Lumiera
The new emerging NLE for GNU/Linux

March 9, 2011 on #lumiera
20:00 GMT - 23:30 GMT

Participants

  • Christian Thaeter (cehteh)

  • Francesco Siddi (fsiddi)

  • Hermann Voßeler (ichthyo)

  • Odin Hørthe Omdal (Velmont)

  • Raffaella Traniello (raffa)

  • Stefan Kangas (skangas)

The New Website

The new website is finally online.

How do we proceed with the graphical layout?

21:37 <fsiddi> the template code is not ready yet
21:37 <fsiddi> and there are some incorrect uses of tags in the current one
21:37 <fsiddi> so i will go on coding a static html layout
21:38 <fsiddi> then will submit it to your critique and then we can make a template out of it
...
21:50 <ichthyo> fsiddi: I liked the way you provided a slightly larger content area
                           for the tutorial part

Discussion about the IFrame

The only serious alternative to IFrames seems to be SSI, since we want to keep generating pages with asciidoc, and not have to regenerate the entire website just because one page changed.

Conclusion was that we will stick with IFrames for now and change later when/if that becomes a problem.

Old stuff in the repository

There is a folder /attic on the website.

This information is supposed to be merged with the rest of the website, and then deleted from its current location. This work is on-going and considered a "background task".

IRC Logs

A discussion was brought up about how to handle IRC logs.

For privacy reasons we do not want to save them routinely in public locations. If we put them on the web they will most probably be kept by Google until the sun burns out.

robots.txt might be respected by Google or not. Surely some other search engines will not respect it. Private logs or other logs not publically available are fine though.

22:40 <cehteh> well ... IRC should *not* be for persistent documentation .. we shall
               force/educate ourself to document things (and decisions) properly, irc
               should be volatile
...
22:41 <ichthyo> in the past, sometimes we had very "contentful" dicussions
22:41 <ichthyo> in that cases we just saved a transcript
...
22:44 <ichthyo> so personally I prefer to make transcripts of the *really* contentful
                discussions and put them in the public documentation
22:44 <ichthyo> "transcript" means cleaned up irc log
22:44 <cehteh> yes
22:45 <cehteh> so for meetings we should prolly do that
22:47 <cehteh> not further edited but just leave all join/parts/noise and silly
               comments out
22:47 <cehteh> (and also rants and stuff)
22:48 <cehteh> only conclusions and most important arguments shall stay
22:48 <cehteh> maybe reorder it a bit as it happens that we talk about 2 things at the
               same time
22:48 <cehteh> but also not too much work on it
22:49 <cehteh> and more importantly: document decisions and proposals in a more
               offical way: that is 'rfc'
22:49 <skangas_> Agreed. It is more important to preserve information in an accessible
                 way for the future than focusing on mundane details like
                 join/part/order/ranting etc.
22:50 <skangas_> This means I personally always prefer shorter summaries of a
                 discussion to IRC logs, given that they do not leave anything
                 important out.

VoIP discussions over Mumble?

There seems to be ongoing discussions. on using VoIP more in Debian GNU/Linux.

It was decided that while the idea is nice, it will not be used for developer meetings. We might use it for developer discussions outside meetings if we feel the need though. This will probably be more common as the number of developers increase.

Also, the point was brought up that GNU Emacs has really nice support for several people sharing the same session. Everyone gets their own point, and with some hacks it is even possible to color them differently. This needs to be done on an untrusted remote server though — Emacs can do anything and it is thus highly insecure.

Version numbering

Hermann has written an rfc for version numbers.

