Date: Fri, 31 Oct 1997 09:54:38 +1300 From: physmsa AT cantua DOT canterbury DOT ac DOT nz (Mr M S Aitchison) Subject: Re: thoughts and questions. To: opendos AT delorie DOT com Message-id: <199710302054.JAA22416@cantua.canterbury.ac.nz> Precedence: bulk > From: sb1 AT sisna DOT com > > Marc Perkel wrote some ideas about the future of DOS back in 1991. > He basically said make an int that would tell the system when called to > run the new dos features. Then create API that would let us have an > open architecture with emulators for other os etc. > Also if the int was not called to inable the new extentions it would be > just like good old dos. That sounds reasonable. Of course a lot of people will want just a good DOS (perhaps because they run on lean hardware), and that is fine... OpenDOS already has some good extensions (which could be further developed) which don't take up lots of RAM the way they do in Win95 or Unix. But to go much beyond traditional DOS we need to start with a clean sheet of paper and come up with a new API (even if we end up borrowing from Java etc). Win95 is one way to mesh old and new; Linux with Dosemu is another (and, I think, better in concept as well as implementation). > Hopefully we can get more of the power of unix with the ease of use like > dos. It amazes me why we haven't seen a really new text interface (command.com, "CLI" or "shell" in Unix-speak).. they are all ports or modifications of old ones (command.com came from CP/M, which was heavily influenced by old PDP11 stuff; bash, ksh, tcsh all trace back to old Unix shells, and back to Multics. Not only is there a need for a new nice command language (that makes it easy to use voice input, and other new technology) the very act of taking an old standard and modifying it means you loose standardisation. I don't particularly like old Unix idea of abbreviating commands to two letters (mv for move, cp for copy, ls for list) - I prefer to have sensible names and let the user alias the ones they often need, but it is worse having *some* commands heavily abbreviated and others using long names with different Capitalisation depending on what was in fashion at the time of writing. > Will we ever get to run windows programs from the DOS prompt? I guess it is up to Wine, Wabi, bochs, twin to do this part. I guess we could start on a sister project called "OpenWIN" but Sun beat us to the name! OS/2 (v5) is likely to be the best way to go if you want to run Windows programs; I think a good policy stance is to run a thin X11-compatable client alongside OpenDOS... with an applications server elsewhere on the net. > Will we ever be able to use the win .dll files etc? > What about winsock for the net? There have been very few new programs, especially in networking, that use the old DOS-based standards... it would be nice to take advantage of new program development work. Shared libraries and DLLs have a big attraction, but often they are part of huge apps that would be unattarctive to most OpenDOS users. I firmly believe the answer is to create a compiler system that generates programs easily (=can get a 1Mb zip file for free and start creating programs at home after 15 minutes reading) that work natively/quickly under OpenDOS but also work under Win 3/95/NT, OS/2 and of course under Linux within Dosemu. I am tempted to say this should involve Java somehow. I think the world still needs a good popular compiler, not Visual BASIC or Delphi or C++ but something as small, easy and importantly: respected as Turbo Pascal used to be. Have it automatically create programs that run really well under OpenDOS (and pretty well under other operating systems - in the same way that a program might detect if you have a numeric copro and emulate it if you don't). Coupled with this, Caldera could pay the authors of the best editor, mail program, etc created using this (so long as they are a superset of the previous distribution's standard) to be included in each revision of the standard commercial OpenDOS distribution. ------------------------------------------------------------------------- Mark Aitchison, \_ Phone: +64 3 364-2947 home 337-1225 Dept of Physics & Astronomy,