Roy T. Fielding wrote:
On Dec 12, 2007, at 11:34 AM, Stefano Mazzocchi wrote:
I feel the need to outline, though, that the "no release" rule for labs
was created to avoid people from using labs as a way to route around the
incubator, yet 'code-less' labs are able, in fact, to 'release' their
own documentation artifacts, without oversight or the need for a
community around the effort.
Let me repeat: I think that Webarch and code-less labs are a *good
thing* and should be encouraged in any way possible, as long as the
product of their effort is clearly marked as coming from a lab (which is
the case for Webarch) and not used as a way to abuse the apache branding
to give resonance to an individual committer's effort.
That's a good point. And I think it's sometimes lost.
The labs should not be an obstacle for /a participant/ to publish /their/
work. Not as an "ASF release", obviously, but as their own work.
Not necessarily through the ASF at all, although people.apache.org/~me/
should be is clearly enough that person's work, perhaps with additional
disclaimers as you hint. This should be refined and tweaked.
Certainly, publishing externally shouldn't be a problem. CLA grants the
ASF a license. It doesn't strip the author of their copyright.
In fact, when we talk about licensing, if Sam wants to go and license
their work under the GPL as well as the AL, there's really nothing that
prevents that (nor should there be). How a labs project starts doesn't
have to reflect its final disposition by the author?
I'm just mentioning this so that both labs@ and board@ consider this
food for thought and in no way I want such food for thought endanger the
ability to use labs in the way Roy has done... but it's my duty, as PMC
chair, to report changes in the project evolution that might require, in
the future, board attention that this is the closest thing to that (even
if, right now, does not require any action from any part and it's
unlikely that it will require any in the future either).
However, I should also note that the IETF httpbis WG (or at least the
chairs in consultation with the ADs) have decided that apache.org
is not a sufficiently "neutral" player to host such work in the future,
so my partitioned drafts, issues, and related stuff are presently
being moved to tools.ietf.org. The webarch http lab has served its
purpose, but is about to be closed (essentially graduating to a spec
project on some SDO's infrastructure).
This was very interesting, can you clarify what aspects of the ASF were
determined to be not-"neutral"? The fact that the ASF implements specs?
I could understand that perspective. Was it the patent clause of the
license? Or was it something more that the board should be concerned about?
As for the rest of webarch, I've pretty much concluded that it is
pointless to host a project on infrastructure that doesn't allow
the creation of project-specific mailing lists. Labs is a nice idea
but is overly constrained. For example, consider the website issue.
Is it really okay for me to just publish stuff like
http://labs.apache.org/webarch/http/draft-fielding-http/
without it being considered a "release"? I honestly don't know.
I mean, I know it is safe for me to do so given the nature of
the content, but that may not be true for other labs, or even for
webarch if I put the waka spec up on the site (the original plan).
One nice caviat there, you named it 'draft-'. Consider if it became
a final document, are you prevented from publishing it there, but
allowed to have published it throughout the draft process? Interesting
tangle.
And what happens to that content when the lab is closed? The two
main reasons for having it at Apache are so that others can
work on it (e.g., Julian) and so that the version history is
permanently available (for legal and historical reasons). I think
that may have stretched the labs idea too far.
Collaboration with non-committers is a huge hassle, I'd agree. And
when it outgrows one or two authors into many, labs probably isn't
the place for it.
In order for me to do webarch at Apache, I need a TLP with its own
mailing lists, a stable location for subversion/jira/website URIs,
and the ability to release "products" with licenses that are more
lenient than the Apache License. I don't see a way of getting
there from here. Creating my own SDO would be easier.
I find this especially interesting, given your involvement in very
carefully crafting the AL 2.0. What you author, of course you may
relicense differently externally to the ASF, and the labs concept
is really ideal for this (limited authors, easier licensing).
Ignoring the infrastructure issues; what can the ASF provide in terms
of a more lenient license, without exposing end users to more
significant risks?
Feel free to fork these diverse thoughts into board or labs threads
as you will.
Bill
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
