April 29, 2010
Moritz Warning wrote:
> Walter takes any possible copyright taint very serious.

Yes, I do. It is extremely important for D's future that D and Phobos be clear of any intellectual property legal problems.

Phobos is now under the Boost license, which is the most liberal one we could find that was in common use and meets with corporate and lawyer approval. All submissions to Phobos must be under this license.

> Better someone told W about it before he does an emergency blow up of phobos. ;)

Yes, I'm glad to hear about it sooner rather than later.

> Maybe you can talk to the Tango devs to clear up this matter?

I suggest that the Tango devs convert the Tango modules that can get full agreement by their respective devs be converted to the Boost license. The Boost license is free of the legal problems that BSD has, and is compatible with the Phobos license.
April 29, 2010
Andrei Alexandrescu дµ½:

> Hello,
> 
> 
> Following the great work that Masahiro Nakagawa and SHOO offered to Phobos, we are inviting them to join the Phobos developer ranks.
> 
> Once the community review of the proposed libraries concludes with approval, Masahiro and SHOO will be able to commit the modules to Phobos themselves.
> 
> In case you two accept the invitation, please create accounts on dsource.org if you don't have and the let me know of your IDs.
> 
> 
> Andrei

port tango code to phobos will be infringement  ?

please W don't maintain(update)  d1

April 29, 2010
On Thu, 29 Apr 2010 09:24:22 -0700, Walter Bright wrote:

> Moritz Warning wrote:
[..]
>> Maybe you can talk to the Tango devs to clear up this matter?
> 
> I suggest that the Tango devs convert the Tango modules that can get full agreement by their respective devs be converted to the Boost license. The Boost license is free of the legal problems that BSD has, and is compatible with the Phobos license.

As far as I have heard, Tango changed it's license to be compatible with
Phobos in the first place. But Phobos then changed it's license and now
it's incompatible again.
What were the reasons for Phobos to change the license?
I suspect is was discussed before, do you have a link?

thanks,
mwarning
April 29, 2010
Moritz Warning wrote:
> On Thu, 29 Apr 2010 09:24:22 -0700, Walter Bright wrote:
> 
>> Moritz Warning wrote:
> [..]
>>> Maybe you can talk to the Tango devs to clear up this matter?
>> I suggest that the Tango devs convert the Tango modules that can get
>> full agreement by their respective devs be converted to the Boost
>> license. The Boost license is free of the legal problems that BSD has,
>> and is compatible with the Phobos license.
> 
> As far as I have heard, Tango changed it's license to be compatible with Phobos in the first place. But Phobos then changed it's license and now it's incompatible again. 

That is 100% incorrect. Tango always used a more restrictive license than Phobos. Tango has always been able to use Phobos code, but the reverse does not apply.

> What were the reasons for Phobos to change the license?
Phobos was mostly public domain, which has legal problems (eg in Japan).
The boost license is the closest equivalent to public domain.
April 29, 2010
Moritz Warning wrote:
> On Thu, 29 Apr 2010 09:24:22 -0700, Walter Bright wrote:
> 
>> Moritz Warning wrote:
> [..]
>>> Maybe you can talk to the Tango devs to clear up this matter?
>> I suggest that the Tango devs convert the Tango modules that can get
>> full agreement by their respective devs be converted to the Boost
>> license. The Boost license is free of the legal problems that BSD has,
>> and is compatible with the Phobos license.
> 
> As far as I have heard, Tango changed it's license to be compatible with Phobos in the first place.

Tango is originally based on Phobos code, and I gave explicit permission for it to be incorporated into the Tango project & BSD license, but the BSD license does not permit code to flow the other way without the explicit permission of the Tango devs.

Some code has moved back to Phobos, in particular Sean & Don's work, because Sean & Don are the developers of that code and it is their prerogative to do what they please with it.


> But Phobos then changed it's license and now it's incompatible again. What were the reasons for Phobos to change the license?
> I suspect is was discussed before, do you have a link?

Phobos was formerly actually a collection of different licenses, Phobos 1.0 still is. Some was public domain.

The reason it was switched (for Phobos 2) to Boost was:

1. Boost is corporate and lawyer approved, making it a no-brainer for commercial, professional use of Phobos

2. Boost is the most liberal license we were able to find

3. Public domain is not recognized in many countries

4. Having one license for Phobos makes it much easier to manage and deploy

The perennial problem with the BSD license is the binary attribution clause. Tango believes it has a solution to this by embedding the appropriate string in object.d, but I don't know if this has been legally tested and it still puts a constant burden of explanation on the Tango team.

It's just a problem that I can see no reason to adopt.
April 29, 2010
"Walter Bright" <newshound1@digitalmars.com> wrote in message news:hrcbrr$2t7e$1@digitalmars.com...
> Moritz Warning wrote:
>
>> Maybe you can talk to the Tango devs to clear up this matter?
>
> I suggest that the Tango devs convert the Tango modules that can get full agreement by their respective devs be converted to the Boost license. The Boost license is free of the legal problems that BSD has, and is compatible with the Phobos license.

