I met Steve Bourne a while back and asked him why he chose to write sh with all those ALGOL-like macros. His answer was terse—he just wanted to use the syntax he was most familiar with.
I did that kind of thing when I first started using C (my attitude then was that C was fine, but it should have looked more like BCPL). I stopped after a short time.
FrankWilhoit 3 days ago [-]
C for people who don't like C. At CompuServe, much of the legacy code base was in BLISS and so there was a header that said
#define BEGIN {
#define END }
etc. etc. etc.
azinman2 20 hours ago [-]
First time I’ve heard of BLISS. Did you work there? What was used after bliss? Is there anything non-legacy about compuserve?
lproven 8 hours ago [-]
BLISS is not a place, it is (was) a DEC programming language:
For all that DEC is nearly forgotten now, it had far more effect on modern computing than IBM ever did.
DEC OSes are why CP/M and MS-DOS looked and worked the way they did. Even DEC GUIs: OS/2 Presentation Manager fed back into Motif, which lent VMS DECwindows and DEC Ultrix their look and feel, and all this influenced Windows 3.
DEC VMS was a huge influence on Windows NT.
The hardware too: the DEC PDP-7 and PDP-11 shaped Unix, and much of the Unix and Linux design is down to DEC hardware influences.
How PCs look and feel and work is because of DEC, not IBM. Nothing about the PC is at all IBM-like; IBM's OSes are profoundly different, from the 5100 and APL all the way up to modern Z Series mainframes.
Regarding CI$, many of the things that were odd about CompuServe, even the user account names, are because it ran on PDP-10 kit:
I did work there. BLISS was originally an academic project (at Carnegie-Mellon) whose point of departure was this. Suppose a compiler to emit machine code comparable, for size and speed, with the best hand-tuned assembler: then what kind of high-level language could be the input to that compiler? What features must it [not] have? The findings were documented in a book, The Design of an Optimizing Compiler by William Wulf et al., published 1971.
keybored 1 days ago [-]
> #define LOOP for(;;){
Predicted Rust.
Rendered at 19:55:36 GMT+0000 (Coordinated Universal Time) with Vercel.
Setting my User-Agent to match my Firefox instance fixes it.
https://news.ycombinator.com/item?id=45800777
#define BEGIN { #define END }
etc. etc. etc.
https://en.wikipedia.org/wiki/BLISS
For all that DEC is nearly forgotten now, it had far more effect on modern computing than IBM ever did.
DEC OSes are why CP/M and MS-DOS looked and worked the way they did. Even DEC GUIs: OS/2 Presentation Manager fed back into Motif, which lent VMS DECwindows and DEC Ultrix their look and feel, and all this influenced Windows 3.
DEC VMS was a huge influence on Windows NT.
The hardware too: the DEC PDP-7 and PDP-11 shaped Unix, and much of the Unix and Linux design is down to DEC hardware influences.
How PCs look and feel and work is because of DEC, not IBM. Nothing about the PC is at all IBM-like; IBM's OSes are profoundly different, from the 5100 and APL all the way up to modern Z Series mainframes.
Regarding CI$, many of the things that were odd about CompuServe, even the user account names, are because it ran on PDP-10 kit:
https://www.inwap.com/pdp10/compuserve.txt
Predicted Rust.