One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.

I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:

1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.

2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.


The bottom line is the Tango devs should get the final say on what is infringing and what isn't, and we won't relicense infringing code into Phobos without their explicit permission.

I'd much rather never get near legal entanglements/issues, but for copyright inifringement to occur there has to be actual copying. Parallel evolution is allowed.  Of course, the fundamentally boils down to unvalidatable claims on both sides of the issue.

Sigh,
Brad

On Wed, 28 Apr 2010, Walter Bright wrote:

> One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.
> 
> I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:
> 
> 1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.
> 
> 2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.
> 
> 
> The bottom line is the Tango devs should get the final say on what is
> infringing and what isn't, and we won't relicense infringing code into Phobos
> without their explicit permission.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos
> 

Now I'm glad I never looked at Tango. I suggest you post this to the newsgroup and we give up on Shoo's code. I don't empathize with the Tango fellows keeping their precious locked because it's very difficult to frame that action as having D's community interest at heart. To be frank their whole motivation looks petty and political to the extreme, particularly because it's not a rocket science library, it's a God damn date and time routines we're talking about.

To me there's only way out of this: define artifacts that are so much better than Tango's, it would be impossible to them to claim we stole from them.


Andrei

On 04/28/2010 09:38 PM, Walter Bright wrote:
> One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.
>
> I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:
>
> 1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.
>
> 2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.
>
>
> The bottom line is the Tango devs should get the final say on what is
> infringing and what isn't, and we won't relicense infringing code into
> Phobos without their explicit permission.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Years ago, the lawyers told me that "I've never looked at that source code" is an effective defense for parallel evolution. Things get a lot harder if one has. That's why I never look at, say, the gcc source code.

What I'd really like to see is Tango move to the Boost license as well. This will eliminate this regrettable problem of incompatible licenses, which is holding D back. The "two standard libraries" comes up every single time D gets mentioned in other forums, and it's a convenient excuse people use for not looking further at D.


Brad Roberts wrote:
> I'd much rather never get near legal entanglements/issues, but for copyright inifringement to occur there has to be actual copying. Parallel evolution is allowed.  Of course, the fundamentally boils down to unvalidatable claims on both sides of the issue.
>
> Sigh,
> Brad
>
> 

The Boost.Date_Time library seems like a good starting point.

We could port all or parts of it to D without having to worry about licence issues, and it seems to contain all we need and more.  Does anyone have any experience using it?

-Lars


On Wed, 2010-04-28 at 22:59 -0500, Andrei Alexandrescu wrote:
> Now I'm glad I never looked at Tango. I suggest you post this to the newsgroup and we give up on Shoo's code. I don't empathize with the Tango fellows keeping their precious locked because it's very difficult to frame that action as having D's community interest at heart. To be frank their whole motivation looks petty and political to the extreme, particularly because it's not a rocket science library, it's a God damn date and time routines we're talking about.
> 
> To me there's only way out of this: define artifacts that are so much better than Tango's, it would be impossible to them to claim we stole from them.
> 
> 
> Andrei
> 
> On 04/28/2010 09:38 PM, Walter Bright wrote:
> > One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.
> >
> > I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:
> >
> > 1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.
> >
> > 2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.
> >
> >
> > The bottom line is the Tango devs should get the final say on what is
> > infringing and what isn't, and we won't relicense infringing code into
> > Phobos without their explicit permission.
> > _______________________________________________
> > phobos mailing list
> > phobos at puremagic.com
> > http://lists.puremagic.com/mailman/listinfo/phobos
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

On 4/29/10 07:19, Walter Bright wrote:
> Years ago, the lawyers told me that "I've never looked at that source code" is an effective defense for parallel evolution. Things get a lot harder if one has. That's why I never look at, say, the gcc source code.
>
> What I'd really like to see is Tango move to the Boost license as well. This will eliminate this regrettable problem of incompatible licenses, which is holding D back. The "two standard libraries" comes up every single time D gets mentioned in other forums, and it's a convenient excuse people use for not looking further at D.

These two links discuss a possible change of the Tango license:

http://www.dsource.org/projects/tango/forums/topic/786 http://www.dsource.org/projects/tango/ticket/1701


> Brad Roberts wrote:
>> I'd much rather never get near legal entanglements/issues, but for copyright inifringement to occur there has to be actual copying. Parallel evolution is allowed. Of course, the fundamentally boils down to unvalidatable claims on both sides of the issue.
>>
>> Sigh,
>> Brad
>>

Having looked at Boost.Date_Time some more, I see that it's a quite big library, and a lot more than we need.  Still, it is a (safe) place to look so we don't reinvent the wheel.

