The short story: Upgrade to beta gems or wait for 0.13.1 during the weekend. The full story: http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the-slowdown-blues-after-going-013/ -- David Heinemeier Hansson http://www.loudthinking.com -- Broadcasting Brain http://www.basecamphq.com -- Online project management http://www.backpackit.com -- Personal information manager http://www.rubyonrails.com -- Web-application framework
David Heinemeier Hansson wrote:> The short story: Upgrade to beta gems or wait for 0.13.1 during the weekend. > The full story: > http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the-slowdown-blues-after-going-013/This suggests that you released code that hadn''t been tested in beta. Is that so? Justin
On Jul 8, 2005, at 6:33 PM, Justin Forder wrote:> David Heinemeier Hansson wrote: > >> The short story: Upgrade to beta gems or wait for 0.13.1 during >> the weekend. >> The full story: >> http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the- >> slowdown-blues-after-going-013/ >> > > This suggests that you released code that hadn''t been tested in beta. > Is that so?Justin, everything pre 1.0 IS beta. - Jamis
Jamis Buck wrote:> On Jul 8, 2005, at 6:33 PM, Justin Forder wrote: > >> David Heinemeier Hansson wrote: >> >>> The short story: Upgrade to beta gems or wait for 0.13.1 during the >>> weekend. >>> The full story: >>> http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the- >>> slowdown-blues-after-going-013/ >>> >> >> This suggests that you released code that hadn''t been tested in beta. >> Is that so? > > > Justin, everything pre 1.0 IS beta.Thanks, Jamis. I don''t want to get entangled in a debate about the meaning of words. You have a mechanism for allowing users to test code before it is released, and it doesn''t seem to have been used as well as it could. You are doing great work, but that doesn''t mean that there''s nothing that could be improved. Justin
On Jul 8, 2005, at 7:29 PM, Justin Forder wrote:> Jamis Buck wrote: > >> On Jul 8, 2005, at 6:33 PM, Justin Forder wrote: >> >>> David Heinemeier Hansson wrote: >>> >>> >>>> The short story: Upgrade to beta gems or wait for 0.13.1 during >>>> the weekend. >>>> The full story: >>>> http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the- >>>> slowdown-blues-after-going-013/ >>>> >>>> >>> >>> This suggests that you released code that hadn''t been tested in >>> beta. >>> Is that so? >>> >> Justin, everything pre 1.0 IS beta. >> > > Thanks, Jamis. > > I don''t want to get entangled in a debate about the meaning of > words. You have a mechanism for allowing users to test code before > it is released, and it doesn''t seem to have been used as well as it > could. > > You are doing great work, but that doesn''t mean that there''s > nothing that could be improved.You''re right, Justin, and I apologize for my hastily composed reply. However, we haven''t used a formal beta testing process before (though we will move more in that direction, soon, as we ramp up towards the 1.0 release), mostly because each release IS the beta release. We''re adding features, tweaking things, and yes, things are bound to break, but that''s what the release is for. We test as well as we can "internally", and when we feel it is "stable enough", we release. When the larger user base starts hammering on it, the bugs get discovered, logged, and fixed. You''ll notice the slow down has already been identified and resolved in the beta gems. As David announced, the 0.13.1 release is imminent. It''s all iterative. It''s about getting real as quickly as possible. If 0.13 is too unstable for someone, they can always stay with 0.12.1 until the next release of Rails. It certainly doesn''t hurt any feelings for someone to be cautious. (But implying that the core developers are irresponsible is never a very friendly thing to do. Just because our process is not the process you are used to, does not mean the process is any less valid.) - Jamis
Justin Forder wrote:> This suggests that you released code that hadn''t been tested in beta. > Is that so?This is not the official process (there is none that I know of). Since 0.8 I have always waited for the 0.x.1 release. In every case the .0 release had problems which were fixed only a couple of days later. I think it would have helped if the release would simply be called "0.x.0 Beta 1..n" with a call to action for everybody to try it. After the smoke clears call it the "0.x.0 Stable". Just to avoid the confusion. Although I can surely live with the inofficial system, knowing that any problems will be addressed in no time. Sasa
