Re: 11.5 Stability ?

On Jun 11, 8:23 pm, Fernando Nunes <domusonl...@gmail.com> wrote: > PeterP wrote: > > On Jun 10, 11:36 pm, Fernando Nunes <domusonl...@gmail.com> wrote: > >> Jim Tranny wrote: > >>> On Tue, Jun 10, 2008 at 5:51 AM, Fernando Nunes <domusonl...@gmail.com> wrote: > >>>> PeterP wrote: > >>>>> So when is 11.5 Stable enough to use? > >>>>> Perhaps that should read - When is 11.5 stable enough to feel safe to > >>>>> use. > >>>>> Knowing a bit about the bugs in early versions, I'm wondering how 11.5 > >>>>> is shaping up? Normally I'd say something like XC3 or greater - Maybe > >>>>> IBM should release 3 fp's in the first month so people take it up, > >>>>> similar to how they confuse you with DB2 fp/version numbering. > >>>>> I guess you could go down the path of install on a dev system and see > >>>>> if anything breaks badly. Just wondered thoughts & experiences of > >>>>> peoples here. > >>>>> i.e. if your not a 24x7 mission critical bank using strange datatypes > >>>>> and new features then it's fine to try at FC1. > >>>>> TIA. > >>>> I will not talk specifically about 11.50.FC1. But as a general rule, I would > >>>> avoid using an early version in a production mission critical system... Not > >>>> only Informix, but any other software in the world. Specially when it contains > >>>> new features. > >>>> In any case, there were good FC1 versions and bad FC1 versions... There were > >>>> historically bad FC* versions also... The main principle should be the same for > >>>> any version: Test it as much as you can in your environment... All versions > >>>> have bugs. Some of them are nasty, others are "cosmetic", and many of them are > >>>> transparent if you're lucky enough to not use the feature or not hit the > >>>> conditions that make them visible... > >>>> That said, I have no knowledge of any nasty bug with 11.50, but this may be > >>>> just because not many people are using it... It was launched just a month ago, > >>>> so possibly nobody is using it in critical systems... > >>>> One good point that may not be obvious for the customers: When IBM finds a bug > >>>> in one of the families (9.x, 10.x, 11.10, 11.5) the bug is tested against the > >>>> other versions. So, it's correct to assume that 11.5 includes fixes done in > >>>> late 11.1 and 10 for example. Obviously, due to product launch and code > >>>> freezing, possibly  not all of them were applied. > >>>> My advice: Test it, check the bug alerts and other IBM notes. Besides, there > >>>> are a number of new features that should keep you busy for some weeks in order > >>>> to get used to them. > >>>> Regards > >>>> _______________________________________________ > >>>> Informix-list mailing list > >>>> Informix-l...@iiug.org > >>>>http://www.iiug.org/mailman/listinfo/informix-list > >>> It is increasingly difficult to use the methodology of waiting for a > >>> couple of fixpacks before adopting a new release.  For instance there > >>> was not a lot of time from when version 10 that version 11.1 was > >>> released (at least when you look back at Informix' history).  And now > >>> version 11.5 comes out when version 11.1 was barely out of the door > >>> and it is my understanding that the 11.1 line will not continue, so > >>> anyone looking to upgrade to version 11 should be testing 11.5 > >>> instead. > >>> As for myself, I'll be following Fernando's advice. > >> Well, that was true for 11.10. Not so for 10. 10 is currently at FC8 which is a > >> reasonable number of fixpacks. The 11.10 -> 11.50 cycle was in fact very short > >> (only one year). The fact is that there were features too important to appear > >> in a 11.10 FC3/FC4... Let's wait and see, but it's possible tat newer versions > >> take a little longer. Meanwhile for customers using the main platforms, don't > >> forget the PID (Post Interim Drops), identifiable by major.minor.F*W*... These > >> Ws include *cumulative* fixes. I personally think think brings great value for > >> customers. And this decreases the need to build specific patches > >> (major.minor.F*X*), so it's also good for IBM. > >> So, it's possible we see a 11.50.FC1W1 before a 11.50.FC2. Keep an eye on > >> FixCentral. > > >> Regards- Hide quoted text - > > >> - Show quoted text - > > > Fernando, that is excellent advice in your first post. > > > Jim, where did you get V11.1 will not be around much longer? IBM seems > > to be saying it's the one that'll be around the longest > > > Looking at the PLC page -http://www-306.ibm.com/software/data/support/lifecycle/ > > And policy type (Standard and Extended) - > >http://www-306.ibm.com/software/support/lifecycle/lc-policy.html > > > V9.40 - 30.04.2008 > > V10.0 - 30.09.2010 > > V11.1 - 30.09.2012 - Due to extended policy 5yrs > > V11.5 - 30.09.2011 - Assuming the 3 yr policy rule > > > So if you don't need the 11.5 features, at the moment, 11.1 is going > > to be around longer? Very strange but an assumption based on a policy. > > I think that what Jim was saying is that we won't see much more 11.10.FC* > versions... Not that it will be unsupported. > > > Does anyone know how the code is branched for 11.5? Is it 11.1 + the > > new features or are there major re-writes in all areas of the code. > > And how is 11.1 Different from 10 which is fairly stable. I'm sorry if > > thats not very clear I'm after more of an engineering answer. > > Is 11.5 really 11.1.FC4 but had new features so it got called Cheetah > > 2. If thats not the case, then how come IBM is maintaining so many > > code lines particularly by including V11.1 in the extended policy? > > I can't speak for the development/IBM, but AFAIK, IDS 11.5 is not a code > re-write. It's based on 11.10 with new features. > This is not an Engineering answer of course. But what I mean is that 11.10 and > 11.50 share most of the code as far as I can tell and with all the disclaimers > applied... > > Regards.- Hide quoted text - > > - Show quoted text - IDS 11.50 lives on a bit longer. It seems the IBM PLC page was incorrect, do IBM really want people to use Informix? Apart from the PLC mistake (I wonder if 4GL will also be fixed tomorrow), It's now 33% more expensive than DB2 I'm told - that said, IBM would need to be together enough to actually invoice you. From IBM: "I have verified IDS 11.10 and 11.50 support policy with Engineering. They confirmed that the information on the PLC website was incorrect. Both IDS 11.10 and 11.50 for both Enterprise and Express Edition both have support policy of 'E'. That means they are all guaranteed to be supported for 5 years plus 3 years Service Extension. The information in PLC website will reflect changes in 24 hours."