walt2@ihuxl.UUCP (Walt Kurszewski) (10/14/86)
Fellow Netlanders, I'm about to drop $1045 on a 1040ST with a color monitor. Being a conscientious consumer (tight wad), I've got a few questions. I know the following topics have been kicked around this group in previous months, so feel free to mail your replies to me. What's in the developers kit? Tons of (useful?) documentation, a C compiler, an assembler, a linker. Does it include a communication package to upload and download files from the UNIX system at work? How buggy/fast is the C compiler that comes with the developers kit? 300 frogskins is quite a bit, would I get my monies worth with this? My other options for a C compiler are Lattice (slow, but I can get it for ~$125) or Megamax (fast but I don't know how buggy or how much it costs). I'll need an assembler that I can link with the C compiler. Is there any public domain C/assembler/linker? I also need a communication package. From what I understand, a vt100 emulator has communication capabilities that will work on the UNIX systems at work. Does a public domain vt100 emulator exist? How do I get my paws on it if I can't upload/download? What are good/bad documentation sources? Books? Magazines? I'm new to the PC world and I'm pretty much lost. If any experienced 1040ST person would be willing to mentor me via e-mail I'd appreciate it. Thanks. Walt Kurszewski ihuxl!walt2
appelbau@topaz.RUTGERS.EDU (Marc L. Appelbaum) (10/14/86)
For up/downloading with a UNIX you could use KERMIT, I don't know if it is included with the kit or not. If not you could get from a local user group. To ul/dl with kermit is fairly simple. At the UNIX end you type kermit to get into kermit. To send files type send filename, to recieve files type get files. It's basicly the same on the ST end. -- Marc L. Appelbaum Arpa: appelbau@blue.RUTGERS.EDU Uucp: ...{allegra| harvard| seismo| sri-iu| ut-sally}!topaz!ru-blue!appelbau
jons@islenet.UUCP (Jonathan Spangler) (10/17/86)
In article <1978@ihuxl.UUCP> walt2@ihuxl.UUCP (Walt Kurszewski) writes: >My other options for a C compiler are Lattice (slow, but I can get it for ~$125) >or Megamax (fast but I don't know how buggy or how much it costs). I'll need >an assembler that I can link with the C compiler. Is there any public domain >C/assembler/linker? > Saw an add for the Mark Williams C for the ST. Sells for $179.95 and includes MicroEMACS, 600 pages of docs, assembler, linker and archiver, and the Microshell Command Processor (is this Beckmeyer Tools?) Also: complete K/R "C" + extensions, 8 register variables, full AES & VDI libraries, UNIX-compatible libraries for portability, 300 Atari-specific routines, one-step linking & compiling with cc command, English error messages link-link error checking. Includes utilites: egrep,sort,diff,cmp,pr,tail,uniq,wc and more. Don't have it, but it looks really good. PLEASE do not ask me anything else about this -- I am just passing information along. Aloha, Jonathan Spangler {ihnp4,vortex,dual}!islenet!jons
braner@batcomputer.TN.CORNELL.EDU (braner) (10/24/86)
[] From what I've heard, Mark Williams C is great. The remaining problems are supposed to be solved in the next release. BUT: (THE CATCH!!!) You HAVE to have a hard disk to make any serious use of it!!! It is FAR too slow on floppies (to my taste). (And you HAVE to have TWO DOUBLE-SIDED floppies if you want to go that way...) It is too slow on a hard disk, too! So in my opinion Megamax is still the only viable option for "personal" (low-budget) use, and is superior to MWC in speed of compilation which is important to all users. Consider: Megamax is so compact I put the compiler, linker, libraries, microEMACS, utilities (from micro-C-Shell) in a (576K) RAM disk, and still have some 250K free RAM disk space left! In the rest of my 1040's RAM I have micro-C-Shell resident, and enough space to run all those things one at a time. Compiling and linking (all the way to a .prg file) a small (2-page) program takes about 15 seconds, all done in RAM. (MWC on a hard disk is an order of magnitude slower.) I can put the source for microEMACS (180K) on the RAM disk, and compile it completely in RAM! Now if only somebody would figure out how to make Megamax and I/O redirection from within Micro-C-Shell work together... - Moshe Braner
iarocci@eneevax.UUCP (Bill Dorsey) (10/26/86)
