Bill Rizzi has hit upon an interesting point. Has anyone moved from
TOPS-20 to UNIX and liked it? To be specific- does anyone believe he/she
is more productive working under UNIX after using TOPS-20?
-------
Sure, I have. I used TOPS-20, and before that TOPS-10, for many years
before beginning with UNIX. I can see that anyone who thinks that
command completion and online help is the principal measure of an
operating system or programming environment would love TOPS-20. On
the other hand, anyone who really uses the system, system calls and all,
could get real tired of TOPS-20 and BLISS. I did.
Indeed, it was the severe shortcomings of TOPS-20 as a serious programming
environment that motivated a group at Yale to invest what seem like man-
centuries in developing the Yale Tools package.
If you want to see the difference, try writing your favorite UNIX utilities
in BLISS or MACRO-10 using TOPS-20 system calls and (echh..) byte pointers.
Don't forget 5 7-bit chars per word, left justified, and be sure to include
the precious command completion and "?" functions.
Talk's cheap....
John G. Zornig
Ye Olde Venerable
BBN Communications Corp.
I worked for a couple of years on TOPS-20 and its predecessor
TENEX before moving to UNIX. I found UNIX to lead to a large
increase in my productivity almost immediately. The big thing
that TENEX lacked that UNIX has is pipes and all the programs on
UNIX know how to use them--I spent a long time trying to teach a
large application program I was using to take it's input from
another process (since that time the application has been moved
off of TENEX and onto UNIX), and never really succeeded. That
comes for free on UNIX.
Of course, we cheated at BBN and rewrote the terminal driver to
allow wakeup-sets and stop-at-the-bottom-of-the-page mode,
CRT-oriented erasing, etc. (this was back in the days of V6).
(Pipe everything through MORE? gag. Build MORE into readnews and
news and the mail-reading program? double-gag.)
When my TOPS-20/TENEX friends ask me why I prefer UNIX, I tell
them about a 25-LINE LONG SHELL file I once wrote in about an
hour (I made lots of refinements) which goes and gets the past
few day's messages off of MIT-AI's bulletin board (running FTP
from a pipe, using AWK to generate FTP's command-input) and puts
them into uniquely-name files in a scratch directory, so I can
look at them at my leisure.
How long do you think it would take to write a similar program on
TOPS-20? I'd guess probably two or three days of very heavy
MACRO-10 hacking...Of course you'd probably never undertake the
task in the first place since it would seem so daunting.
But it's true that there are lots of lessons that UNIX could
learn about paging, process-control, user-interface,
documentations, etc., etc., from TENEX, and MULTICS. I was
struck by the initial responses to Scott Cooper's (?) article
("How long are we going to put up with this crap?").
The responses I mean are the ones of the "Oh, yeah? Well name an
operating system that's better!"
About the ONLY operating system I've ever used with a worse user
interface and which was more poorly-documented than UNIX was
MIT's ITS system, which was abysmal in both those respects. Come
on, folks, is UNIX the first "commercial-grade" operating system
you've ever used? (ITS is by no means commercial-grade.)
There are some really neat ideas out there in OS land. Don't
think UNIX has it all. Look around a little.
Yes.
To be specific- does anyone believe he/she is more productive working
under UNIX after using TOPS-20?
Yes.
I work for a few years at school on TOPS-10. When we got a V6 UNIX system
circa 1975 and put it up on an 11/34 I was blown away. It was far more
powerful and easy to use than this 1 million dollar DEC-10. Anytime
I have to use a -10 or -20 it's a royal pain.
The only thing I missed was command/filename recognition and completion...
so I implimented it (in a better manner than tenex or tops).
I've posted same to net.sources.
Ken Greer
--
Ken Greer
I have, I still love lots of the things TOPS-20 does that UNIX does not,
like command completion, and file name completion, but there is no reason
that if someone felt like it, unix could not do the same. the only thing
that unix could not do, is deal with '?' in the same manner, for, reasons
I need not mention to you folks.
But TOPS-20 was written for a different use I beleive. And it is my
opinion that BOTH are winning. But that UNIX SEEMS to do lots more then
tops-20. (but, when you look close, that works both ways, but for god's sake,
they are VERY different Operating Systems).
Yours In Hacking(c),
-Shawn
CopyRight 1983, All Rights Reserved,
Sundar Iyengar
sundar.Case@UDel-Relay
decvax!cwruecmp!sundar
I would however like to point out that this discussion would be more
appropriate on net.cog-eng than net.unix-wizards. I thus recommend that
all further discussion on this topic be confined to cog-eng.
John Diamant
Usenet: ...decvax!cwruecmp!diamant
CSNet: diamant@Case
ARPA: diamant.Case@UDel-Relay
/Joe Kelsey
May I remind you that not everyone who is on Unix-Wizards has access
to the Usenet and/or its mailing lists. I, for one, am finding much
of this discussion pertinent to my applications.
--Lynn
-------