www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Phobos licensing

reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Phobos' std.loader contains this language in the source file:

  *   (ii) Any derived versions of this software (howsoever modified)
  *        remain the sole property of Synesis Software.

It doesn't define 'derived' so I suspect this could be interpreted quite 
broadly by the courts, particularly given the "howsoever modified" bit 
which seems to be saying "if you were thinking of interpreting 'derived' 
narrowly -- DON'T."

I don't think draconian licensing terms like these belong in the 
standard library.  I suggest we ask Synesis for more liberal licensing 
terms or else remove the file from Phobos until it can be reimplemented.

--bb
Jun 01 2007
next sibling parent reply Ant <duitoolkit yahoo.ca> writes:
Bill Baxter wrote:

 Phobos' std.loader contains this language in the source file:
 
  *   (ii) Any derived versions of this software (howsoever modified)
  *        remain the sole property of Synesis Software.
 
 It doesn't define 'derived' so I suspect this could be interpreted quite 
 broadly by the courts, particularly given the "howsoever modified" bit 
 which seems to be saying "if you were thinking of interpreting 'derived' 
 narrowly -- DON'T."
 
 I don't think draconian licensing terms like these belong in the 
 standard library.  I suggest we ask Synesis for more liberal licensing 
 terms or else remove the file from Phobos until it can be reimplemented.
 
 --bb


I asked that question 3 years, 1 months and 3 days ago.

I moved to tango now.

Ant
Jun 01 2007
parent reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Ant wrote:

 Bill Baxter wrote:

 Phobos' std.loader contains this language in the source file:

  *   (ii) Any derived versions of this software (howsoever modified)
  *        remain the sole property of Synesis Software.

 It doesn't define 'derived' so I suspect this could be interpreted 
 quite broadly by the courts, particularly given the "howsoever 
 modified" bit which seems to be saying "if you were thinking of 
 interpreting 'derived' narrowly -- DON'T."

 I don't think draconian licensing terms like these belong in the 
 standard library.  I suggest we ask Synesis for more liberal licensing 
 terms or else remove the file from Phobos until it can be reimplemented.

 --bb


 
 I asked that question 3 years, 1 months and 3 days ago.


Glad you're keeping track. ;-)
It was actually brought to my attention by an old NG post somewhere that 
I stumbled across with regard to making D an official GCC language. 
That was also a few years old.

Looks like std.windows.registry is also affected in addition to std.loader.
Interestingly neither module appears in the online docs.  So maybe we're 
  not supposed to consider these modules "official" Phobos members?


 I moved to tango now.


Tango doesn't seem to have either a windows registry module or a dynamic 
library loader.  So if those licenses are your only worry you can just 
delete those two files from your copy of Phobos.

I haven't looked too closely at it, but what's there in std.loader 
doesn't look very difficult to come up with a replacement for.  It seems 
to just be a wrapper around dlopen/dlsym and the equivalent Win32 thing. 
  There are probably a dozen different implementations of similar (and 
some more complete) things out on the web.

I'm jumping the gun here, but here are some thoughts on where one might 
look for a replacement:  Many things that come to mind have a GPL 
license (FOX,GTK,Qt), and so are out.  But wxWidgets has a dynlib module 
with I think an acceptable license,
http://www.wxwidgets.org/manuals/stable/wx_wxdynamiclibrary.html#wxdynamiclibrary
however, the code is pretty messy. Some of the code is here,
http://cvs.wxwidgets.org/viewcvs.cgi/wxWidgets/src/common/dynlib.cpp?rev=1.118&content-type=text/vnd.viewcvs-markup
but apparently the Unix and Windows specific versions are elsewhere.

But it's jumping the gun because I suspect Matthew Wilson might be 
willing to change the license.  At least I'm guessing so, because 
std.openrj is also by him, and it has a reasonable license.

--bb
Jun 01 2007
next sibling parent reply Jari-Matti =?ISO-8859-1?Q?M=E4kel=E4?= <jmjmak utu.fi.invalid> writes:
Bill Baxter wrote:


 I'm jumping the gun here, but here are some thoughts on where one might
 look for a replacement:  Many things that come to mind have a GPL
 license (FOX,GTK,Qt), and so are out.


