[tahoe-dev] questions about development priorities for Tahoe-LAFS, summer 2009

[Zooko O'Whielacronx]

Folks:

Here are just a few scattered notes about how I'm not sure who is  
going to move Tahoe forward in which ways next.  Your opinions about  
what should be done, and your offers to do them, would be gratefully  
accepted.

First of all, about me.  I'm interviewing for new jobs (if you were  
considering offering to hire me, now is the time!).  As part of this  
process I'm asking potential employers if they will give me a degree  
of moral support to work on Tahoe in my "spare time".

I'm also encouraging Peter Secor in his efforts to find renewed  
funding for Tahoe development, either on a commercial basis from  
someone who intends to make money with this technology, or on a  
philanthropic basis from someone who thinks that this technology  
could improve people's lives.

It could turn out that within a week or two from writing this letter  
I will take on some exciting new job which is not directly related to  
Tahoe and my contributions to Tahoe will lessen dramatically.

Now about Brian: it looks like he is still planning to do lots of  
hacking on Tahoe this summer, and he's an awesome hacker, so you can  
expect top-quality work from him.

Kevin Reid is not going to work on a Tahoe GUI this summer after all  
-- the Google Summer of Code people asked him to work on a remote  
object protocol for caja instead.  (Which is cool, but I'm still  
disappointed.)

The student-and-mentor team of Vincent Chow and Nathan Wilcox appear  
to be determined to marry a decentralized revision control tool such  
as Bazaar or Mercurial with Tahoe this summer.  It seems to me that  
it would be easy to make a proof-of-concept for that but hard to make  
a solid, high performance, usable tool.  I can't wait to see what  
they do!

Shawn Willden is presumably still in his secret underground lair in  
Utah hacking on GridBackup.  He hasn't been very active on this list,  
but he has published his git repo...   Hm, looks like he hasn't  
updated that git repo since he first committed to it in February.  By  
the way, here is a podcast of his presentation about GridBackup to  
the Utah Python group: http://podcast.utos.org/index.php?cat=gridbackup

Mahmoud More wants to work on decentralizing the introducer.  If one  
confines oneself to implementing ticket #68, comment #11, *and  
nothing else* then this is a small, confined project that a student  
like Mahmoud could accomplish in one summer.  However, Mahmoud would  
need a mentor to guide him, and because of my job situation I might  
not have the time to commit to that.

François Deppierraz is eager to fix #534 (encoding issues).  He has  
submitted a patch, which I didn't apply because it is a large patch  
and it takes me a while to get my head around it and to check if all  
of its changes are covered by unit tests.  I started splitting it up  
into smaller patches, but, again, I don't really have time to do more  
on that right now, so if François or anyone else could look at those  
patches and make sure that everything that we care about is tested  
with a unit test then that would be forward progress!

I might be forgetting other people who are currently working on Tahoe  
improvements.  If so, I apologize and please speak up.


Now, per my earlier cryptic notes about Tahoe development strategy  
[1], my next priorities for Tahoe are:

  * short-length feedback loop with actual users
  * inclusion in Debian & Fedora
    - pycryptopp linkage
  * low hanging fruit of performance
  * encoding
  * volunteergrid

If anybody wants to help out with things along these lines, speak  
up!  Let's make a goal to have a new stable release of Tahoe,  
containing *some* kind of improvements that people would care about,  
having no known regressions, and having thorough unit tests for all  
changed code, in June 2009 -- about two months since the v1.4 release.

Regards,

Zooko

[1] http://allmydata.org/pipermail/tahoe-dev/2009-March/001506.html

tickets mentioned in this message:

http://allmydata.org/trac/tahoe/ticket/68#comment:11 # distributed  
introduction COMMENT ELEVEN
http://allmydata.org/trac/tahoe/ticket/534 # "tahoe cp" command  
encoding issue