> We had a turbogears planning meeting last week, and discussed a lot of > pieces of the turbogears future. I'll provide a quick summary here, > followed by a slightly organized and cleaned up version of the IRC log for > those who want all the details. > > So, here's what was decided:
<snip> > - We will work out the details of 1.5 and later releases as they happen, > but for now all promotional activity will be around the 2.0 codebase. <snip> > And here's the details: <snip> > Part 5: 1.x discussion > > [3:17PM] "let's have a time-box for every point.," said scfe. > [3:17PM] "like 10 minutes for TG1," said scfe. > [3:17PM] "ok," markramm says "so let's sepend 10 min on tg1" > [3:18PM] "Here's the deal. 1.1 is pretty much final," said markramm. > [3:18PM] "so we have a 1.x release that uses genshi and sa by default," said > markramm. > [3:18PM] "and provides a way to update all your tests to something that'll > be tg2 compatible," said markramm. > [3:18PM] "point here," elpargo says "I think we should have 1.1final before > 2.0final" > [3:19PM] "I agree," markramm says "and I think florent will make that > happen" > [3:19PM] "so what happens to 1.x after 1.1?," asked markramm. > [3:19PM] "Florent has suggested that it go into maintenance mode on the > whole 1.x branch." said markramm. > [3:19PM] "but we have a 1.5 branch which is pretty close to complete," said > markramm. > [3:20PM] "and some recent work has been going into adding a tm2 based > transaction system, to that branch" said markramm. > [3:20PM] "I think this work is useful to some people," markramm says "and if > it provides a stepping stone to 2.x that's a good thing" > [3:21PM] "but the long-term viability of that code line is somewhat limited > by the fact that it's trapped at less than 2.x," said markramm. > [3:21PM] "this is not really that different than python," said markramm. > [3:21PM] "which is now trapped at less than 3.x," said markramm. > [3:21PM] "because of 3.0," said markramm. > [3:22PM] "I think this may become a problem at some point in the future," > markramm says "and my proposed solution is to allow the cp3 enthusiasts to > continue this code line as long as they want" > [3:22PM] "markramm, well the thing is that I don't see tg1.5 as a step up to > 2.0, the whole CP tools concept is very different from anything on tg2.," > said elpargo. > [3:22PM] "but probably it will have to be under a new name after 1.5 or > 1.6," said markramm. > [3:22PM] "Yea they could be doing this in a more tg2 ish way," said > markramm. > [3:23PM] "using repoze.tm2 as middleware," said markramm. > [3:23PM] "what about releasing 1.1/2.0 first," scfe asks "then do 1.5 and > check afterwards if there is a community willing to push 1.5 forward?" > [3:23PM] "ok," markramm says "back to my previous line of thought" > [3:23PM] "1.5 branch may have a long and happy life," said markramm. > [3:24PM] "looking at it from an outsiders perspective," elpargo says "this > is like changing peak.rules and RuleDispatcher." > [3:24PM] "the filters->tools change is fairly minor. the important thing is > the 1st class wsgi support of cp3," said kkuhl. > [3:24PM] "the common-tg dev will say huh? I didn't knew about that.," said > elpargo. > [3:24PM] "if there is," scfe says "let's make a plan under which > name/umbrella 1.5 can develop forward." > [3:24PM] "but at the point where it stops being a transition point to 2.0," > said markramm. > [3:24PM] "it should start maintaining that long and happy life under a new > name," said markramm. > [3:25PM] "that is not to say that 1.5 should be that point," said markramm. > [3:26PM] "but there are things that are in discussion for a 1.6 branch or > whatever that change the API," said markramm, "of expose." > [3:26PM] "kkuhl, my biggest issue with CP in general, is that a lot of > things are done "the cp way", and the original TG >1.0 did some weird things > on CP2, so a lot of people are used to them. I recently started a port of a > tg1.1->2.0 app and the biggest problem was how to translate X into Y.," said > elpargo. > [3:26PM] "I'm supportive of those ideas," markramm says "but I don't want > 1.6 to be less tg2 like than 1.1" > [3:26PM] "i'd say there's some room for innovation outside of just being a > stepping stone," kkuhl says "but yeah, the main thrust of 1.5 should be > adopting the rest of the 2.0 tools" > [3:27PM] "of course, innovation is good. But compatability is also > important." said markramm. > [3:27PM] "CP3 pushed that even further rather than closer.," said elpargo. > [3:27PM] "And as long as we don't start evolving the user facing API's in > different directions they can continue to live in the same "namespace"," > said