FOX and GTK+ are licensed under LGPL, not GPL.
Jun 02 2007
parent Bill Baxter <dnewsgroup billbaxter.com> writes:
Jari-Matti Mäkelä wrote:

 Bill Baxter wrote:
 

 I'm jumping the gun here, but here are some thoughts on where one might
 look for a replacement:  Many things that come to mind have a GPL
 license (FOX,GTK,Qt), and so are out.


 
 FOX and GTK+ are licensed under LGPL, not GPL.


Yes but for the purposes of this discussion it's a distinction without a 
difference.  Either way it's no good for Phobos.

--bb
Jun 02 2007
prev sibling next sibling parent reply Walter Bright <newshound1 digitalmars.com> writes:
Bill Baxter wrote:

 But it's jumping the gun because I suspect Matthew Wilson might be 
 willing to change the license.  At least I'm guessing so, because 
 std.openrj is also by him, and it has a reasonable license.


If someone wants to take the lead on this, and either ask Matthew to 
change the license, or replace those modules, I'll be happy with either. 
As it stands, none of them are lynchpins of Phobos, and can be simply 
dropped without affecting the rest.
Jun 02 2007
parent reply =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Walter Bright wrote:


 But it's jumping the gun because I suspect Matthew Wilson might be 
 willing to change the license.  At least I'm guessing so, because 
 std.openrj is also by him, and it has a reasonable license.


 
 If someone wants to take the lead on this, and either ask Matthew to 
 change the license, or replace those modules, I'll be happy with either. 
 As it stands, none of them are lynchpins of Phobos, and can be simply 
 dropped without affecting the rest.


AFAIK, Matthew *did* change the license on those, and GDC followed this.
Not sure about std.openrj, but definitely so with registry and loader...

See DMD Bug #321: http://d.puremagic.com/issues/show_bug.cgi?id=321
and Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues

--anders
Jun 02 2007
parent reply Bill Baxter <dnewsgroup billbaxter.com> writes:
Anders F Björklund wrote:

 Walter Bright wrote:
 

 But it's jumping the gun because I suspect Matthew Wilson might be 
 willing to change the license.  At least I'm guessing so, because 
 std.openrj is also by him, and it has a reasonable license.


 If someone wants to take the lead on this, and either ask Matthew to 
 change the license, or replace those modules, I'll be happy with 
 either. As it stands, none of them are lynchpins of Phobos, and can be 
 simply dropped without affecting the rest.


 
 AFAIK, Matthew *did* change the license on those, and GDC followed this.
 Not sure about std.openrj, but definitely so with registry and loader...
 
 See DMD Bug #321: http://d.puremagic.com/issues/show_bug.cgi?id=321
 and Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues
 
 --anders


Great.  Thanks for that info.  But where is the paper trail showing that 
Matthew agreed?  Is there a newsgroup post you can point to?

Closest thing I could find was this 
http://www.digitalmars.com/d/archives/12879.html
in which Mr. Wilson declares "I don't have a problem with donating my IP 
to Phobos".  But that's not a specific statement that any particular IP 
*has* been donated to Phobos.

--bb
Jun 02 2007
next sibling parent =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Bill Baxter wrote:


 AFAIK, Matthew *did* change the license on those, and GDC followed this.
 Not sure about std.openrj, but definitely so with registry and loader...

 See DMD Bug #321: http://d.puremagic.com/issues/show_bug.cgi?id=321
 and Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues


 
 Great.  Thanks for that info.  But where is the paper trail showing that 
 Matthew agreed?  Is there a newsgroup post you can point to?


Nope, I only know that the license changed from watching the GDC diffs.
GDC 0.18, http://gdcmac.sourceforge.net/diffs/gdc-0.17-to-0.18.diff.gz

It would be good if this was cleared up and sorted out or files deleted.
In fact, I even thought this was a showstopper issue before DMD 1.0...

