-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi everybody!
This mail is about to short term and longer term plans. Lets start with short
term stuff:

What do you think of releasing 1.9.3 tomorrow? Is the game currently in a
"stable enough" state or do you have any changes lying around that you need/want
to incorporate? Please ping me if you have anything that should stop me from
releasing tomorrow.

Now the longer term plans:

Normally we tend to get a new stable series out every year, usually sometimes in
the timeframe from Febuary to April. Somehow I am not 100% sure if we should do
so this release cycle, too. I somehow have the impression that from the coding
side (eg due to the work involved in mentoring GSOC) there were not too many
"larger changes" yet beside what was done as part of GSoC. From the GSoC work
(regarding my own perception, might be wrong, please correct me if I am!) this
is the current status:
- - "Eclipse plugin" is "standalone" and thus not really bound to the release
cycles of Wesnoth itself (beside it targeting current dev versions).
- - "Persistent variables" seem to work, at least for singleplayer. No idea if
anything is open there.
- - "Whiteboard" appears to be working, too. No idea if anything is open there 
and
how much polishing it could use.
- - "Asynchronous Network Architecture (ANA)" seems to work okayish for "end
users" but tends to be crashy on the server side and thus requires some serious
bug fixing.

Beside this there is still the new lobby which is still not really 100% usable.
Other (possible) new features:
- - new addon server (not ready yet, should not depend on the release schedule
since it is run on serverside anyway)
- - the new/updated terrains which make the game *really* look different, but
animations put some extra strain on the CPU (water!)
- - added the campaign "Dead Water" which seems to be working nicely
- - Before starting the 1.9.x series we also considered adding a new faction for
multiplayer, this has not happened yet and I don't know the status of the work

In general I have the impression that we are not really ready to go into any
kind of feature freeze "soonish" enough to release in the timeframe of February
to April (don't forget that eg translation teams need some significant string
freeze, I'd say that some 6 to 8 weeks would make sense as minimum!).

The main reason of releasing at "early time of the year" is that this way we are
in "heavy development" when Google Summer of Code starts so that huge and
groundbreaking features can be added. So if we want to participate in GSoC 2011
(not sure if there will be one and if we are accepted) we should either release
before April or after GSoC, which ends late in August (meaning that when adding
some feature/string freeze the release could be some time in November/December).

Personally I don't feel too comfortable with the idea of releasing soonish with
the amount of features/changes that we have so far. Some more reasons why I'd
prefer moving the release schedule to "later":
- - Our devs tend to be more active in the winter months (many are from 
Europe/the
northern hemisphere and when it is rather hot outside people tend to enjoy the
weather instead of coding...). Why not allow larger features to happen in this 
time?
- - With the 1.8 cycle we had a rather longish feature freeze resulting in 
several
devs starting branches and working on them, which is not as nice as the "real
work" happening directly in trunk.
- - GUI2 is still in a "needs work and is just partly done" state. maybe when
allowing "large term changes" we have it finished in the next stable series?
- - alink started to work on an OpenGL port of Wesnoth. Starting rather hacky 
just
trying to 1:1 replace sdl functionality with OpenGL functionality where
required/useful. This work currently resides in an extra branch and is far from
finished/polished. Maybe with a move cycle this work can be completed and 
tested.

If OpenGL and GUI2 were done I'd even label the release 2.0, since those are
really groundbreaking things and combined with the other changes since 1.0 we
have *really* done huge tasks.

What do you think regarding changing the release cycle to not release in the
timeframe Febuary-April 2011? Yes, personally I love the system "release early,
release often". But getting out a new stable series does always require some
additional efforts that I currently think would be better spent in getting the
existing features into a state where we can say "done!". I'd love to hear from
you regarding the release plans.


As you saw mentioned above GSoC is also somehow relevant to the release plans.
If we participate we need to allow students to commit stuff somehow. Since those
are usually not tasks like "polish ABC" but more like "create some new feature"
or "reimplement ABC in a sane way" they can basically not happen in mainline
during a feature freeze. And in case that the students vanish shortly after GSoC
I would not like to have the work (that we could later on use) happen in an
extra branch that would be hell to merge back. The general question that came to
my mind is:

"Should we (try to) participate in GSoC 2011?"

Pro:
- - Possibly brings us new contributors.
- - New larger features get tackled and implemented.
- - It is a cool idea, offering students a nice summer job.

Con:
- - Requires developers resources. Those devs that mentor in GSoC have less time
for their own "projects" inside Wesnoth.
- - Many students leave shortly after GSoC, the work is done, but some polishing
is still required. Often some "other" dev has to handle this, which possibly
requires some time to get used to the code.
- - If we don't participate in GSoC some other project could participate in our
place and experience GSoC.
- - We have to accommodate GSoC somehow in our release cycle/plans.

What do you think, should we participate in the next GSoC or not?


So much about those questions. Please do also consider the mail I sent some mins
ago regarding FOSDEM. Some of the stuff can most likely be discussed at FOSDEM,
too (eg GSoC, especially since most of our GSoC mentors so far tend to come to
FOSDEM, too). I'd love to hear what you think regarding a new stable series as
well as GSoC and no idea what other planning we require.
Cheers,
Nils Kneuper aka Ivanovic
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.16 (GNU/Linux)

iEYEARECAAYFAkz+ItcACgkQfFda9thizwUEBQCgk92LHL1e+qSCYRPDvQxJ3nra
cacAn3SCpvsERjnv3P9CkMoN4no/LfOZ
=t1yL
-----END PGP SIGNATURE-----

_______________________________________________
Wesnoth-dev mailing list
[email protected]
https://mail.gna.org/listinfo/wesnoth-dev

Reply via email to