Cadol, anyone?
Brent Laminack
(4 to 8 user) system using 8088s and 64k RAM, 10 Meg 14" disk, etc.
It had a particularly noxious language: CADOL. All modules had to
compile into less than 256 bytes, only 8 levels of subroutines, etc.
CADO eventually got bought by Contel systems. Does anybody know how
CADO and CADOL died?
Thanks,
Brent Laminack (br...@cc.gatech.edu)
Zaq
> Anybody remember CADO systems? In the late 70's it was a multiuser
> (4 to 8 user) system using 8088s and 64k RAM, 10 Meg 14" disk, etc.
> It had a particularly noxious language: CADOL. All modules had to
> compile into less than 256 bytes, only 8 levels of subroutines, etc.
> CADO eventually got bought by Contel systems. Does anybody know how
> CADO and CADOL died?
We use a successor to it. I believe a company called VersysS now owns
it. The guy who showed me everything I know about it, now works
(again) for a company called Benchmark Systems out of Lynchburg VA, he
can go through the entire history of it.
Anyway, when I first got hear about 5 years ago, it was being emulated
under AT&T unix and has since been upgraded to AIX.
The version we have still has the same old Basic language with 20-some
odd numeric registers, 240 bytes of alpha registers (spread across
A-C3), some flag and global registers and 5 256 byte buffers.
I believe the latest version has added a number of registers (more
alpha, numeric and a few more buffers), but we don't have that
version. From what I heard, Benchmark usually avoids the new
registers if possible for backwards compatibility (why have them?).
It has been expanded in a number of areas, for instance, we have a 64
user system. It also now supports multiple segments per program,
effectively eliminating the 256 byte program limit.
I am really not sure about the 8 levels of subroutines, but since I
never ran into that problem, I guess that it has been eliminated.
We also have an old CADOS machine for development, as the name
implies, it runs under dos,... man does it compile SLOW.
I just jumped over to www.versyss.com, but they don't have any good
info on the product on the web page, and I just don't remember
benchmark systems web address off the top of my head.
There are a lot more things you can say about it...
The 256 byte buffers are also the max size of a data record. Data
records are variable length, and fields are not named, you have to
know that date of birth is '4 alpha skips, 15 numeric bytes, 2 more
alpha skips and that it is stored as an alpha'... code would look
something like this (it has been a while)
READ 5
INIT R
SKIP A 4 ! Skip 4 variable length alpha fields
SKIP N 15 ! skip 15 bytes
SKIP A 2 ! 2 more alphas
A = R(A) ! read the alpha and store it into A
What else... There are NO local variables, no recursion, no floating
points, you had 10 byte (I think) numerics registered where you had to
emulate decimals on the fly with multiplication and division. Also,
there is no (usable) indexing system.
Oh yeah, their method for terminating strings. Half the ascii chart
was non-terminated strings (128-256) and the other half was
terminated. So if you hex view a datafile, everything EXCEPT the
final terminating character is in upper ascii.. highly unreadable, I
think I like C/C++'s null terminated strings better.
... but, it was simple...