On Jul 8, 2005, at 7:03 PM, Sascha Ebach wrote:> Justin Forder wrote: >> This suggests that you released code that hadn''t been tested in beta. >> Is that so? > > This is not the official process (there is none that I know of). > Since 0.8 I have always waited for the 0.x.1 release. In every case > the .0 release had problems which were fixed only a couple of days > later. I think it would have helped if the release would simply be > called "0.x.0 Beta 1..n" with a call to action for everybody to try > it. After the smoke clears call it the "0.x.0 Stable". Just to > avoid the confusion. Although I can surely live with the inofficial > system, knowing that any problems will be addressed in no time.Sascha has a pragmatic outlook. As we approach 1.0 you''ll see more release candidates to encourage folks to hammer the latest code. But, like many things version-numbered, this comes down to psychology; don''t get tripped up on it. We welcome your eyeballs on the beta gems and your patches in the Subversion repository. Best, jeremy
Jamis Buck wrote:> On Jul 8, 2005, at 7:29 PM, Justin Forder wrote: > >> Jamis Buck wrote: >> >>> On Jul 8, 2005, at 6:33 PM, Justin Forder wrote: >>> >>>> David Heinemeier Hansson wrote: >>>> >>>> >>>>> The short story: Upgrade to beta gems or wait for 0.13.1 during >>>>> the weekend. >>>>> The full story: >>>>> http://weblog.rubyonrails.com/archives/2005/07/08/feeling-the- >>>>> slowdown-blues-after-going-013/ >>>>> >>>> >>>> This suggests that you released code that hadn''t been tested in beta. >>>> Is that so? >>>> >>> Justin, everything pre 1.0 IS beta. >> >> Thanks, Jamis. >> >> I don''t want to get entangled in a debate about the meaning of words. >> You have a mechanism for allowing users to test code before it is >> released, and it doesn''t seem to have been used as well as it could. >> >> You are doing great work, but that doesn''t mean that there''s nothing >> that could be improved. > > You''re right, Justin, and I apologize for my hastily composed reply. > However, we haven''t used a formal beta testing process before (though > we will move more in that direction, soon, as we ramp up towards the > 1.0 release), mostly because each release IS the beta release. We''re > adding features, tweaking things, and yes, things are bound to break, > but that''s what the release is for. We test as well as we can > "internally", and when we feel it is "stable enough", we release. When > the larger user base starts hammering on it, the bugs get discovered, > logged, and fixed. You''ll notice the slow down has already been > identified and resolved in the beta gems.Thanks. I only became aware of the beta gems recently, when it became necessary to use them in order to follow the second beta of the Agile Rails book. There are perhaps 2,500 people in the same position. Unless someone explicitly tells us otherwise, "beta" has a well-understood meaning. People will be a bit miffed if released code shows regressions relative to the beta, and they won''t be delighted by finding new features in the release. New features mean code that hasn''t been through the beta process.> As David announced, the 0.13.1 release is imminent. It''s all iterative. > It''s about getting real as quickly as possible. If 0.13 is too unstable > for someone, they can always stay with 0.12.1 until the next release of > Rails. It certainly doesn''t hurt any feelings for someone to be > cautious. (But implying that the core developers are irresponsible is > never a very friendly thing to do. Just because our process is not the > process you are used to, does not mean the process is any less valid.)I apologise if I appear to have been criticising the developers. Release management is a process in its own right. Looking at it critically is part of the "due diligence" that anyone evaluating Rails for use in non-trivial applications should go through. regards Justin PS I really should get some sleep - it''s 3:30am here.
Jeremy Kemper wrote:> > On Jul 8, 2005, at 7:03 PM, Sascha Ebach wrote: > >> Justin Forder wrote: >> >>> This suggests that you released code that hadn''t been tested in beta. >>> Is that so? >> >> >> This is not the official process (there is none that I know of). >> Since 0.8 I have always waited for the 0.x.1 release. In every case >> the .0 release had problems which were fixed only a couple of days >> later. I think it would have helped if the release would simply be >> called "0.x.0 Beta 1..n" with a call to action for everybody to try >> it. After the smoke clears call it the "0.x.0 Stable". Just to avoid >> the confusion. Although I can surely live with the inofficial system, >> knowing that any problems will be addressed in no time. > > > Sascha has a pragmatic outlook. As we approach 1.0 you''ll see more > release candidates to encourage folks to hammer the latest code. But, > like many things version-numbered, this comes down to psychology; don''t > get tripped up on it. We welcome your eyeballs on the beta gems and > your patches in the Subversion repository. > > Best, > jeremyThanks Jeremy. Goodnight from a rather edgy London. Justin