> When you're selecting a library to solve a particular problem, you > normally have to do some research and evaluate more than one library > so, for me, the activity of the project and software versions > supported are part of that necessary research. I can't imagine "just > using" some random library without doing some legwork...?
I appreciate what you're saying, and I think it is true to some extent, but on the other hand, I think in practice many people experiment with libraries to see if they work, and use the libraries without being absolutely sure that they work perfectly. You may think I'm doing it wrong, but I don't think I'm alone at all. I feel that Flaubert's quote really reflects the situation. For example, I recently used the Brent function minimization algorithm in commons math (which admittedly has a decent reputation). I didn't review the source code of the implementation carefully -- it's too complicated -- but as far as I can tell, the code works well for my application, and that part of my app passes various tests. The risk of bugs from that library is nonzero, but acceptable to me. For clojure to break backwards compatibility is levying an extra cost beyond these usual risks. I agree, as Andreas suggested, that there is a cost to the speed of innovation in the language if one maintains backwards compatibility, but on the other hand I think breaking backwards compatibility imposes a cost to the community and slows down library innovation. We are comparing the costs to two different groups -- the language developers and the users, and I am a user whining about the pain. I just hope Clojure's re-birthing pains will be minimized in the future. ;) > Phil's suggestion of adding metadata to Clojars might help Yes, I think it will be very helpful to mark which jars are compatible with 1.2 or 1.3 -- I suppose this can be extracted from the dependencies vector in the project.clj file. And Stu's suggestion of detect-and-warn or detect-and-fail are also very valuable, since right now the situation is that mysterious errors occur without warning. Arthur -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en
