Phil Dawes wrote: > One of the things I'm loving about factor is its ability to change > implementation direction on a sixpence. I don't know how long this can > last, but for now it's a massive plus for the language. This came up on the IRC channel today. Factor is changing quickly and a lot of code has to be updated because I prefer to get the language design right before we start seeing many more users. Right now I estimate about 10-20 people use Factor on a regular basis, and while we still have a small core group of users we need to perfect the language design and get the heavy lifting of the major features out of the way. This also means that these people who do use Factor need to give lots of feedback about what they like and what they dislike --because if you dislike something and don't speak up, chances are it will get baked in permanently!
My agenda includes adding new-slots, multi-methods and inheritance to the core. After that I think it will be time to start focusing on performance, stability, and peripheral libraries. Ed has proposed some stack effect changes to core collection words and Dan is toying with generic collection protocols but I'm still undecided as to whether these ideas are worthwhile, and if adding them now will postpone 1.0 too far. Perhaps both can be merged into one proposal: a generic collections vocabulary implementing new stack effects. This would go in extra/ and people would be free to use it but the core code would not change and existing code using core sequence and collection words wouldn't need updates either. Ed and Dan, what do you think? Personally I like improving the language but I don't get too excited about updating lots of my own code either. So at some point in the near future, we have to say that enough is enough and focus on getting 1.0 out of the door. > The Gambit > Scheme folks are missing a big trick trying to move to a cpan style > library system too early rather than just encouraging everybody to dump > their code into the distribution and managing via dvcs. A package manager will definitely be needed at some point. I would greatly prefer if core libraries such as collections, parsers, web stuff, multimedia and so on was part of the core repository, however people will want to start building applications at some point and shipping them all with the distribution will start to be unwieldy, both due to size concerns and mismatches between the Factor development process and how other people want to manage their own projects. Slava ------------------------------------------------------------------------- This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Factor-talk mailing list Factor-talk@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/factor-talk