It looks like the Tango devs are pretty much settled on BSD-only with some hack to get around the binary attribution thing: http://www.dsource.org/projects/tango/ticket/1701  (*Shrug*, well, at least it's not as insanely verbose and impenetrable as Apache 2.0...)

I *hate* licenses...(That's why I use the zlib one, none of the public domain problems, all of the freedoms that I've been told Boost offers, and none of Boost's idiotic over-verbosity.)

-------------------------------
Not sent from an iPhone.


April 29, 2010
Nick Sabalausky wrote:
> It looks like the Tango devs are pretty much settled on BSD-only with some hack to get around the binary attribution thing: http://www.dsource.org/projects/tango/ticket/1701

The problem with the hack is everyone trying to use the library professionally will (if one is lucky) ask about the binary attribution thing, and one would have to convince them that the hack takes care of it. This puts a never ending burden on the team.

If one is unlucky, they'll just avoid the library because of that license, and you'll never hear from them about the lost opportunity.

Those options are bad and worse, hence changing the license is a much more attractive proposition going forward.
April 29, 2010
Nick Sabalausky wrote:
> "Walter Bright" <newshound1@digitalmars.com> wrote in message news:hrcbrr$2t7e$1@digitalmars.com...
>> Moritz Warning wrote:
>>
>>> Maybe you can talk to the Tango devs to clear up this matter?
>> I suggest that the Tango devs convert the Tango modules that can get full agreement by their respective devs be converted to the Boost license. The Boost license is free of the legal problems that BSD has, and is compatible with the Phobos license.
> 
> It looks like the Tango devs are pretty much settled on BSD-only with some hack to get around the binary attribution thing: http://www.dsource.org/projects/tango/ticket/1701  (*Shrug*, well, at least it's not as insanely verbose and impenetrable as Apache 2.0...)
> 
> I *hate* licenses...(That's why I use the zlib one, none of the public domain problems, all of the freedoms that I've been told Boost offers, and none of Boost's idiotic over-verbosity.)

Yeah, we all feel the same way.
But I don't think the boost license is verbose. It's 4% of the length of the GPL:

zlib:     957 characters
boost:    1361 (1/3 of which comes from US legal requirements).
Apache2:                 9219
Academic free license3: 10332
GPL 3:                  32069
April 29, 2010
"Don" <nospam@nospam.com> wrote in message news:hrclc9$gjg$1@digitalmars.com...
> Nick Sabalausky wrote:
>> "Walter Bright" <newshound1@digitalmars.com> wrote in message news:hrcbrr$2t7e$1@digitalmars.com...
>>> Moritz Warning wrote:
>>>
>>>> Maybe you can talk to the Tango devs to clear up this matter?
>>> I suggest that the Tango devs convert the Tango modules that can get full agreement by their respective devs be converted to the Boost license. The Boost license is free of the legal problems that BSD has, and is compatible with the Phobos license.
>>
>> It looks like the Tango devs are pretty much settled on BSD-only with some hack to get around the binary attribution thing: http://www.dsource.org/projects/tango/ticket/1701  (*Shrug*, well, at least it's not as insanely verbose and impenetrable as Apache 2.0...)
>>
>> I *hate* licenses...(That's why I use the zlib one, none of the public domain problems, all of the freedoms that I've been told Boost offers, and none of Boost's idiotic over-verbosity.)
>
> Yeah, we all feel the same way.
> But I don't think the boost license is verbose. It's 4% of the length of
> the GPL:
>
> zlib:     957 characters
> boost:    1361 (1/3 of which comes from US legal requirements).
> Apache2:                 9219
> Academic free license3: 10332
> GPL 3:                  32069

Saying a license isn't verbose because it's much shorter than the GPL is like saying a particular restaurant is good just because it's better than eating out of a dumpster.

Besides, when 2/3 of...anything...is made up of sentences that are more than 60 words each (I counted), it's just plain badly written, period. (Seriously, 60+ words per sentence?! And the first one ends with a colon, so it's easy to argue it's one 120+ word sentence. Talk about run-on unreadability!) And then, naturally, the other 1/3 is all-caps.

Seriously, were they *trying* to prevent people from understanding it? If so, I don't think they could have done a better job. (At least not without hiring the FSF's "Let's do everything we can to enure our profession is needed as much as possible" lawyers.)


April 29, 2010
On 04/29/2010 01:54 PM, Walter Bright wrote:
> Nick Sabalausky wrote:
>> It looks like the Tango devs are pretty much settled on BSD-only with
>> some hack to get around the binary attribution thing:
>> http://www.dsource.org/projects/tango/ticket/1701
>
> The problem with the hack is everyone trying to use the library
> professionally will (if one is lucky) ask about the binary attribution
> thing, and one would have to convince them that the hack takes care of
> it. This puts a never ending burden on the team.
>
> If one is unlucky, they'll just avoid the library because of that
> license, and you'll never hear from them about the lost opportunity.
>
> Those options are bad and worse, hence changing the license is a much
> more attractive proposition going forward.

OTish: What would be funny is if the tango team didn't take the pains to ensure the compiler didn't strip the string out during optimization. Or compress it. Or obfusticate it.