so sure I
> should change my licence because it is a better business model for a downstream commercial
> exploiter.
>    

I agree with that - the LGPL is bending over backwards (more than) enough.

COMP is saying "bend over backwards more so that I can exploit you 
harder".  Frankly, what a nerve they have!  They get fantastic software 
for free and they give us... what exactly? Where's the balance?

Paul.

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Geoffrey Hutchison | 7 Jul 2010 16:12
Picon
Favicon
Gravatar

Re: Concerns about restrictions imposed by LGPL

A company COMP wishes to use LIB in its products in a statically linked manner
and asserts that this is not possible with LGPL code.
Their assertion is incorrect. The LGPL does not "know" of any difference between static linking and dynamic linking. In either method, the code "mixes" on the binary level.
distribute the LGPL'd portion as a dynamically linked library
which the end user could replace if desired or a API-based rebuild system. This may apparently have undesirable commercial consequences for COMP.
The only requirement of the LGPL is that COMP must distribute the code for LIB, or optionally whatever portion of LIB they actually use. For exempt, if they only use PMRCODE, then they only need to make that available *upon request*. Now some companies (e.g., RedHat) make all their GPL and LGPL code available. But technically, COMP only needs to respond to e-mail or written requests.

Also, if COMP makes code changes to LIB or PMRCODE or whatever, they must make these available to the community.

The main difference between the LGPL and BSD licenses is that the LGPL requires COMP to make the code and any modifications available upon request to the community. The BSD allows Microsoft to take the TCP/IP stack and not distribute any changes they make.

I don't have a chance right now, but the FSF has a nice FAQ about their different licenses, and I'm sure this is one of the common cases addressed.

Best regards,
-Geoff

P.S. While I'm considering making my code available under the BSD license, I generally *like* the LGPL because it requires changes go back to the community. I'm just getting tired of explaining licenses. IANAL.

---
Prof. Geoffrey Hutchison
Assistant Professor, Department of Chemistry
University of Pittsburgh
Office: (412) 648-0492

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Blueobelisk-discuss mailing list
Blueobelisk-discuss@...
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss
Konstantin Tokarev | 7 Jul 2010 16:48
Picon
Favicon

Re: Concerns about restrictions imposed by LGPL

> The only requirement of the LGPL is that COMP must distribute the code for LIB, or optionally whatever
portion of LIB they actually use. For exempt, if they only use PMRCODE, then they only need to make that
available *upon request*. Now some companies (e.g., RedHat) make all their GPL and LGPL code available.
But technically, COMP only needs to respond to e-mail or written requests.

LGPL requires not only availability of sources, but also possibility to relink application with newer
version of library. If static linking is used, COMP will need to provide needed object files and
instructions for building

--

-- 
Regards,
Konstantin

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Geoffrey Hutchison | 7 Jul 2010 17:20
Picon
Favicon
Gravatar

Re: Concerns about restrictions imposed by LGPL


On Jul 7, 2010, at 10:48 AM, Konstantin Tokarev wrote:

> LGPL requires not only availability of sources, but also possibility to relink application with newer
version of library. If static linking is used, COMP will need to provide needed object files and
instructions for building

Right, I guess that's true.

-Geoff

---
Prof. Geoffrey Hutchison
Department of Chemistry
University of Pittsburgh
tel: (412) 648-0492
email: geoffh@...
web: http://hutchison.chem.pitt.edu/

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Craig James | 7 Jul 2010 17:25
Favicon

Re: Concerns about restrictions imposed by LGPL

On 7/7/10 8:20 AM, Geoffrey Hutchison wrote:
>
> On Jul 7, 2010, at 10:48 AM, Konstantin Tokarev wrote:
>
>> LGPL requires not only availability of sources, but also possibility to relink application with newer
version of library. If static linking is used, COMP will need to provide needed object files and
instructions for building
>
>
> Right, I guess that's true.

Only if they ask.  They never do.

Craig

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Peter Murray-Rust | 7 Jul 2010 17:26
Picon
Picon
Favicon
Gravatar

Re: Concerns about restrictions imposed by LGPL

Many thanks for the rapid and valuable response - please keep mailing. The issue appears to be primarily over static vs dynamic linking

P.


--
Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
CB2 1EW, UK
+44-1223-763069

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Blueobelisk-discuss mailing list
Blueobelisk-discuss@...
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss
Paul Emsley | 7 Jul 2010 17:36
Picon
Picon
Favicon

Re: Concerns about restrictions imposed by LGPL

On 07/07/10 15:12, Geoffrey Hutchison wrote:
>> A company COMP wishes to use LIB in its products in a statically linked manner
>> and asserts that this is not possible with LGPL code.
> Their assertion is incorrect.

While formally this is true, I didn't take it to be strict in the sense 
that you seem to have done, I thought that "this is not possible" meant 
that "management does not allow it", i.e. internally, it is "not 
possible" - which could well be the case.

> The LGPL does not "know" of any difference between static linking and 
> dynamic linking. In either method, the code "mixes" on the binary level.

Agreed.

Paul

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
Peter Murray-Rust | 8 Jul 2010 11:47
Picon
Picon
Favicon
Gravatar

Development of Chemistry Laboratory in Second Life

I've mailed and blogged about this before ...

We now have two new members of our AMI project aimed at creating an intelligent fumehood, and where possible as part of the Blue Obelisk community and including SL . I'd like to introduce them 

"Peter Matthews" pm418 at normal place
"Matthew Smith" ms813 at normal place

Peter and Matthew have just finsihed their second year in Natural Sciences (mainly chemistry) at Cambridge and are now working with us over the summer (with support from JISC). They are working on approaches to visualisation in chemistry coordinated by Joe Townsend.

They've got themselves some avatars in SL but would appreciate some guidance as to what the Blue Obelisk has already set up and who are the best people to approach on particular topics.

Maybe this is an opportunity for an introductory meeting of some sort?

P.


--
Peter Murray-Rust
Reader in Molecular Informatics
Unilever Centre, Dep. Of Chemistry
University of Cambridge
CB2 1EW, UK
+44-1223-763069

------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
Blueobelisk-discuss mailing list
Blueobelisk-discuss@...
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss

Gmane