Hi,
it's likely much of April will be spent on polishing and releasing
Tails 0.11.
Once this is done, we will have shipped a few major new features into
the wild. Once this is done, I hereby propose we pause implementation
of big new features for a short while, and instead focus on two areas
where, I think, we are lagging behind:
* re-scaling our _infrastructure_
* connecting back to our immediate _surroundings_
This proposal is not aimed at preventing anyone from implementing new
features. It's about outlining a general direction I propose we
collectively take.
This proposal is not about organizing work or matching people to
tasks. It's about outlining a vision.
Contrary to my usual release schedule proposals, this one contains no
time indication of any kind. We'll set those later, once and if we
agree on the general vision.
My immediate desire is to hear answers and feelings about this general
vision. I'll understand commenting and debating on details as implicit
agreement on the general vision ;)
Re-scaling our _infrastructure_
===============================
Most of our current infrastructure has slowly grown from what was
initially setup back in the early days of a project called amnesia.
Since then, well, many things happened. It's almost a miracle that our
infrastructure has scaled this well until now. But the limits of the
old design are becoming apparent in various places. Most of the needs
were thought through already, some of the future plans are crystal
clear, what is now needed is to spend serious time on it, implement
designed bits and design others for the next few years:
* APT repository -
https://tails.boum.org/todo/APT_repository/
This blocks the Git split, the move to Icedove, and possibly
todo/distribute_source.
* split Git -
https://tails.boum.org/todo/split_Git/
Our current Git is too big, and mixes stuff that hardly belongs
together. Post-0.11 will be a great time to rewrite history, as we
won't have that many unmerged development branches.
* improve the forum -
https://tails.boum.org/todo/improve_the_forum/
Moving to a better web forum will get us a less cluttered Git
history, better user experience, and hopefully even better
community dynamics.
* automated builds and tests
https://tails.boum.org/todo/automated_builds_and_tests/
This will become huge, but should start small.
* more stable WhisperBack SMTP relay
https://tails.boum.org/todo/more_stable_WhisperBack_SMTP_relay/
Bring some more hardware up and running,
refactor the Puppet recipes into a dedicated module.
* document and clarify how our infrastructure works
Some parts are very well documented, some are not at all.
It should become easier to share infrastructure maintenance work.
* miscellaneous other stuff:
https://tails.boum.org/todo/locking_down_this_website/
https://tails.boum.org/todo/better_web_interface_to_translate_this_website/
* more?
Connecting back to our immediate _surroundings_
===============================================
Many, among the software projects that are Tails upstream, are
currently preparing new stable releases. Time is getting tight to make
sure their upcoming releases fits Tails needs.
* live-build 3.x -
https://tails.boum.org/todo/live-build_3.x/
Tails relies on Debian Live. Heavily.
Tails 0.11 will ship live-boot and live-config 3.x,
but we're still using live-build 2.x.
We need to convert our configuration tree to the (deeply
incompatible) live-build 3.x format. There is no turning back.
All unmerged branches shall be converted, if possible at the
same time. Just like the Git split, I think post-0.11 is a great
time to do so... and, along the way, fix any bug in live-build 3.x
we find.
* Debian Wheezy
Tails relies on Debian. Heavily.
Debian Wheezy is supposed to be frozen in June.
We need to check what, in the current state of Debian Wheezy, is not
fit for Tails, and fix it or have it fixed.
Next step: to build Wheezy test images.
* AppArmor -
https://tails.boum.org/todo/Mandatory_Access_Control/
I've started to work on getting Debian Wheezy some AppArmor support.
If the Wheezy freeze is not postponed, June is the deadline to get
such things into Debian. Ping me if you're interested in testing
stuff.
* Tor 0.2.3.x
We need to make sure the next major Tor stable release will be great
for Tails, with a focus on the separate streams features.
Next steps: everyone of us, let's run Tor 0.2.3.x and use the
separate streams feature. Put it into Tails experimental.
* Vidalia 0.4
... is supposed to be released as stable in the next few months.
We need to make sure it will be great for Tails, with a focus on the
areas it's lacking for us (bridges support) and where we patch it
(see our patches). It would be great if we could ship Vidalia 0.4
without any custom patch, wouldn't it?
Next steps:
- run Vidalia 0.3.x from Debian experimental
- test the branch that implements Tor#2905
- look at our patches, and see what we could drop, have merged, or
implement as a plugin
* more?
Cheers!
--
intrigeri
| GnuPG key @ https://gaffer.ptitcanardnoir.org/intrigeri/intrigeri.asc
| OTR fingerprint @ https://gaffer.ptitcanardnoir.org/intrigeri/otr.asc