markramm. > [3:27PM] "Yeah," kkuhl says "1.0 did some weird things on top of CP2. Some > of those we eliminated (testutils), and some of them fumanchu incorporated > back into CP3 (reloader)" > [3:27PM] "elpargo: how so?," asked kkuhl. > [3:28PM] "but if innovation leads to larger differences," markramm says > "rather than smaller, that's when a friendly fork is needed." > [3:28PM] "Does that make sense?," asked markramm. > [3:28PM] "yeah," kkuhl says "absolutely" > [3:29PM] "now for marketing message stuff," said markramm. > [3:29PM] "markramm, but that will still place the issue of, should I port to > 1.5 or 2.0," said elpargo. > [3:29PM] "true," said markramm. > [3:29PM] "and we should give clear direction on that," said markramm. > [3:29PM] "first of all 2.0 will be out next month," said markramm. > [3:29PM] "sorry," Francois_jersey says "from a marketing point of view, I do > not understand why 1.5 or 1.6, but a clear message would be 2.0+" > [3:30PM] "our clean message should be port to 2.0 as soon as you can," said > markramm. > [3:30PM] "to be honest releasing 1.5 or 1.6 will be similar to the > zope2/zope3 problem.," said elpargo. > [3:30PM] "yea," said markramm. > [3:31PM] "people could see it as zope3 is so awesome and great but all the > big players are still at zope2.," said elpargo. > [3:31PM] "i think 1.5 is more like five," percious asks "no?" > [3:31PM] "that problem existed as soon as 2.0 was decided upon. putting out > further 1.x releases doesn't make it any larger," said kkuhl. > [3:31PM] "right," markramm says "but the thing is that there's not nearly > the switching cost between 1.1 and 2.0 as the zope 3 issue" > [3:31PM] "zope 3 is a hugely different beast than zope 2," said markramm. > [3:32PM] "yea," elpargo says "that's true." > [3:32PM] "tg2 is different," markramm says "but I would note that some > people were was annoyed when they took the intro to tg2 class because it was > "all exactly like tg1. ;)" said markramm. > [3:33PM] "kkhul: i agree," said markramm. > [3:33PM] "the problem is not going to go away magically," said markramm. > [3:33PM] "to be honest I'll love to stop asking on the mailing list," > elpargo asks "are you using tg1.x or tg2?" > [3:33PM] "and it was created when we switched from kid-->genshi," markramm > says "so-->sa, cp --> pylons" > [3:34PM] "we will have tg 1.x users for a long," markramm says "long time" > [3:34PM] "markramm, with the exception (it seems) that not everyone agreed > with the cp -> pylons change.," said elpargo. > [3:34PM] "the only reasonable way to stop asking that is to make a 2x > mailing list," said markramm. > [3:34PM] "well," markramm says "absolutely everybody agreed to it at the > time" > [3:34PM] "there were no objections," said markramm. > [3:35PM] "and of course the technical difficulty of supporting cp and pylons > with the same code base :)," said elpargo. > [3:35PM] "Though I registered pygears.org and was prepared to release tg2 > under a different name," said markramm. > [3:35PM] "I didn't knew that :)," said elpargo. > [3:36PM] "got to go... see you tomorrow at the sprint part 2," albertov says > "bye people" > [3:36PM] "ok so wrapping up.," said elpargo. > [3:36PM] "and actually that's what I proposed first was creating pygears," > said markramm. > [3:36PM] albertov left the chat room. > [3:36PM] "fact: CP3 based TG's are "castrated" by the 1.x numbering.," said > elpargo. > [3:36PM] "not making pygears the next version of turbogears," markramm says > "it's just that alberov and dangoor convinced me that it should be tg2" > [3:36PM] "I think we're used our 10 minutes.. the important thing is that we > all recognize that 2.0 is the future of TG. That was decided long ago & > won't change," said kkuhl. > [3:36PM] "agreed," said markramm. > [3:37PM] "and I think that we all need to be on the same page about 1 other > thing," said markramm. > [3:37PM] "1.5+ is going to live," said markramm. > [3:37PM] "thanks. :-)," said kkuhl. > [3:37PM] "that has been said several times.," said elpargo. > [3:37PM] "sorry to be anal," elpargo says "but those seem contradicting > arguments." > [3:37PM] "and it's OK for it to live in the TG namespace as long as the API > is more like tg2 than tg 1.1," said markramm. > [3:38PM] "but it will not be advertised on tg.org as the future of tg," said > markramm. > [3:38PM] "they don't contradict as long as they're moving towards > each-other," said kkuhl. > [3:38PM] "and we will be telling people that tg2 is the future," said > markramm. > [3:38PM] "and that they should port to tg2 not 1.5," said markramm. > [3:38PM] "kkuhl, that's the thing. that's not possible because one will have > to drop it's "core component"," said elpargo. > [3:38PM] "unless