just because we have enough power available to be sloppy with
our programming doesn't mean we should do so...

Of course we shouldn't. Don't agree with everything in hutches
text though. I made my own statements in the article at my site
( http://f0dder.has.it , articles, "packing, data handling and stuff" .)
I don't believe that to advocate sloppy coding... quite on the contrary.
I suggest people to read both hutches article and mine, and make
up their own minds. Tired of flamewars.
Posted on 2002-02-25 22:09:50 by f0dder
I read yours first f0dder... my decision is quite concise ;)
Posted on 2002-02-25 22:15:25 by SpooK

Tired of flamewars.
Then put away your lighter, friend.
There is enough warmth here for all. :)
Posted on 2002-02-25 22:31:33 by bitRAKE
Nah... keep that lighter out and start singing "Give Peace a Chance" ;)
Posted on 2002-02-25 23:14:34 by SpooK
Well, being one of those that has been programming on more platforms, in more languages than I can remember, for the past 3+ decades, I still find myself coming back to things like, well, things like movsd. I wonder why that is. :grin:

Maverick,

I've been involved in a couple of "home grown" languages. One was a terminal control language, similar to CICS on the mainframe. The other a composition, or typesetting language, also mainframe based. Both were full featured languages, with verbs like IF, MOVE, DIVIDE, PERFORM, READ, etc. The composition language had specific verbs, like LEAD, FONT, and SETLINE. Same thing for the terminal system, with verbs specific to the application.

Funny thing is that they both started life as macro based assembly language on the mainframe. The macro processor on the S/390, ever since the old days of the S/360, is much more powerful than that provided by MASM. We wrote a MOVE macro. We wrote an IF macro. They did all the necessary data conversion, sometimes generating inline code, sometimes calling a proc for complicated functions. Speaking of data conversion, we invented a new data type for the terminal system. It was a compressed date, that packed a binary YYMMDD date into a 16 bit field. When it was moved to a character field, for example, we would "unpack" it and convert it back to decimal digits. A side benefit, our Y2K problem was minor. :cool:

The terminal system never made it off the mainframe. But the composition language was ported to the PC. We wrote a compiler, that read the mainframe "assembly" source and generated x86 object code. It really wasn't "assembly" source any longer, it was it's own language. Programs were written without a single line of "native" S/390 assembly code. It looks more like COBOL than assembly.

Today, it's still our primary composition language. We still enhance it. When we do, we start on the mainframe, maybe with a new macro, maybe updates to existing macros, usually a combination.of both. Once we beat it up on the big box, then we update the compiler on the PC.

I get a kick out of people that say assembly language isn't well suited for large applications. They must have never seen a mainframe! Take a look at OS/390, once describes as the most complicated software ever written. 100% assembly language. Or look at CICS. Sure, it has a fancy front-end for COBOL, and another fancy front-end for PL/1, and another for C, but the guts, 100% assembly language.

Now, why was I here again... ... ...

:)
Posted on 2002-02-25 23:31:48 by S/390

Funny thing is that they both started life as macro based assembly language on the mainframe. The macro processor on the S/390, ever since the old days of the S/360, is much more powerful than that provided by MASM. We wrote a MOVE macro. We wrote an IF macro. They did all the necessary data conversion, sometimes generating inline code, sometimes calling a proc for complicated functions.
That's the same genesi of my HLA (High Level Assembly) for Amiga.. started as a set of MACROs for AsmOne (absolutely great assembler + editor + source level debugger.. now if the PC only had such a tool..), but then I preferred to write a true compiler (first in Amos :grin: then rewritten in HLA itself). The PC language instead is following a different road.. it even changed name, I'm not going to write it though because otherwise I risk that Randall uses this name too (JOKING!!) :grin:

Greets,
Maverick
Posted on 2002-02-26 03:34:38 by Maverick