--
--dmg

---
Daniel M. German
http://turingmachine.org

I believe we already discussed this to some degree and decided that we would not enter the arena of word equivalents with the exception of spelling variations for known American-British English.  Although there are certainly plenty of words that seem “safe” to equate, that begins to feel like a slippery slope in terms of the potential for different meanings or the license author’s intent getting altered.  More specifically, many licenses have definitions for such words.  Past and present tense can also be tricky, both from deciding how and when it’s okay to use either tense.  All in all, I think it’s best to take a very conservative approach to what we will “replace” in terms of templatizing the licenses.  We can always expand, but it would be very hard to roll it back later on.  Just my two cents.

Certainly the BSD and Apache 1.1 licenses are problematic, because they are often verbatim with the exception of the author’s name in the third clause (BSD) and the third, fourth, and fifth clause (Apache 1.1) as well as the disclaimer sections for both.  We are working through how/where to capture this, as well as copyright notices in general, though.



Jilayne


On 3/23/11 10:11 AM, "Kate Stewart" <kate.stewart@...> wrote:

Thanks Daniel,

    Will look into adding this after we can get the guidance from the lawyers as to what varients are equivalent.  ;)   

Kate


--- On Wed, 3/23/11, dmg <dmg@...> wrote:

From: dmg <dmg@...>
Subject: Re: Today's Agenda Legal WorkStream
To: "Esteban Rockett" <mgia3940@...>
Cc: spdx-legal@..., spdx@...
Date: Wednesday, March 23, 2011, 9:48 AM



On Wed, Mar 23, 2011 at 7:40 AM, Esteban Rockett <mgia3940@... </mc/compose?to=mgia3940@...> > wrote:

Agenda Legal WorkStream

(4) Review and Discuss --- Guidelines for "Templatizing" licenses

Please goto:  http://spdx.org/wiki/guidelines-templatizing-licenses

I think it is going in the right direction. You should look at the way Ninka normalizes test  for in-file licenses (BSD, MIT, etc).
For the BSD and MIT variants the problems are not only spelling, but in some cases, variants in grammar: past vs present
tense, use of word "software", "product", "program", "library" in the same place.

---

_______________________________________________
Spdx mailing list
Spdx@...
https://fossbazaar.org/mailman/listinfo/spdx

Jilayne Lovejoy |  Corporate Counsel
jlovejoy@...

720 240 4545  |  phone
720 240 4556  |  fax
1 888 OpenLogic  |  toll free
www.openlogic.com

OpenLogic, Inc.
10910 W 120th Ave, Suite 450
Broomfield, Colorado 80021

--
--dmg

---
Daniel M. German
http://turingmachine.org

These variant licenses would simply end up needing to be added as a “nonstandard” license, meaning the SPDX generator would not be able to use the standardized SPDX license list shortname for that license, but it wouldn’t (shouldn’t) be tagged as unknown in such a case.


On 3/23/11 10:54 AM, "dmg" <dmg@...> wrote:

On Wed, Mar 23, 2011 at 9:48 AM, Jilayne Lovejoy <Jlovejoy@...> wrote:

I believe we already discussed this to some degree and decided that we would not enter the arena of word equivalents with the exception of spelling variations for known American-British English.  Although there are certainly plenty of words that seem “safe” to equate, that begins to feel like a slippery slope in terms of the potential for different meanings or the license author’s intent getting altered.  More specifically, many licenses have definitions for such words.  Past and present tense can also be tricky, both from deciding how and when it’s okay to use either tense.  All in all, I think it’s best to take a very conservative approach to what we will “replace” in terms of templatizing the licenses.  We can always expand, but it would be very hard to roll it back later on.  Just my two cents.

I agree. I wonder if a solution is to allow the specification of "variant" of a license. Perhaps a way to say: the closest license is this one, with a text similarity metric of X, and where the differences are  such and such. Otherwise a lot of BSD and MIT licensed files will be listed as unknown/other.


--dmg
 

Certainly the BSD and Apache 1.1 licenses are problematic, because they are often verbatim with the exception of the author’s name in the third clause (BSD) and the third, fourth, and fifth clause (Apache 1.1) as well as the disclaimer sections for both.  We are working through how/where to capture this, as well as copyright notices in general, though.

Jilayne Lovejoy |  Corporate Counsel
jlovejoy@...

720 240 4545  |  phone
720 240 4556  |  fax
1 888 OpenLogic  |  toll free
www.openlogic.com

OpenLogic, Inc.
10910 W 120th Ave, Suite 450
Broomfield, Colorado 80021

Philip Odence twisted the bytes to say:


Philip> Yes, and let's not forget this same point when we are talking about the process of adding new license to the standard list. The discussion is never
Philip> whether the license can be included in an SPDX file—it always can be—the only issue from the SPDX file creator's perspective is whether it matches a
Philip> license on the standard list with a predefined short name, or whether they have to go the one extra step of including the license text in Section 4.
Philip> Licensing Info and create their own local short name.

One interesting aspect of licenses that are "almost" a match is that, if
they are only listed as "unknown", the knowledge/analysis performed by
the SPDX-file's author might be lost (why is such license unknown).

Perhaps a "comment" field attached to any attribute (file, package, etc)
that can have a license would be very useful to include a rational.

--dmg


--
--
Daniel M. German
http://turingmachine.org/
http://silvernegative.com/
dmg (at) uvic (dot) ca
replace (at) with @ and (dot) with .