Hi Jilayne, I can see why SPDX would want to try to have some consistent approach to forming identifiers for related licenses (although I think the value of having a standard identifier is more important than consistency across multiple identifiers).
My view is that an SPDX short identifier has a specific primary purpose which comes out of the SPDX specification, and a definition which among other things is dependent on the SPDX matching guidelines. It is not meant to supplant all other names of a license for all purposes, which if possible should reflect prevailing usage of the community (or author) using a license. Of course as SPDX gets more successful it may start to influence how licenses are named in common discourse, and I already see this happening to some extent. I see no reason why SPDX identifiers can't coexist with sometimes-different "common usage names" of licenses, and I myself have already gotten comfortable with using SPDX short identifiers in some situations and "common usage names" in others. I am uncomfortable however in getting ahead of the community and actively trying to stamp out the traditional or common usage names of licenses. Of course none of that is relevant to BSD+Patent which is a new license that hasn't yet been used by any projects. I think "BSD-2-Clause-Patent" is OK for an SPDX short identifier if that's what this group wants to use for that purpose, and whatever this group decides to use will be noted on the OSI web page for BSD+Patent. But, subject possibly to some preference by McCoy as license author, I am disinclined to change the name "BSD+Patent" itself. It is one of the best names for a new open source license I've seen in a really long time and has a catchiness that may help with adoption of the license, which I consider a really valuable addition to the set of OSI-approved licenses. To instead call it, for *non-SPDX* purposes, "BSD-2-Clause-Patent", or even "2-Clause BSD plus Patent License", would be (I feel) to substitute for a name that has charm, beauty, vitality and magic something cold, soulless, gray and bureaucratic (qualities that are absolutely fine within the sphere of activity that SPDX is concerned with). (If SPDX calls BSD+Patent "BSD-2-Clause-Patent", wouldn't consistency suggest changing the identifier for Clear BSD to "BSD-3-Clause-No-Patent"?) Richard On Fri, Jun 02, 2017 at 05:08:31PM -0600, J Lovejoy wrote: > Hi All, > > I probably should have explained this to begin with and Richard has now > provided some key background, which I’ll add to here: > > OSI has adopted (and did endorse via a joint public announcement that was > probably back in 2011) the SPDX identifiers. This is implemented on the OSI > list via the URL and indication in parenthesis next to the full name. This > was all very easy in the beginning, as the OSI list did not change for some > time, so once we were all “caught up” from both sides - that is, SPDX ensure > all OSI-approved licenses were on the SPDX License List; and OSI adopted the > short identifiers from SPDX. > > More recently, as OSI started to approve licenses again, we realized we > needed some better cross-communication such that: SPDX would know when OSI > added a new license; SPDX could then create a short identifier; OSI could use > the short identifier in the same way as described above: and SPDX could > ensure that new OSI-approved was added to the SPDX License List. We > discussed this process about a year or so ago and it was also decided that > where OSI got a new license up on their website prior to SPDX adding it to > the SPDX License List / assign a short identifier, then OSI could add that > reference (parenthetical) to their webpage and change the OSI URL as needed > to all be consistent. > > So, that is where we are now. SPDX License List needs a full name and a > short identifier to be consistent with the protocols that we have used so > far; the short identifier being more critical. We have also tried to conform > with the wishes of the license author (or license author representative, as > in this case) as well. > > The addition of this license does not raise any compelling reason to change > existing short identifiers of licenses that have long been on the SPDX > License List here. This license is like others already on the list that use > some aspect of one of the BSD licenses and builds upon that. We have a > pattern of short identifiers for such a situation, so we ought to carry on > that pattern as that is what people will expect and being consistent is key! > Along the same lines, considering that “+” and “with” (as well as “or” and > “and”) are license expression operators, we aim to avoid using these in a > license short identifier as that would obviously cause confusion (and in full > names to some extent). > > Hence: the suggestion was/is: > > Short identifier: BSD-2-Clause-Patent > (patent can be capitalized or not, but we have capitalized > other such words in BSD* short identifiers, so I’d vote for being consistent > there too. I don’t really see any reason for something different here and it > seems like we have some amount of consensus here, so can we call this done > and dusted?) > > Full name: BSD 2-Clause Patent > (this would be consistent with other full names on the SPDX > License List. Note, the full name is not as critical as this is generally not > used in the same was as short identifiers, however consistency here is also > nice for viewing on the list, etc. I see that OSI has listed the full name > as “BSD+Patent” - I’d urge McCoy and OSI to reconsider this as, even though > it’s not a short identifier, it could cause confusion due to the “+” sign > being used. > Note: OSI now lists the full name for BSD-3-Clause as: "3-Clause BSD > License", whereas SPDX has the full name as: "BSD 3-clause "New" or "Revised” > License” > SPDX choose that name as following the lead of OSI and how they listed the > full name on their list of licenses circa 2011 when OSI adopted the SPDX > identifiers: > https://web.archive.org/web/20110605024926/http://www.opensource.org/licenses/alphabetical > > <https://web.archive.org/web/20110605024926/http://www.opensource.org/licenses/alphabetical> > > I don’t know why OSI has since changed the full names, but again, I don’t > think the full name is as critical (especially in order of the words) so long > as it is sensibly recognizable. If OSI was to follow it’s own current naming > convention for BSD licenses, it might use "2-Clause BSD plus Patent License” > or something along those lines. > > So, can we: > 1) align on the short identifier; and > 2) come up with a full name that is either the same or reasonably similar on > both lists, and consistent with the pattern as evidenced on both lists (and > make McCoy happy)? > > Remember the goal here is easy, consistent, and reliable identification of > licenses. > > Thanks, > Jilayne > > SPDX Legal Team co-lead > opensou...@jilayne.com > > > > On Jun 2, 2017, at 3:51 PM, Richard Fontana <font...@opensource.org> wrote: > > > > On Fri, Jun 02, 2017 at 05:16:01PM -0400, Wheeler, David A wrote: > >> J Lovejoy: > >>> Specifically, when adding other BSD-x-Clause licenses, we have tried to > >>> follow the same pattern for the identifiers as it aids in identifying > >>> what exactly the license is, which I think everyone finds helpful! Hence > >>> the use of BSD-x-Clause-<extra> was intentional and thus, why I suggested > >>> such a pattern here. > > > > I think it is possible that this effort to systematize the identifiers > > for textually-related licenses, if taken too far, can actually cause > > greater confusion or have other questionable unintended > > consequences. One case where I worry this is happening is the > > so-called Clear BSD license, which SPDX calls "BSD-3-Clause-Clear". > > > >> I agree, I think something like "BSD-2-Clause-Patent" would be the better > >> choice: > >> * It's more consistent with the other licenses > >> * All the existing tools can handle that, even if they can only handle > >> identifiers (not expressions) > >> * Using "-with-" is hard to distinguish from the operator "... WITH ..." > >> when spoken > >> * Using "...plus..." is hard to distinguish from terminating "+" when > >> spoken > >> > >> I hope that SPDX & OSI will agree on a short name, too...! > > > > Well, OSI doesn't have a notion of an official, OSI-endorsed short > > name. OSI has been prominently noting the SPDX short identifiers on > > the opensource.org web pages for the licenses and that commitment > > should continue. I don't think the OSI should do more than that, > > though. > > > > Richard > > > _______________________________________________ Spdx-legal mailing list Spdx-legal@lists.spdx.org https://lists.spdx.org/mailman/listinfo/spdx-legal
