On Fri, Oct 2, 2009 at 10:54 AM, John A. De Goes <j...@n-brain.net> wrote:
> On Oct 1, 2009, at 12:13 AM, Curt Sampson wrote: > > And as far as something like dealing with a changing language and >> libraries, the mainstream already has well-established and popular >> techniques for doing just: agile development. >> > > A project manager's worst nightmare: > > "Sorry boss, but we're just not going to be able to meet that deadline, > because, well, a language extension we were using was dropped from the > language, and the syntax for some core operators was changed. Not only is > our code broken, but many of the libraries we were using are broken. Don't > worry, though, we're 'agile', so our unit tests will tell us when our code > is working again." > While I agree that it probably isn't the right idea to say that we are Agile, so it is safe for us to build on a foundation that is constantly shifting underneath us, this same argument came up from Credit Suisse regarding the standardization of Haskell' at ICFP 06. At the time, as I recall, they were limiting themselves to Haskell 98 + Addenda. I argued that they should be interesting in having _more_ such standardization efforts to bring into the fold more features that they can rely upon. Even so, Haskell' includes only one breaking change (dropping n+k patterns) at this time and really how often do language features get dropped from GHC? (And at the time even the n+k change was laughed at by the audience as a joke proposal, not one that anyone was serious about -- how times have changed). There have been a couple of quirky changes in how big scary types involving scoped type variables change. Otherwise it has been far more stable and consistent over the last 11 year run than any non-toy compiler that *I* can think of. Heck, think how different your C compiler is now than it was in 1999. It feels, to me that there are more breaking changes in just upgrading to, say, C99 than there have been over the entire post-Haskell 98 life of GHC. -Edward Kmett Big business demands stability. > > Regards, > > John > > > > _______________________________________________ > Haskell-Cafe mailing list > Haskell-Cafe@haskell.org > http://www.haskell.org/mailman/listinfo/haskell-cafe >
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe