> Pier P. Fumagalli wrote: > James Duncan Davidson <[EMAIL PROTECTED]> wrote: > > On 12/18/00 8:28 AM, "[EMAIL PROTECTED]" <[EMAIL PROTECTED]> wrote: > > > >> To come back out of the blue and announce that Ant is out of control and > >> must be reigned in is more than a little arrogant IMHO. > > > > So, I take it that you would answer yes to my "Am I being arrogant pig?" > > question. :) > > > > I said that because other people have said it to me. Jon, Pier? Ya there? > > There's two people right there. > [snip] > Yep... Kinda late on emails (around 500 yet to read on this list) but > somehow here. What I told James (since he's the person closer to Ant I see > most) is that I hated how that little creep changes every single attribute > on every single release... And I have to go ahead and modify my bazillion of > makefiles.
There were just two releases. And how many build files do You have exactly? > Gee, just take a look at the other build tools available: GNU Make. I > recompiled recently my old mod_jserv (almost 2 years without a change) with > the latest version of GMAKE and IT WORKS! It doesn't say "that **** is > deprecated, use another". > IIRC, make is over 20 years old, Ant not even 20 months! Do You really expect backwards compatibility from day 1? Besides, no one was forcing You to switch to a newer version. Although I admit that changing the properties behaviour forth and back caused some hassle, I think You are overdoing it a lot. > I stopped posting on this list when someone started saying that they wanted > to put JavaScript in makefiles, and to repeat myself, I said "I'm going to > checkout my own local copy of Ant and maintain it". That's what I did, > because over here, in ANT land things change way too often. Didn't You notice that the current committers blocked that request successfully? > Now, when I ranted with James and he later on showed me his proposal, I was > so happy to see that someone was "re-analyzing" the problem from bottom up, > and trying to bring some order I almost cried. The same thing was done on ant-dev, with the only difference that more than one person was contributing to that "re-analyzing". > And my 0.02$ are, from what I read here, everyone is scared that James is > going to take over the world, scared to loose their own little piece of > "kingdom" because the new "dictator" is arising, without even listening to > what he has to say "technically". Quite the opposite. When James announced he would spend more time and effort to Ant, everyone applauded. The dispute started when he claimed to have special rights to determine the direction of Ant. And, frankly, this is a disputable question (I'm appreciating a *lot* what he has done for Open Source in the meantime). Three weeks ago no one on this list had any "kingdom". Stefan and Conor (and others) were directing the Ant 1.2 progress without showing any "royal" attitudes, while a lot of people were discussing the direction Ant 2.0 should take, without anyone dominating. Then James started claiming he had the right to dominate. Unfortunately instead of resolving that dispute directly, people were starting to publish lots of proposals to block James's claims. We still discuss technical arguments on the surface, while the real issue remains unsettled. > I listened, it made sense, it's what I want, so I would > love to see it happen... I think one of the problems is that when Ant started it was a helper tool for the Jakarta projects, but in the meantime a lot of people use it for other purposes successfully. As I see it, one of the most fundamental differences between James's vision for Ant and the vision of some other committers is that in James opinion Ant should focus on Java projects (with support for other uses seen as 'nice to have' at best), while others are heading towards a more general build tool or even task engine (with support for Java projects the most important requirement). > I didn't cry when Giacomo came on > the Cocoon mailing list and took over "my" Cocoon 2.0 (as Stefano didn't do > it when I revolutioned his Cocoon 1.x). Revolutions happen, and if we don't > care about loosing our "prestige" or "leadership" but look at their > technical value, we'll find out that most of them are very valid... If I'm not completely wrong every single committer has agreed to a revolution for Ant 2.0 and this decision was made before James showed up on ant-dev again. Besides, no committer even dared to present his proposal until James checked in AntEater, although they had already done their prototypes. I interpret this as sign of modesty, not of hunting for "prestige". I really can't see Your point here. Honestly, I feel Your statements are more than biased, and You do no one a favour with it, not even James. Instead You have largened the gap between the 'Jakarta VIP group' and the 'Ant committers group', and that is a pity. You seem to take the position 'either James gets control or Ant will fall apart', but the reality has already proven You wrong. James has done great work in inventing and designing Ant, but the current committers have also done great work in making Ant useful for a large count and broad spectrum of users, and they already had started a successful discussion about the next-generation Ant. And while I'm convinced that James can contribute a lot to the future of Ant, I'm also convinced that its future will be even brighter if *all* committers finally manage to join forces. I hope that all contributers will soon start to resolve the status questions instead of hiding behind pseudo-technical arguments, as is happening now. My (biased) opinion is that James would have to take the first step by explicitly giving up his claim to be the decisive voice. I see neither a formal rule nor a strong enough moral obligation that would support this claim. I do hope that this is no 'either I'm the leader or I'll leave' decision for him. The next step should be getting consensus about vision and main requirements. If these steps are taken successfully, I'm optimistic that all those bright brains on ant-dev will be able to find the best design for Ant 2.0 easily. Wolf
