OSI Prez confronts irate users over 'badgerware' license
Quick action needed after nine months
Open Source Initiative (OSI) President Michael Tiemann has responded to queries about the organization's decision to "rush through" approval for a new badgerware license by saying that the OSI board did its best for "the community."
Last week, the OSI slapped the official open source tag on Webtwopointzero-y start-up Socialtext's Common Public Attribution License (CPAL). The move stands as significant, since it provides OSI's strongest backing to date of so-called attribution-style or badgeware licenses. It's also of note because the OSI approved last-minute changes to the license during an early morning board meeting without running the alterations by "the community," as is standard practice.
A handful of OSI watchers complained about the rapid approval. Developer Andrew Oliver, for example, removed himself from an OSI mailing list in protest. And Matthew Flaschen chimed in to the message board asking: "Since when has OSI approved a version of a license that license-discuss hasn't seen? I don't like this part at all."
The developers begged Tiemann to give his side of the story, and today Tiemann dished the dirt on the mailing list.
In Monday's board meeting the topic of whether to approve the CPAL was considered. We took the advice of the License Approval Chair, which was to approve the CPAL.
We know that there was not 100 per cent consensus among those who subscribe to license-discuss, but we felt that there was reasonable consensus that the safe harbor of the license was sufficiently broad to meet the requirements of the OSD (open source definition) and therefore approve the license.
Whereas there was significant and broad consensus *against* the predecessor version, the response by many to the current version was a stark contrast that we read as general acceptance. In the course of taking that recommendation, which I agreed with, I observed what I considered to be a mismatch between the text of the license and the intentions discussed, debated, and before us to approve.
I suggested that it would be better to correct the error in the license as long as such correction did not in any way affect the substance of the discussion of that license. If the error had in any way violated the OSD, I would have sent the license back to the review process, as I did when the SimPL license was (erroneously) submitted with incorrect text. But this error related to a potential for confusion where none need exist, in a license whose terms were otherwise recommended for approval.
I thought it better to make a change and approve a license with less potential for confusion than to approve the license and deal later with the confusion, or force yet another delay on an already long-delayed process simply to recount votes that had already been cast. The board discussed this and agreed.
It's crucial to note that the OSI spent about nine months debating either variants of CPAL or the license itself in a public forum. As Tiemann points out, many open source watchers first portrayed a badgerware license as a horrible thing to thrust upon "the community", only to later welcome the license with
CPAL should drive some measure of consistency among the badgeware license crowd. Companies such as SugarCRM and Centric CRM - and many others - have crafted various versions of the Mozilla Public License (MPL) that include so-called attribution clauses unique to their wares. As a result, scores of attribution - or badgeware - licenses have been thrust at customers - none of them OSI approved.
The attribution clause is one workaround for software makers looking to receive credit when companies pick up their code. It requires customers to display a logo and/or text noting who contributed to writing the code in question. In addition, CPAL has a "network use" clause that forces service providers and others to make clear their use of open source code, even though they are not redistributing software in the traditional sense often covered by open source licenses. So, even if you run software on a server and only deliver a service to end users, you must credit the original creator of the code, under CPAL.
It's expected that companies such as SugarCRM will modify their old attribution licenses to fit CPAL. We've taken the liberty of dubbing CPAL a badgerware license in honor of Socialtext's fluffy nature.
We can understand the OSI watchers' concern with the license's last minute approval. The changes to the original text, however, were beyond minor - just a few words. In addition, the board had followed and participated in the CPAL debate for months. Surely, "the community" must put same faith in the board to do the right thing at crunch time. ®