In article <1277@batcomputer.TN.CORNELL.EDU> braner@batcomputer.UUCP (braner) writes: >[] > >From what I've heard, Mark Williams C is great. The remaining >problems are supposed to be solved in the next release. > >BUT: (THE CATCH!!!) You HAVE to have a hard disk to make any >serious use of it!!! It is FAR too slow on floppies (to my taste). >(And you HAVE to have TWO DOUBLE-SIDED floppies if you want to go >that way...) It is too slow on a hard disk, too! > Depends on what you mean by slow. Mark Williams C is about two and a half time faster than Alcyon C. While not nearly as fast as Megamax C, at least it doesn't resrict you to 32k code segments! You DO NOT have to have two double-sided floppies. Mark Williams has released a version that runs on two single-sided floppies. And if you have a 1040ST or a 520ST w/ 1 meg, you only need one drive (either single-sided or double sided) to use the compiler. To do this, you put the compiler disk in the floppy and put your source code on a ramdisk along with whatever other programs you want that will fit. >So in my opinion Megamax is still the only viable option for >"personal" (low-budget) use, and is superior to MWC in speed >of compilation which is important to all users. > Speaking of low-budget use, Mark Williams C lists for $20 less than Megamax C. And Mark Williams Company gives dealers a larger margin on their product than Megamax; I bought Mark Williams C for $130 (list=$180) whereas I have yet to see Megamax for less than $190 (list=$199). >Consider: Megamax is so compact I put the compiler, linker, libraries, >microEMACS, utilities (from micro-C-Shell) in a (576K) RAM disk, >and still have some 250K free RAM disk space left! In the rest of my >1040's RAM I have micro-C-Shell resident, and enough space to run >all those things one at a time. Compiling and linking (all the way to a >.prg file) a small (2-page) program takes about 15 seconds, all done in RAM. >(MWC on a hard disk is an order of magnitude slower.) I can >put the source for microEMACS (180K) on the RAM disk, and compile >it completely in RAM! > >Now if only somebody would figure out how to make Megamax and I/O redirection >from within Micro-C-Shell work together... > >- Moshe Braner Nice, but with Mark Williams C you get a powerful Make utility, a small but useful shell program, and an assortment of unix-like utilities to round off the package into a powerful full-featured development system. I'm not disputing that the Megamax C compiler is a good compiler, but I do disagree with Moshe's conclusion that Mark Williams C is only good for developers. If compiling speed isn't the only consideration you buy a C compiler for, then I would highly recommend Mark Williams C compiler. -- ------------------------------------------------------------------------------- | Bill Dorsey | | 'Imagination is more important than knowledge.' | | - Albert Einstein | | ARPA : iarocci@eneevax.umd.edu | | UUCP : [seismo,allegra,rlgvax]!umcp-cs!eneevax!iarocci | -------------------------------------------------------------------------------
SOFPJF@UOGUELPH.BITNET (Peter Jaspers-Fayer) (10/30/86)
OK, but which C is most Alcyon compatible? I tried to convert something A->MM (the details are a bit fuzz now) having to do with trying to move characters with a pointer into memory based on an offset and some index in a loop. The problem was that Megamax complained that the pointer was not the same type as the long int that was the calculated address. Alcyon accepted this no problems. This was probably 'lint', but I couldn't think of any other way to re-code it. Ideas anyone? I dislike Alcyon for several reasons among which are: 1) Multi-passes, multi little bits and pieces contribute a lot to its slowness 2) The links are a MESS, in Megamax, you just point to the .O files you want linked together (yours not system stuff) and it goes. None of this fiddling with multitudinous libraries and worrying about what order... But unfortunately, a LOT of good software around is still written in it. The .O files are not compatible either! That's not right. I wish there was a standard ST linker, and that PASCAL, MODULA, and WHATever else all produced the same type of object files, which could then be linked together. I know that there's some different parameter passing conventions there, but Am I asking too much? Why not? So what about Mark Williams' and other Cs with respect to source compatability? (The object file compatability is just a dream with multi vendors *sigh*) /PJ