Here's a very brief overview of the basic design concepts: http://www.boost.org/doc/libs/1_42_0/doc/html/date_time.html#date_time.domain_concepts

Possible D translation, here represented by an unnecessarily verbose example:

        TimePoint t1 = Clock.now;
        ...  // Perform some operation.
        TimePoint t2 = Clock.now;

        TimeInterval t12 = t1 - t2;
        TimeDuration deltaT = t12.duration;

        writeln("Operation started at %s and ended at %s.",
                GregorianCalendar.format(t1, "yyyy-mm-dd"),
                GregorianCalendar.format(t2, "yyyy-mm-dd"));
        writeln("Operation took %s milliseconds to complete.",
                deltaT.milliseconds);

-Lars


On Thu, 2010-04-29 at 10:10 +0200, Lars Tandle Kyllingstad wrote:
> The Boost.Date_Time library seems like a good starting point.
> 
> We could port all or parts of it to D without having to worry about licence issues, and it seems to contain all we need and more.  Does anyone have any experience using it?
> 
> -Lars
> 
> 
> On Wed, 2010-04-28 at 22:59 -0500, Andrei Alexandrescu wrote:
> > Now I'm glad I never looked at Tango. I suggest you post this to the newsgroup and we give up on Shoo's code. I don't empathize with the Tango fellows keeping their precious locked because it's very difficult to frame that action as having D's community interest at heart. To be frank their whole motivation looks petty and political to the extreme, particularly because it's not a rocket science library, it's a God damn date and time routines we're talking about.
> > 
> > To me there's only way out of this: define artifacts that are so much better than Tango's, it would be impossible to them to claim we stole from them.
> > 
> > 
> > Andrei
> > 
> > On 04/28/2010 09:38 PM, Walter Bright wrote:
> > > One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.
> > >
> > > I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:
> > >
> > > 1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.
> > >
> > > 2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.
> > >
> > >
> > > The bottom line is the Tango devs should get the final say on what is
> > > infringing and what isn't, and we won't relicense infringing code into
> > > Phobos without their explicit permission.
> > > _______________________________________________
> > > phobos mailing list
> > > phobos at puremagic.com
> > > http://lists.puremagic.com/mailman/listinfo/phobos
> > _______________________________________________
> > phobos mailing list
> > phobos at puremagic.com
> > http://lists.puremagic.com/mailman/listinfo/phobos
> 
> 
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

Lars Tandle Kyllingstad wrote:
> The Boost.Date_Time library seems like a good starting point.
>
> We could port all or parts of it to D without having to worry about licence issues, and it seems to contain all we need and more.  Does anyone have any experience using it?
>
> 

I have no experience with it, though C++ Boost libraries tend to be of unusually high quality. The compatible license also makes them ideal candidates for conversion to D.

I know people who swear by Boost.Date_Time. I think starting from it is a great idea.

Andrei

On 04/29/2010 09:24 AM, Walter Bright wrote:
>
>
> Lars Tandle Kyllingstad wrote:
>> The Boost.Date_Time library seems like a good starting point.
>>
>> We could port all or parts of it to D without having to worry about licence issues, and it seems to contain all we need and more. Does anyone have any experience using it?
>>
>
> I have no experience with it, though C++ Boost libraries tend to be of
> unusually high quality. The compatible license also makes them ideal
> candidates for conversion to D.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos

That's effing ridiculous.  The D community could really do without this kind of behavior.

Sent from my iPhone

On Apr 28, 2010, at 7:38 PM, Walter Bright <walter at digitalmars.com> wrote:

> One of the Tango developers called me today. There are 5 developers of the Tango time library, and they feel that the Phobos time lib submission is close enough to theirs to be considered an infringement on their license. The Tango license is the BSD license, which does not permit others changing the license, such as to Boost which is the Phobos license.
>
> I am not qualified to compare the two source code bases and make a legal determination if there is infringement or not. And quite frankly, I don't want to split legal hairs about it against the Tango developers' wishes. I've invited the Tango devs to subscribe to this mailing list, and I hope we can come to a resolution:
>
> 1. I think the best solution would be for Tango to relicense the time module under the Boost license, which would require the agreement of the five time module developers. Then, the Phobos version would include them as authors and they'd share in the copyright.
>
> 2. Next would be if the Tango developers who do agree to the Boost license would identify their contributions, those would get authorship & copyright credit, etc. Tango developers who do not agree would identify code they consider infringing, and that code would be removed from the Phobos version, and possibly reimplemented by someone who has not looked at the Tango version.
>
>
> The bottom line is the Tango devs should get the final say on what
> is infringing and what isn't, and we won't relicense infringing code
> into Phobos without their explicit permission.
> _______________________________________________
> phobos mailing list
> phobos at puremagic.com
> http://lists.puremagic.com/mailman/listinfo/phobos