they have a compelling reason to go to 1.5," said markramm. > [3:38PM] "it's similar to saying that eventually django and pylons will > merge.," said elpargo. > [3:39PM] "We won't merge," said markramm. > [3:39PM] "but we can live happily side by side as long as 1.x moves towards > 2.x," said markramm. > [3:39PM] "then "the future" is pointing two ways.," said elpargo. > [3:39PM] "Perhaps the 1.5 moving out to live on it's own will happen at > 1.5," said markramm. > [3:40PM] "so there will not be a tg 1.5 release," said markramm. > [3:40PM] "but I'm not convince that that's necessary," said markramm. > [3:40PM] "and you have to take into account that it will probably continue > to live wether we want it to or not," said markramm. > [3:40PM] "imho frameworks can not go away that simple. That's why i'm +1 for > TG 1.5.," said scfe. > [3:40PM] "and the cost of a fork right at the release time of 2.0 is much > larger than the cost of it after 2.0 is established," said markramm. > [3:41PM] "I don't like that idea," elpargo says "TG dev community is small. > Making a "fork" will be harmful." > [3:41PM] "and friendly fork or not," markramm says "there is a cost to the > fork" > [3:41PM] "Yet we can't force people to not work on a CP3 based TG.," said > elpargo. > [3:41PM] "or a pylons TG...," said percious. > [3:41PM] "So," markramm says "the best solution going forward is to let them > ;)" > [3:41PM] "what about worrying about this when the point comes?," asked scfe. > [3:41PM] "What is the benefit of a CP3 TG compared to Pylon?," asked > Francois_jersey. > [3:41PM] "the one thing we haven't talked about lately re: 1.5 is that > without it," kkuhl says "we're asking fumanchu to continue to support CP2, > which is rather rude" > [3:41PM] "just have the message '2.0 is the future'.," said scfe. > [3:42PM] "scfe, Francois_jersey you guys may not be aware of this but there > is working code for that :)," said elpargo. > [3:42PM] "Francois_jersey: if you don't know the answer to that, use TG 2.0 > :-)," said kkuhl. > [3:42PM] "right," said markramm. > [3:42PM] "the future of tg is tg2," said markramm. > [3:43PM] "asking question for marketing purpose to make sure that people > move to tg2," said Francois_jersey. > [3:43PM] "If you have a large investemet in CP," said markramm. > [3:43PM] "1.5 will be easier for you to move too," said markramm. > [3:43PM] "If TG2 grows the community marketing-wise, as is hoped, then 1.5 > will end up a historical footnote," said khrushchev. > [3:43PM] "that's it's main advantage for existing users," said markramm. > [3:43PM] "if the CP3 people agree to work under those conditions then I'm > fine with it.," said elpargo. > [3:44PM] "I think they are all agreed," said markramm. > [3:44PM] "but I could be wrong," said markramm. > [3:44PM] "I want to do everything we can to support 1.5 short of hobbling > 2.0," said markramm. > [3:45PM] "ok timebox for 1.x well past over. :) " said markramm. > [3:45PM] "and I kind of expected us to blow that one out," said markramm. > [3:45PM] "Agreed for the time being :-) There may come a day when I think > 1.5 deserves better marketing, but that day isn't today," said kkuhl. > [3:45PM] "And I think 1.5 can get better marketing from the TG community," > said markramm. > [3:46PM] "particularly after 2.0 is established enough that we won't pollute > our main message," said markramm. > [3:46PM] "right," said kkuhl. > [3:47PM] "so," markramm says "there's a lot of potential for everybody to > win." > [3:47PM] "kkuhl, just to be clear, I'm not against 1.5. I just prefer at > this point in time to make sure 2.0 is a success.," said elpargo. > [3:47PM] "if the overall TG market share grows," markramm says "and some of > that goes to a 1.6 based fork" > [3:47PM] "that's a good thing," said markramm. > [3:47PM] "for the whole community," said markramm. > [3:47PM] "if the fork creates confusion and shrinks the community that's a > bad thing," said markramm. > [3:48PM] "so we all benefit when we all work together ;)," said markramm. Thanks a lot Mark for this transcript! I'd like to ask the dev people behind the 1.x branch what their thoughts are, I assume not everyone took part in the above discussion. Is there a lot term commitment to the 1.x branch? Should the 1.x branch get closer and closer to the 2.x or not necessarily and should it evolve on its own? Will/should tg fork? Any thoughts I'd appreciate very much! Cheers, Daniel -- Psss, psss, put it down! - http://www.cafepress.com/putitdown --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "TurboGears" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/turbogears?hl=en -~----------~----~----~----~------~----~------~--~---
