jhc@mtung.UUCP (Jonathan Clark) (06/12/85)
<> Please God save us from any more high-level assemblers, like SMAL, or PL/Z, or anything else (not macro assemblers, though. They are very useful - in the right context). My immediate reaction when faced with a SMAL program is to get the assembler output and then convert that into 'real' assembler code. SMAL just gets in the way of the underlying machine - although you can write 'high-level' you still have to know what's going on underneath (for example, you have to know what instructions the SMAL is translated into so that you can jump on the correct flag (carry, overflow, etc)). Personally I think this is bad, and that a C and assembler mix is a reasonable compromise between development time, software quality, speed and codesize, and so on. Another favourite moan while I'm on the subject - I do hate writing in (cross-) assembler under UNIX because the opcodes and assembler format are always different from the manufacturers. Anyone know whether this was done to avoid copyright problems, to keep all the assemblers looking the same, or just to discourage people from writing in low-level languages? -- Jonathan Clark [NAC]!mtung!jhc
oz@yetti.UUCP (Ozan Yigit) (06/19/85)
In article <573@mtung.UUCP> jhc@mtung.UUCP (966-Jonathan Clark) writes: > >Please God save us from any more high-level assemblers, like >SMAL, or PL/Z, or anything else (not macro assemblers, >though. They are very useful - in the right context). > Uhm.. The first thing a macro hacker does is to write a package of structured macros. There is about ten of them for PDP-11s. I have seen about three on VAX/VMS. That is getting *high level*, wouldn't you say ??? > >My immediate reaction when faced with a SMAL program is to >get the assembler output and then convert that into 'real' >assembler code. SMAL just gets in the way of the underlying >machine - although you can write 'high-level' you still >have to know what's going on underneath (for example, you >have to know what instructions the SMAL is translated into >so that you can jump on the correct flag (carry, overflow, >etc)). Are we talking about SMAL/80 ??? It has 1-to-1 correspondance with the underlying machine, (Z80 in this case), and it generates what it means. LOOP and IF-THEN structures are straight forward, and no *secret* code is generated for them. Even if the conveniences of LOOP and IF-THEN constructs are not available, it seems to me that smal/80 style is better than the ancient practice of creating assemblers with barely understandable 3-5 character mnemonics. Naturally, 3-5 character mnemonics are supposed to be "more efficient" to write for the programmer. Think of how much time one spends in *commenting* each and every line. (sigh!) Oz (whizzard of something or another, no doubt..) Usenet: [decvax|ihnp4|allegra|linus]!utzoo!yetti!oz Bitnet: oz@yuleo | oz@yuyetti
jhc@mtung.UUCP (Jonathan Clark) (06/21/85)
<> Oops. I was talking about an internal AT&T language called SMAL/80, which is not and never will be a product. That does have all the problems I was moaning about earlier. This other SMAL/80 I have no experience with. -- Jonathan Clark [NAC]!mtung!jhc