On 12/27/05, Steve Freitas <[EMAIL PROTECTED]> wrote: > Patrick Giagnocavo wrote: > > Some people will want a small footprint, starts-fast implementation, > > bytecoded OK. Others will not care about startup time as much as > > good native code compilation, robust threads, etc. > > > > How to please everyone with one implementation? > > Simple. You don't. > > Surendra Singhi wrote: > > Having choice is not something which is bad. In fact the more the > > better. > > Surendra, there are plenty of situations in which having too much choice > is a bad thing. Because when you decide to do too many of them at once, > you end up doing all of them badly. And that's what the open source Lisp > community, if I can define them that way and that broadly, has done. > > Guys, I need only to quote myself: > > > > Trying to keep too many options open can prevent one from taking > > > advantage of any at all. > > Perfect is the enemy of the good, and right now the former is winning. > > Peter cited Python. Python's quite popular. Yet there's only one real > (read: broadly usable) implementation. It doesn't have native threads. > It doesn't have native compilation. It doesn't do serious recursion. I > want those things. That doesn't stop me from using it in most of my work > projects. Why? Because everybody is unified behind it, so the libraries > are comprehensive, the documentation is plentiful, and across multiple > platforms it Just Works. The result is that it's painless to get started > with, and what always seem like a zillion deployment details that give > me great pause in Lisp don't give me a moment's worry with Python. Can > you honestly say that about any OS Lisp right now? > > People, this is how Lisp ended up completely missing the dynamic > language renaissance of the last ten years in the first place, with > everybody going their own direction, leading to a division of efforts > until the most basic of things -- like providing a decent source of > libraries -- became too much for the community to do. > > If we're gardners, we'll help the garden grow. Do you want a garden made > of a thousand bonsai trees that won't survive the first hailstorm? Or a > tall, strong oak that will survive the winter? Well, the only way to get > the oak is to make sure an implementation Just Works. > > Steve
I haven't tended any bonzai trees in my lifetime, but I do know oaks take quite a while to grow. I don't think all of us unifying behind a particular implementation will do anything but cause a good portion of the gardeners to quit and make their own groups around the implementations they use. I think we should all just do the things that we can reasonably do now, and over time it is possible we will converge on one implementation (and it damn well better be clisp). For now, its better to just find a project you can do, and then do it. _______________________________________________ Gardeners mailing list [email protected] http://www.lispniks.com/mailman/listinfo/gardeners
