I just want to make one thing clear related to this: if _I_ make changes that are not backward compatible then it's because they are really important and no one should question them; if anyone else makes (or proposes) a change that I don't like and I can raise the backward compatibility flag then I will, and you should respect that and just don't do whatever the thing is. Don't worry, I'll be sure to make strong statements and appeal to authority and popularity of patterns so that you can justify it to whoever you feel responsible to. Of course, those outward reasons are the very things that you'll never be able to argue against, no matter how inapplicable or extreme or pragmatically unhelpful they might be.
Now, if anyone disagrees with my position on this then that's fine, as long as you don't feel like you can behave this way and that you never have behaved this way. Any challengers? -Not David P.S. On a more serious note, I'm not a huge fan of universal backward compatibility (or of changes that cause hidden breaking of backward compatibility... better to break things completely than make things seem like they are working even though existing code fails, like type conversion changes and such). Sometimes it takes a while to figure out the best way to do things, and why keep the problematic and bad ways of doing things when we do figure it out? If a hexagonal wheel works so much better than a square wheel that square wheels don't make sense anymore, to the point where they caused more problems than they solved, why not encourage everyone to change and just stop supporting the bad design? And yes, sooner or later we're gonna figure out that a good round shape is better than a hexagon and then we'll have to change again... and it'll be WELL worth it and users will thank us a lot for fixing those issues. Just because we haven't fixed a bad idea for a long time doesn't make it any better of an idea. On Apr 20, 2010, at 11:56 PM, Ashish Vijaywargiya wrote: > For now we can create new target and can keep old as it is. And as the > time passes we can recommend the use of ant run-install-demo instead > of ant run-install. > > After certain duration, lets say after 6 month or one year we can > deprecate the use of ant run-install from the system. > > I think we can start *VOTE* on this. > Thoughts? > > -- > Ashish > > On Wed, Apr 21, 2010 at 10:17 AM, Ashish Vijaywargiya > <[email protected]> wrote: >> +1. >> >> -- >> Ashish >> >> On Wed, Apr 21, 2010 at 10:13 AM, Raj Saini <[email protected]> wrote: >>> In my opinion run-install is misleading as it really does not tell that it >>> is going to install demo data and cause confusion to new users as it has >>> happened with me in the past. For the benefit of old and new users, I >>> suggest to add new command run-install-demo and leave the run-install as it >>> is for a while and document them to tell what exactly they do. There are >>> numerous examples of this in the Linux/Unix world. >>> >>> Thanks, >>> >>> Raj >>> >>> On 21/04/10 09:39, Jacopo Cappellato wrote: >>>> >>>> On Apr 21, 2010, at 1:15 AM, Ean Schuessler wrote: >>>> >>>> >>>>> >>>>> Jacopo Cappellato wrote: >>>>> >>>>>> >>>>>> This is just *your* opinion and I respect it (even if comparing this to >>>>>> altering an api is ridiculous)... but please quit with the "teacher/guru" >>>>>> mode... >>>>>> >>>>>> >>>>> >>>>> Why is it ridiculous to think of shell script parameters as an API? You >>>>> would surely be surprised if "ls" became "rm" one day, as an extreme but >>>>> valid example. >>>>> >>>> >>>> As a new user of the "OFBiz Linux OS" I would be also very surprised if >>>> the "rm" command really was a "rm -r" command; in my opinion this is worth >>>> of a change, even if old users would have to read one line of documentation >>>> to learn the change. >>>> But really, it is not very important, I am more than happy to leave things >>>> as is because I understand that this could be annoying for old users and >>>> frankly speaking I don't have time and energy to further discuss this. >>>> >>>> Jacopo >>>> >>>> >>>>> >>>>> I think we can safely regard shell scripts as a class of >>>>> program. >>>>> >>>>> Regarding "run-install", we've set an expectation that run-install will >>>>> give you a demo system and that could throw people off. Changing it >>>>> doesn't seem hazardous but I'm not clear that it adds value. >>>>> >>>>> -- >>>>> Ean Schuessler, CTO >>>>> [email protected] >>>>> 214-720-0700 x 315 >>>>> Brainfood, Inc. >>>>> http://www.brainfood.com >>>>> >>>>> >>>> >>>> >>> >>> >>
