On Sat, Dec 17, 2011 at 6:02 PM, Dennis E. Hamilton <dennis.hamil...@acm.org> wrote: > Um, are you expecting that the first podling release of binaries will not be > a beta? >
I've learned to rely on testing, not expecting. Let's see what we find out from a good test pass. 3.4 has already been through one beta, but we have made significant changes in some areas, These areas will need to be tested more deeply. And we'll want to see if any critical bugs have been reported in from the earlier beta. I'm not a big fan of using beta releases as substitute for testing, though they are great for feature feedback. But the AOO 3.4 release is not going to differ much functionally from the earlier beta. So I think our emphasis should be testing and fixing bugs until 3.4 reaches a level of quality that we can be proud to release and recommend for use without hesitation. -Rob > - Dennis > > -----Original Message----- > From: Rob Weir [mailto:robw...@apache.org] > Sent: Saturday, December 17, 2011 13:59 > To: ooo-dev@incubator.apache.org > Subject: Re: A timeline for an Apache OO release > > On Sat, Dec 17, 2011 at 4:29 PM, Dennis E. Hamilton > <dennis.hamil...@acm.org> wrote: >> I think it is irresponsible to not mitigate risk by having multiple levels >> of fall-back in place always. Not installing a 3.4 atop a 3.3 is one of >> those safeguards. It is foolish not to take that precaution. It honors >> users by allowing them to compare based on *their* use cases and decide >> when, if ever, to remove a previous version. >> > > The user is free to pick another directory at install time, right? So > they can make the choice that works best for them. > > If you want to improve on the install logic that has worked well for > OpenOffice for past releases, then that is wonderful as well. There > is always room for improvement and patches are welcome. > > But my issues was more with your assertion that 3.4 should come with > "a gigantic disclaimer against production use". I think that is an > irresponsible statement, considering you have not seen or tested a 3.4 > candidate release yet. > > -Rob > > > >> For me, it is always appropriate to leave a previous .x release of a >> productivity product installed. As a matter of policy, I would never >> silently uninstall anything. That is regardless of the presumed quality of >> the new release. >> >> My intention is to safeguard the user first, no matter what my level of >> confidence (or hubris) might be. I am not presuming anything about the >> quality of any non-existent release. I am expressing a principle that does >> not move our risk of error onto the user if at all possible and practical. >> >> - Dennis >> >> -----Original Message----- >> From: Rob Weir [mailto:robw...@apache.org] >> Sent: Saturday, December 17, 2011 12:11 >> To: ooo-dev@incubator.apache.org >> Subject: Re: A timeline for an Apache OO release >> >> On Sat, Dec 17, 2011 at 2:25 PM, Dennis E. Hamilton >> <dennis.hamil...@acm.org> wrote: >>> The releases have to be rebranded anyhow, because they are currently Oracle >>> branded. I think having it be OpenOffice.org 3.4 and installed over >>> OpenOffice.org 3.3 is a very risky idea. The quick-release cycle may be >>> great >>> for our teething; users should not have to suffer any of the consequences. >> >> >> [ ... ] >> >> I think it is irresponsible for anyone to make statements about the >> quality or the suitability for production use of a release they have >> not yet seen, not installed, and not tested. Let's wait to see a >> release candidate before we start issuing speculative predictions that >> have no factual basis. >> >> -Rob >> >> [ ... ] >> >