--anders

PS. Probably easiest to ask David Friedman and Matthew Wilson directly ?
Jun 02 2007
prev sibling parent reply Carlos Santander <csantander619 gmail.com> writes:
Bill Baxter escribió:

 Anders F Björklund wrote:

 Walter Bright wrote:


 But it's jumping the gun because I suspect Matthew Wilson might be 
 willing to change the license.  At least I'm guessing so, because 
 std.openrj is also by him, and it has a reasonable license.


 If someone wants to take the lead on this, and either ask Matthew to 
 change the license, or replace those modules, I'll be happy with 
 either. As it stands, none of them are lynchpins of Phobos, and can 
 be simply dropped without affecting the rest.


 AFAIK, Matthew *did* change the license on those, and GDC followed this.
 Not sure about std.openrj, but definitely so with registry and loader...

 See DMD Bug #321: http://d.puremagic.com/issues/show_bug.cgi?id=321
 and Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues

 --anders


 
 Great.  Thanks for that info.  But where is the paper trail showing that 
 Matthew agreed?  Is there a newsgroup post you can point to?
 
 Closest thing I could find was this 
 http://www.digitalmars.com/d/archives/12879.html
 in which Mr. Wilson declares "I don't have a problem with donating my IP 
 to Phobos".  But that's not a specific statement that any particular IP 
 *has* been donated to Phobos.
 
 --bb


I knew I'd seen this somewhere:
http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmars.D&article_id=31544

-- 
Carlos Santander Bernal
Jun 02 2007
next sibling parent Bill Baxter <dnewsgroup billbaxter.com> writes:
Carlos Santander wrote:

 Bill Baxter escribió:

 Anders F Björklund wrote:

 Walter Bright wrote:


 But it's jumping the gun because I suspect Matthew Wilson might be 
 willing to change the license.  At least I'm guessing so, because 
 std.openrj is also by him, and it has a reasonable license.


 If someone wants to take the lead on this, and either ask Matthew to 
 change the license, or replace those modules, I'll be happy with 
 either. As it stands, none of them are lynchpins of Phobos, and can 
 be simply dropped without affecting the rest.


 AFAIK, Matthew *did* change the license on those, and GDC followed this.
 Not sure about std.openrj, but definitely so with registry and loader...

 See DMD Bug #321: http://d.puremagic.com/issues/show_bug.cgi?id=321
 and Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues

 --anders


 Great.  Thanks for that info.  But where is the paper trail showing 
 that Matthew agreed?  Is there a newsgroup post you can point to?

 Closest thing I could find was this 
 http://www.digitalmars.com/d/archives/12879.html
 in which Mr. Wilson declares "I don't have a problem with donating my 
 IP to Phobos".  But that's not a specific statement that any 
 particular IP *has* been donated to Phobos.

 --bb


 
 I knew I'd seen this somewhere:
 http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmar
.D&article_id=31544 


Beautiful.  Good work.  I sure couldn't find that.

So all that's needed is for Walter to copy openrj's license preamble 
over to std.loader and std.windows.resistry.

--bb
Jun 02 2007
prev sibling parent Walter Bright <newshound1 digitalmars.com> writes:
Carlos Santander wrote:

 I knew I'd seen this somewhere:
 http://www.digitalmars.com/webnews/newsgroups.php?art_group=digitalmar
.D&article_id=31544 


I think that covers it. Thanks for rooting it out!
Jun 02 2007
prev sibling parent =?ISO-8859-1?Q?Anders_F_Bj=F6rklund?= <afb algonet.se> writes:
Bill Baxter wrote:


 It was actually brought to my attention by an old NG post somewhere that 
 I stumbled across with regard to making D an official GCC language. That 
 was also a few years old.


For the record D is not "an official GCC language" because the copyright
on the D front-end won't be signed over to the Free Software Foundation.
(this is a requirement for inclusion in GCC, but doesn't fit with DMD)
Note that e.g. GNU Pascal isn't either, even those it *was* signed over.

