Hullo, folks - a number of us have been talking about doing a FAD on
Etherpad (web-based gobby, essentially - I think most people here have
seen it), and it's time to move those planning conversations to a public
mailing list. I thought this one would be the closest to what we needed
for the event/execution planning, since it crosses through multiple
teams. We'll try to move upstream with our development conversations; if
this isn't an appropriate place, please let me know where else I should
be taking it.
The event page is at https://fedoraproject.org/wiki/Etherpad_FAD, and
it's awfully stubby right now, but will be filled in pretty soon.
I should also explain that a bunch of folks involved in this are doing
so as their first major Fedora contribution. All of us have ties to Olin
College, which is the proposed location (Needham, MA, USA). Most of you
already know myself (Olin '07, mchua - "zomg let's see if we can make
upstream a stronger community"/QA person) and Sebastian Dziallas (Olin
'14, sdziallas - packager who's done basically all of
https://fedoraproject.org/wiki/Etherpad so far), but joining us in
* Colin Zwiebel (Olin '12, czwiebel - js hacker and event coordination
* DJ Gallagher (Olin '07, <getting-used-to-IRC> - js hacker and
* Andy Pethan (Olin '11, andyp - interface/user guru).
We plan on drawing in local and not-so-local Fedora (and hopefully
Etherpad, if we can get someone from the original team in!) hackers for
the event (October-something-or-other, likely) and part of our goal is
to get Etherpad deployed in Fedora, at least at a
"ready-to-go-into-staging" level, and in a better state as an upstream
than they are currently (it's nearly impossible to decipher the code,
there is no ticket tracker, the discussion is scattered, there's no
release cycle, etc...) - the functionality/interface of the software is
so strong and such a *great* collaboration tool, particularly for those
not of a development background, and helps new contributors start so
much, that we think it's worth going for. I mean, "First" *is* a
In addition to the actual hacking of Etherpad, we're using this event to
teach Oliners the Fedora/FOSS way of doing things, and how to get
involved in Fedora, with a concrete project folks can dive in on with
their existing skill sets (i.e. "everyone is coming in at a high level
of some skill we need"). Copying Ryan Rix so he's in the loop, since
this seems like something that could potentially fall under the Campus
Ambassadors umbrella (as it's at a college and run by students as part
of FOSS outreach within their community).
So, to resume the conversation, here's a bit of context on where we are
Mainly, I'm concerned that EtherPad may fork a bunch of different
directions or will continue to be a neat, difficult to use project that
will flounder. I mean, it doesn't seem easy to break into the EtherPad
foundation. Communities need to be accessible, its the only way you get
people to contribute patches and ideas. But I think I'm just being
In terms of technical hurdles to community formation, and at the risk of
continuing to sound like a Java fanboy recklessly derailing a Linux con,
I feel like there's a missed opportunity in this not being somewhat
friendlier to the at-large community of young hip developers who know
how to work elements of the common toolchain of Java and Scala. More
specifically: JVM applications aren't meant to be packaged in ways that
negate the cross-platform goodness of the JVM. If the mysqldb setup is
hateful, then developer builds of the app need to be able to forgo that
setup and point to a shared dev database instance. If the app needs to
be packaged with its libraries arranged just so, you write Ant scripts,
which express such tasks in an elegant and compatible way.
There's still a lot of things I don't know or understand, like how
EtherPad integrates a webserver, does it use anything resembling the
Java servlet spec, or am I just thinking in completely the wrong frame
of mind imagining that it would be treated like a Java project that
you'd want to build with Ant or Maven and run on something like Apache
I can only say standardization tends to be win, because code in JRuby or
Groovy or Scala that uses the servlet API can be ported to many
different server front ends, or for that matter the cloud via Google App
Engine. Probably the same thoughts Google had on the matter.
I'm not sure where I was going with this, but I thought I should
volunteer that information. I can't code my way out of a bag in Scala.
Are we hurting for someone who can?
I've read a bit more of the source code now, and I stand corrected on
some of these points. While EtherPad requires the servlet API as a
dependency, it isn't a java webapp, so much as it is a Comet protocol
provider, with an embedded Jetty server as its Comet stack. I may be of
some assistance in explicating the server and servlet components, but I
would not claim to know how the code ought to be packaged or deployed or
whether its dependencies are going to pose problems.
I've been pretty much buried in FPL transition matters this past week,
and I don't think I'll have time to prepare anything for an Insight
meeting this week.
Among other critical path tasks, I need to talk to a few other teams
(notably Design, also Infrastructure) about possible collaboration
points. Design was looking in particular at mail-list front ends, for
I'd like to postpone our next IRC meeting until July 15th, at the
normal time (2:00pm US-Eastern).
Paul W. Frields http://paul.frields.org/
gpg fingerprint: 3DA6 A0AC 6D58 FEC4 0233 5906 ACDB C937 BD11 3717
http://redhat.com/ - - - - http://pfrields.fedorapeople.org/
Where open source multiplies: http://opensource.com