[19:05] <ibeardslee> morning
[19:05] <ajmitch> morning
[19:05] <ajmitch> so, ubuntu as a rolling release...
[19:08] <ibeardslee> as long as the LTS stays stable
[19:08] <ajmitch> that's the point pretty much, that the LTS is kept as the main release
[19:09] <ajmitch> long discussion started on the ubuntu-devel list about it today
[19:10] <ibeardslee> that's the way I reckon it should happen .. stable LTS other 'releases' are bordering on "development", the "let's see how this goes and see it it'll fly in an LTS"
[19:10]  * ajmitch is trying to wade through the thread at the moment :)
[19:10] <ibeardslee> encourage the average home/enterprise user to stick with the LTS
[19:10] <ibeardslee> bleeding edgers roll with the rolling release
[19:11] <ibeardslee> .. my not so humble opinion
[19:40] <hads> Sounds quite good to me.
[20:52] <olly> morning
[20:52] <olly> sounds like debian to me...
[20:53] <ajmitch> though it might be debian with monthly snapshots of testing
[22:06] <snail> what we need is to select distro by section: LTS + rolling games; LTS + rolling java; ...
[22:23]  * thumper avoids commenting
[22:23] <thumper> morning BTW
[22:24] <ajmitch> thumper: now that's not fair :)
[22:25] <thumper> I have too much prior knowledge
[22:25] <thumper> and shouldn't comment
[22:26] <thumper> but I agree that rolling releases are the correct approach
[22:26] <thumper> how else are we going to avoid running out of letters
[22:26] <lifeless> I've been saying for years that debian should stop 'releases'
[22:26] <thumper> I predict a naming scheme change prior to Z
[22:26] <thumper> well, there will still be LTS releases
[22:27] <thumper> like debian stable, no?
[22:27] <ajmitch> that's the current suggestion, yes
[22:29] <lifeless> yeah, I'm more radical
[22:30] <ajmitch> always install from the latest known-good daily image?
[22:37] <ibeardslee> I wouldn't be expecting daily changes on a rolling release cycle
[23:13] <lifeless> ajmitch: right
[23:14] <lifeless> ibeardslee: why not?
[23:15] <ibeardslee> well not as "we must release something today" type daily changes
[23:21] <lifeless> ibeardslee: I don't know what that means.
[23:24] <snail> lifeless: there's pressure to have at least one new package to update every day once you move to a rolling schedule
[23:24] <lifeless> snail: there is?
[23:26] <snail> lifeless: it's one of the standard arguements against rolling release cycles
[23:27] <lifeless> snail: but its an argument against something that doesn't exist?
[23:28] <snail> we see it quite a bit on wikipedia
[23:28] <lifeless> snail: people want to edit a page a day?
[23:31] <snail> lifeless: there are people who want to update the numbers in https://en.wikipedia.org/wiki/List_of_most_viewed_YouTube_videos on a daily basis
[23:31] <lifeless> snail: lol, wow.
[23:31] <lifeless> snail: so consider a distro rolling release; the idea is that you make many small -careful-, -correct- changes
[23:32] <lifeless> each change gets CI tested and promoted
[23:32] <lifeless> and its not 'land in trunk, go to 20M users'
[23:32] <lifeless> its then staged through several successively larger populations looking for errors until it reaches everyone
[23:32] <snail> indeed
[23:33] <snail> but there are some things that are inherently dynamic
[23:33] <lifeless> most users would get batches of things coming through that pass full validation together
[23:33] <snail> things like spam filter data; maps of the world; TZ data;  etc etc
[23:33] <lifeless> only canary populations would get full frequency of updates.
[23:34] <snail> notice that most of that is data, not code
[23:37] <lifeless> sure; but that can also break stuff, so should be in the same regime.
[23:37] <lifeless> And TBH if you look at commercial virus scanners, for isntance - they push that out in realtime.
[23:38] <lifeless> So, I'm not sure why you wouldn't want those things propogating rapidly and efficiently.
[23:38] <ibeardslee> oh you do want that, and security updates