So the GDC is now the "GDC D Compiler" rather than the "GNU D Compiler",
but it is still released Free under the GNU General Public License (v2).
The D specification and documentation is NOT included in this, however.
They are: "Copyright (C) 1999-2007 by Digital Mars, All Rights Reserved"

And Phobos is ("mostly") under the zlib/libpng license or Public Domain.
(That is where mostly means: all files except those few with issues...
See Wiki4D: http://www.prowiki.org/wiki4d/wiki.cgi?PhobosLicenseIssues )
Whereas Tango, as noted on their web page, has a dual AFL/BSD license.

--anders

PS.
D version() is still "GNU". After all, it's the GNU Compiler Collection.
So you should still be using version(GNU) to check for the GDC compiler.
Jun 02 2007
prev sibling parent reply Ant <duitoolkit yahoo.ca> writes:
Bill Baxter wrote:

 Phobos' std.loader contains this language in the source file:
 
  *   (ii) Any derived versions of this software (howsoever modified)
  *        remain the sole property of Synesis Software.
 
 It doesn't define 'derived' so I suspect this could be interpreted quite 
 broadly by the courts, particularly given the "howsoever modified" bit 
 which seems to be saying "if you were thinking of interpreting 'derived' 
 narrowly -- DON'T."
 
 I don't think draconian licensing terms like these belong in the 
 standard library.  I suggest we ask Synesis for more liberal licensing 
 terms or else remove the file from Phobos until it can be reimplemented.
 
 --bb


I just downloaded dmd 1.015.
still contains that license.

:(

Ant
Jun 06 2007
next sibling parent reply Pragma <ericanderton yahoo.removeme.com> writes:
Ant wrote:

 Bill Baxter wrote:

 Phobos' std.loader contains this language in the source file:

  *   (ii) Any derived versions of this software (howsoever modified)
  *        remain the sole property of Synesis Software.

 It doesn't define 'derived' so I suspect this could be interpreted 
 quite broadly by the courts, particularly given the "howsoever 
 modified" bit which seems to be saying "if you were thinking of 
 interpreting 'derived' narrowly -- DON'T."

 I don't think draconian licensing terms like these belong in the 
 standard library.  I suggest we ask Synesis for more liberal licensing 
 terms or else remove the file from Phobos until it can be reimplemented.

 --bb


 
 I just downloaded dmd 1.015.
 still contains that license.
 
 :(
 
 Ant


Not to sound callous, but perhaps it's time the community wrote a more open
version of this?  Last I looked at 
std.loader, it was merely a facade for LoadLibrary/dlopen (and related
functions).

-- 
- EricAnderton at yahoo
Jun 07 2007
parent reply Frits van Bommel <fvbommel REMwOVExCAPSs.nl> writes:
Pragma wrote:

 Ant wrote:

 I just downloaded dmd 1.015.
 still contains that license.

 :(

 Ant


 
 Not to sound callous, but perhaps it's time the community wrote a more 
 open version of this?  Last I looked at std.loader, it was merely a 
 facade for LoadLibrary/dlopen (and related functions).


Why would it be time for that, since the original author has given 
permission to distribute the current version under a BSD-style license 
(as noted elsewhere in this thread) and the only thing left to do is to 
actually change the license text in the source files?
Jun 07 2007
parent Pragma <ericanderton yahoo.removeme.com> writes:
Frits van Bommel wrote:

 Pragma wrote:

 Ant wrote:

 I just downloaded dmd 1.015.
 still contains that license.

 :(

 Ant


 Not to sound callous, but perhaps it's time the community wrote a more 
 open version of this?  Last I looked at std.loader, it was merely a 
 facade for LoadLibrary/dlopen (and related functions).


 
 Why would it be time for that, since the original author has given 
 permission to distribute the current version under a BSD-style license 
 (as noted elsewhere in this thread) and the only thing left to do is to 
 actually change the license text in the source files?


Sorry, I missed that post - I wasn't aware of that fact.  In that case, yea,
the license really needs to go.

-- 
- EricAnderton at yahoo
Jun 07 2007
prev sibling parent Walter Bright <newshound1 digitalmars.com> writes:
Ant wrote:

 I just downloaded dmd 1.015.
 still contains that license.


I know, sorry. I'll get it fixed for 016.
Jun 07 2007