couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Miles Fidelman <mfidel...@meetinghouse.net>
Subject Re: Couch Vision
Date Sun, 04 Mar 2012 22:37:20 GMT
Noah,

Noah Slater wrote:
> Miles,
>
> Thank you for taking the time to write. Apologies for the delayed response.
Thanks for the reply and the details.  I was afraid my comments were 
falling on deaf ears, or that I'd offended folks.

Glad to hear that Couch is still alive and well - I'm about ready to 
commit to using it on a major project, and recent developments have 
given me some pause.  (Also, I just think that Couch is plain cool.)

> It is unfortunate that Damien decided to leave the project, but the ASF is
> set up, specifically, so that projects do not stagnate away when one person
> leaves. Damien was "just another comitter" in the eyes of the Apache way,
> in any case, he had not even been active in the year prior to his
> departure. We're very fortunate actually. CouchDB has a thriving community.
> So I know we'll be fine.

Well sometimes.... I've seen different levels of momentum for different 
ASF projects.  It still seems to be very dependent on having a committed 
core team.  Also, ASF seems to have an awful lot of Java-based projects, 
where Couch is erlang-based.

> We do not have a roadmap at the moment. Which is a bad thing. We should
> have a roadmap. But you suggest that it doesn't look like we're planning to
> do anything besides bug fixes and incremental feature addition. Isn't that
> how most products evolve? What sort of thing were you hoping for? One very
> big change on the (in my head) roadmap is the merging of BigCouch in to
> CouchDB.

It's always struck me that the replication mechanisms are a lot less 
robust than advertised.  Great PowerPoints, but scalability seems 
lacking, and the need to set up pair-wise replication seems 
unmanageable.  I hadn't realized that the BigCouch code was being 
integrated - that's important.  (I keep thinking that a multi-cast 
publish-subscribe update mechanism - something like a gossip protocol - 
would be a key next step.  Something like that may come out of what 
we're working on, but not for a while.)
> 1. Who is really driving CouchDB right now?  Is there an individual or core
>> group who are really providing vision and momentum (some folks are obvious,
>> but where are priorities)?  How do the players overlap with the various
>> other efforts going on.  In other words: what's the ecosystem for Couch and
>> Couch-related software, projects, ....?  (Can't tell the players without a
>> scorecard.)
>>
> With the departure of Damien, I have been talking a lot to Jan about things
> we can do to help kick things up the arse a little bit. We both feel like
> we've been neglectful, and want to use this as an opportunity to double
> down our efforts, and inject this community and this project with a bit of
> love again.
That's good to know.

> We're not the only ones prompted in to action by the recent events.
> Cloudant used Damien's departure as a prompt to announce that they were
> donating BigCouch to the project. I know that other individuals too, from
> the PMC, the committers, and the wider community, have been rallying
> together.
As is that!

> 2.  It would be particularly helpful to those of us new to the project to
>> see a NEW summary of what it is that defines CouchDB, particularly as
>> distinct from the various efforts that have spun out of Couch, and what's
>> the current vision/priority for the future.  What are the core capabilities
>> provided, core mechanisms, .... ?
>>
> That's a great suggestion, and we definitely need something like that.
>
> Watch this space.

Cool.
>> Sort of strikes me a bit like what happened when Ubuntu showed up.  It's
>> still not clear how it relates to Debian.
>
> As an ex-Debian developer, that's a bit of a sore point. ;)
>
As a Debian user, too.  Sigh...

Cheers,

Miles

-- 
In theory, there is no difference between theory and practice.
In practice, there is.   .... Yogi Berra



Mime
View raw message