Draft of "Creating New Projects" Page

[David Recordon]

Last week Danese copied over the Apache Incubation Policy and I've spent the past few days reworking this into a draft for a page on the actual site about starting new projects.  There are still pieces which need to be filled out and I'd love to have help on that.  I'm really trying not to just copy and pase from Apache, rather use their policies as a starting point and tweak to fit specification work and have a different tone to be less of a "normalized policy".

The wiki is a bit funky right now around access permissions, though if you login and request access I'm really quick about approving it.  This should be fixed fairly soon as PBWiki finishes off their changes to v2.

http://open-web.pbwiki.com/Creating+New+Projects

--David

[John Kemp]

Hi David,

Thanks for this. A few quick comments:

i) The proposal says that the "mentor" is drawn from the Foundation
membership - does that mean the mentor could be any member, or in
practice should be a member of the Incubator PMC (for example)?
ii) As I understand the process, once a proposal is ready, a vote takes
place. Let's say the vote passes (according to whatever voting rules are
made) - first - who actually votes on the proposal (the whole OWF
membership - or those subscribed to the incubator list for example)?
Secondly, when the Incubator PMC receives the vote results, does that
mean they "automatically" recommend the project to the Board for moving
beyond incubation? Are there any other, potentially more limited (ie.
board level or incubator PMC) votes?
iii) Who is responsible for establishing "rough consensus" [1] on a
proposal - the mentor, the person proposing, or someone else - can it
vary per project? I would actually suggest a direct edit to the line
where it says "some sort of consensus" to directly say "rough consensus"
to link it to the historical sense of that phrase (again, see [1]).
iv) A related comment is just to ask whether the line "What are the
roles of people involved in a project?" is intended to get someone to
fill in that section? I guess I would say as a start something like

"The proposal should have an editor who is responsible for establishing
rough consensus around the text of the proposal, based on his or her own
edits and those of the other contributors. The editor need not be the
person who initiated the proposal."

As some additional related reading material, the IETF working group
process document, RFC2418 [2] may be helpful.

Regards,

- johnk

[1] <http://en.wikipedia.org/wiki/Rough_consensus>
[2] <http://tools.ietf.org/html/rfc2418>

[David Recordon]

Thanks for the comments, thoughts inline...

On Wed, Aug 13, 2008 at 1:24 PM, John Kemp <jo...@jkemp.net> wrote:

Hi David,

Thanks for this. A few quick comments:

i) The proposal says that the "mentor" is drawn from the Foundation
membership - does that mean the mentor could be any member, or in
practice should be a member of the Incubator PMC (for example)?

I'd think any member of the Foundation...at least early on it won't really make a difference.  Any member should be familiar with how OWF works since they're only a member based on their prior contributions and commitment to what OWF does and how it works.
 

ii) As I understand the process, once a proposal is ready, a vote takes
place. Let's say the vote passes (according to whatever voting rules are
made) - first - who actually votes on the proposal (the whole OWF
membership - or those subscribed to the incubator list for example)?
Secondly, when the Incubator PMC receives the vote results, does that
mean they "automatically" recommend the project to the Board for moving
beyond incubation? Are there any other, potentially more limited (ie.
board level or incubator PMC) votes?

Yeah, part of the doc I didn't fully write is what/where votes happen.  It looks like Apache has changed the way acceptance of new projects works since the last time I went through it, though it feels like this change was made due to how many projects Apache already has.

I'd imagine that at the start members of the Incubator PMC will be roughly the same as members of the Foundation.  I'd imagine that technically votes of Incubator PMC would be the binding votes.  I don't feel strongly here though beyond that it should be as simple as possible.

I think one difference will be that OWF projects will virtually always become top level projects as they'll largely be individual specifications.  Extensions or really related specifications would be the exception to this.
 

iii) Who is responsible for establishing "rough consensus" [1] on a
proposal - the mentor, the person proposing, or someone else - can it
vary per project?  I would actually suggest a direct edit to the line
where it says "some sort of consensus" to directly say "rough consensus"
to link it to the historical sense of that phrase (again, see [1]).

This bit I copied directly from Apache though have no issue with changing the wording.  I think it means that the proposers along with the mentor will determine when the discussion has died down and their is rough consensus on how the proposal should change from how it was initially sent to the vote thread.
 

iv) A related comment is just to ask whether the line "What are the
roles of people involved in a project?" is intended to get someone to
fill in that section? I guess I would say as a start something like

"The proposal should have an editor who is responsible for establishing
rough consensus around the text of the proposal, based on his or her own
edits and those of the other contributors. The editor need not be the
person who initiated the proposal."

Yep, meant to be filled in.  I'd imagine the main roles are Editors, Contributors, Mentors, and I might even call out Implementors not because OWF is about code but because feedback from working implementations is so important to successful specs.
 

[Danese Cooper]

David,

I'm back from vacation now and happy to help.  Sorry I couldn't get all the way through it before I left, but you've made a great start here and it looks good.  In your comments back to John you mentioned that you think the Incubtor PMC will be roughly the same as the Founding Members.  I think the Incubator PMC should be very small initially, on the order of 5-7 people, so would beg to differ on your assumption.  It takes a lot of work to double-check all the crossed "T"s and dotted "i"s of a proposal and that work is best done by a focused and tight group.  If more want to help, let them help first and get added to voting afterwards.  Just my $.02 after watching Apache...

Danese

[Gabe Wachob]

This is a great distillation.

Here's a few questions/comments to start with.

I think for this process, I'd like to see at least some statements about timeframe. How long do discussions on a proposal last? Is there a minimum? (say, 2 weeks) Is there a maximum (say 2 months)? 

Also, I think the language about 1 2 or 3 people isn't neccessary - is there something wrong with a group of 7 that wanted to start a group? I think its good to suggest that one person is almost certainly not enough, but I'd hate to suggest that groups of more than 3 are inappropriate to propose new work... 

On Aug 13, 2008, at 12:34 PM, David Recordon wrote:

[David Recordon]

I don't think we should write a minimum or maximum into this document
as I'd imagine it will range depending on the state of the initial
proposal and what is being proposed. Rather it should be more around
judging when the discussion is no longer new and thus has died down.
I would mind updating the document with some general time frames once
we see it happen a few times.

There should be time lines called out for votes but I think that
applies beyond just the incubator. We'll also need a page like
http://apache.org/foundation/voting.html though hopefully a little
simpler. Apache recommends at least 72 hours for every vote thread.

I'm fine if you remove the size bit, agree that it is more important
to not be one person than if you are seven.

--David

[Ben Laurie]

2008/8/15 David Recordon <reco...@gmail.com>:

>
> I don't think we should write a minimum or maximum into this document
> as I'd imagine it will range depending on the state of the initial
> proposal and what is being proposed. Rather it should be more around
> judging when the discussion is no longer new and thus has died down.
> I would mind updating the document with some general time frames once
> we see it happen a few times.
>
> There should be time lines called out for votes but I think that
> applies beyond just the incubator. We'll also need a page like
> http://apache.org/foundation/voting.html though hopefully a little
> simpler. Apache recommends at least 72 hours for every vote thread.

I'm not sure there's much on that page you'd want to get rid of, though it might be possible to say it more briefly. And perhaps with the fluff at the end instead of the beginning :-)