mikel@Apple.COM (Mikel Evins) (01/17/91)
Archive-name: languages/scheme/survey/1991-01-16 Archive-directory: altdorf.ai.mit.edu:/archive/scheme-7.1/ [18.43.0.246] Original-posting-by: mikel@Apple.COM (Mikel Evins) Original-subject: Re: Looking for Scheme for the NeXT Reposted-by: emv@ox.com (Edward Vielmetti) In article <1991Jan16.204221.23706@grape.ecs.clarkson.edu> anthonjw@clutx.clarkson.edu (Jason W. Anthony) writes: >Could someone point me to a version of Scheme for the NeXT? Either Public >Domain or Commericial would be great. > >I understand there is a language called "T" that is available for the NeXT >and is similar to Scheme. Is it much different? I would be using it for a >class, so a lot of differences might be inconvient. There are several ways to go. You could use T, but T is not quite Scheme. There is a good chance that you would have to expend significant effort in modifying your code for compatibility. T can be had from wheaties.ai.mit.edu. You could use CScheme, also from MIT. CScheme is a portable implementation of Scheme with about a million features. I don't know whether anyone has a binary for the NeXT, or the diffs necessary to compile it, but I don't imagine it would be too difficult. I think you can get CScheme from wheaties as well. You could use ELK. ELK (the Extension Language Kit) is another portable Scheme interpreter, this one designed as an extension language for other applications. It can be compiled as a stand-alone command-line interpreter, however. I ran into an incompatability in the current sources that halted my porting effort. Oliver Laumann, who wrote ELK, reports that he is doing the port himself, so I decided to give up for now and let the expert do it. ELK is available on uunet.uu.net. Oliver Laumann can be reached at net@tub.cs.tu-berlin.de. You could use Oaklisp, if you can find an ftp address for it, and if it can be compiled on the NeXT. It is an object-oriented language designed to be fully R^3 Scheme compatible. I don't know whether anyone has it working on the NeXT, but it is written in C, so the port should not be impossible. It from Carnegie Mellon, but I don't know an ftp site for it. You could use XScheme. Same caveats as ELK and Oaklisp. XScheme is pretty clean code, pretty well documented, so a port shouldn't be too hard. It includes some object-oriented stuff too. You can get XScheme sources from bix, or, I think, from Compuserve. You could buy Chez Scheme. This is areal product on some platforms, noatbly on the VAX. It was designed by Kent Dybvig. I am afraid I have lost my contact information for Kent Dybvig, more's the pity. Chez Scheme has an excellent reputation. They have a version running, but unsupported, on the NeXT. It doesn't support NeXTStep. It costs about $1000. Hopefully, someone else on the net will know the contact info for Kent Dybvig. I believe he's at the University of Indiana. You could wait for me to get Zen done. It's a Scheme programming environment and authoring system with some HyperCard-like characteristics. It is also not even close to done, so don't hold your breath. When I get it done, though, I plan to make it available as shareware. I may substitute ELK for my evaluator if Oliver Laumann gets his port done before I get my feature set fleshed out. Incidentally, I plan to support NeXTStep extensively, and I plan for Zen to be able to generate stand-alone apps (programs that people can run without having a copy of the Zen authoring environment).