23:19 <cehteh> so in my words: we will have a usuall major.minor.patch for releases
23:20 <cehteh> and   major+1~develtag for devel snapshots
23:20 <cehteh> where develtag is a timestamp
23:20 <ichthyo> and we *could* make a stream of development versions
23:21 <cehteh> YYYYMMDD should suffice
23:21 <cehteh> or maybe just monotonic increading from 1
23:21 <ichthyo> well... optional suffix+ number
23:21 <ichthyo> for the rare cases that you roll two releases a day
...
23:20 <ichthyo> and the third idea of my proposal: *not* use major.minor.patch before 1.0
23:24 <ichthyo> i.e. in the whole alpha... beta range
23:24 <ichthyo> we just do 0.01  0.02 ... 0.99
23:25 <skangas_> Does it go without saying that minor versions can go on above 1.9.0
                 to 1.10.0 etc?
23:25 <ichthyo> yes, I didn't mention that, but thats important
23:26 <ichthyo> its not nice, and usually you get a major first, but it can happen
23:27 <ichthyo> so i'd say, we have discussed it now, we could comment on it and then
                revisit it next time (and maybe decide on the proposal then)
23:27 <skangas_> Do we want a stable version and a development version?
23:28 <skangas_> e.g. 1.0 is stable and 1.1 is unstable
23:28 <skangas_> I am thinking about the scheme that wine and linux is using.
23:28 <cehteh> i think linux won a lot with abadon this odd/even practice
23:28 <ichthyo> regarding stable versions
23:28 <ichthyo> personally, I don't like that idea
23:29 <skangas_> cehteh: I did not know they had abandoned it.
23:29 <cehteh> skangas: linux dont use that anymore
23:29 <cehteh> since 2.6 years ago
23:29 <ichthyo> because they tend to rot
23:29 <skangas_> ichthyo: The stable versions?
23:29 <cehteh> you just make a release and done .. everything after that is for the
               next version
23:29 <ichthyo> yes
23:29 <cehteh> the only thing to discuss is how you count there
23:30 <ichthyo> also, it is more in line with the "extreme programming" style
23:30 <ichthyo> i.e. don't do too disruptive changes
23:30 <cehteh> if you want 'pre' or 'rc'
23:30 <ichthyo> but rather refactor often
23:30 <skangas_> OK.
23:31 <ichthyo> there is onre concern though
23:31 <ichthyo> there *needs* to be some stability then
23:31 <cehteh> releases get only bugfixes and *maybe* you can declare some release as
               LTS ... but for lumiera i would think really hard about that and likely
               deny that
23:31 <ichthyo> because people are working with the software
23:31 <ichthyo> but we want to have that with our compatibility scheme
23:31 <cehteh> (our interface system should provide backward compatibility)
23:31 <ichthyo> yes
23:31 <ichthyo> but anyway, it *might* turn out that we fail with that promise
23:32 <ichthyo> and then we'll have to reconsider a "stable" line
23:32 <cehteh> so maintaining some LTS while having another development branch will
               suck developer resources
23:32 <ichthyo> yes
23:32 <ichthyo> so I'd really try hard to avoid that
23:33 <cehteh> well a stable line yes ... that is 'stable' until the next stable
               release is there .. and then 'maintained' as in adding important
               bugfixes for some (but not excessive) time
23:33 <cehteh> we need to care for a stable for sure
23:33 <ichthyo> indeed, but it shouldnt digress too much from the active development
23:33 <cehteh> but we dont want some really old release to get permanent updates and
               care because we declared it as LTS
23:33 <skangas_> ichthyo: Is your RFC open to editing/additions?
23:34 <ichthyo> yes
23:34 <ichthyo> please go ahead, just add comments etc.
23:34 <cehteh> use the rfc.sh comment function :)
23:34 <skangas_> I was just thiking about writing this down in some way, that we
                        specifically do not want this.
23:23 <cehteh> also do we want 'rc' releases?
23:23 <cehteh> or do we just declare certain devel snapshots as rc (i like the later)
23:23 <ichthyo> yes, thats what I'd prefer too
23:24 <cehteh> devel snapshots should be automatic buildable by builddrone
23:25 <cehteh> I would like to make builddrone an automatic staging system
23:25 <cehteh> add some 'rules' there ..
23:26 <cehteh> if build and testsuite succeeds and test-install works and latests
               commit doesnt contain WIP: then stage a devel snapshot
23:27 <cehteh> (eventually or maybe we might even put tests/bugs there which prevent
               certain staging actions, finally all integration and complete releases
               could be automatized)
23:26 <cehteh> remember in test.sh i reserved the 90-99* area for 'bugs'
23:26 <cehteh> we may put some blocking tests there which prevent staging

Next meeting

The next meeting will be at Wednesday April 13 20:00 UTC.