On Sat, Jul 20, 2013 at 7:49 AM, Virgil Arrington <cuyfa...@hotmail.com> wrote: > On Friday July, 19, 2013, Rob Weir wrote: > >> We've discussed AOO 4.0 many times, on the blog and in social media, >> and this has been covered in the press. Yes, we don't issue a press >> release every week or every time we change code indentation, like some >> other projects seem to do. But we do take care of major >> announcements. > > >> I think the pace of development is one reason for the better quality. >> I'd like to release more often as well, but I don't want to >> compromise on quality. But I think there is room for improvement >> here. And we are discussing having a public beta for AOO 4.1. > > > I have complained on the LO user's list of its pace of releasing new > versions. There are several to choose from at its download page, and the > latest often contains bugs that had been fixed in earlier releases. It can > be quite frustrating to download an update only to find a bug that you had > thought was fixed. > > But... > > The slow pace of development at Apache is equally frustrating. AOO 3.4.1 is > a nice program ... except for the inability of the U.S. English version to > properly hyphenate words (See bug 119087). This bug has been around for > years preventing the use of AOO for serious work in America when hyphenation > is required. I assume (perhaps incorrectly) that it will be corrected in > Ver. 4, but it has been frustrating to wait for Apache to release a new > version until it gets everything right. Perhaps some sort of interim release > fixing known and critical bugs could be made. > > Surely there can be some compromise between LO's torrid release pace and > AOO's seemingly non-existent pace. >
I think the compromise then is with quality. Think if it this way: any release has fixed and variable costs. The main fixed cost is testing. Any release, no matter how small, needs to be tested. And given the complexity of AOO (from a code and architecture viewpoint) this means a test of every area of the product. We have over a thousand test cases defined for AOO that we try to run on all major platforms before we release. This is a fixed chunk of work and it can take a couple of months. The variable costs, of course, are the development work that goes into adding features and fixing old bugs. Now, in theory, we could have a release every quarter, but that would mean we do only 1 month of feature work and 2 months of testing. That, I think, would be very inefficient. We could also drop our quality goals and do less testing. Or ship based on dates without any fixed test execution goals. That would allow us to release more frequently as well. I don't think either kind of "compromise" is what our users really want. IMHO, if we want to release more frequently then we need to find a way to accomplish the same quality goals, but in less time. So cut the 2 months of testing down to 1 months, or even less. This could be done, hypothetically, with more test automation and/or more test volunteers. Also, a public beta or "bug finding contests" is not a substitute for formal QA. These things tend to be highly redundant, shallow feature testing. But they can be a good way to get early feedback. Regards, -Rob > Virgil > > --------------------------------------------------------------------- > To unsubscribe, e-mail: users-unsubscr...@openoffice.apache.org > For additional commands, e-mail: users-h...@openoffice.apache.org > --------------------------------------------------------------------- To unsubscribe, e-mail: users-unsubscr...@openoffice.apache.org For additional commands, e-mail: users-h...@openoffice.apache.org