Boston Linux & Unix (BLU) Home | Calendar | Mail Lists | List Archives | Desktop SIG | Hardware Hacking SIG
Wiki | Flickr | PicasaWeb | Video | Maps & Directions | Installfests | Keysignings
Linux Cafe | Meeting Notes | Blog | Linux Links | Bling | About BLU

BLU Discuss list archive


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Launchpad to be free



I too remember VM370/CMS. I was the data Center Manager at Church's=20
Fried Chicken in San Antonio, and we installed VM370 with OS/VS1 as the=20
guest production OS and of course CMS. And, BTW, CMS was originally the=20
/Cambridge Monitor System, developed at IBM's Cambridge Scientific=20
Center with MIT. IBM closed that lab 1992. One thing that I liked about=20
this is that we migrated from Burroughs that had a much different system =

so we virtualized tape drives where we had some jobs that used more=20
drives than we had. Drive allocation on IBM was done at job start where=20
Burroughs was done when the drive was opened. I personally benchmarked=20
that OS/VS1 actually ran faster under VM370 than it did native. One=20
reason was that the VS1 page size was 2K and VM370 was 4K which made a=20
difference on the 370 hardware we had. Additionally, double spooling=20
produced better throughput. The code was all in IBM assembler, and made=20
for interesting reading since my wife did not move up to San Antonio=20
until we sold our house. I also know of other cases where production a=20
OS ran better throughput under VM370 than native. Also note that we=20
installed a number of patches to both VM370 and to OS/VS1. One of these=20
allowed a CMS user to start a compile on the production OS.

/On 02/28/2009 10:25 PM, Jack Coats wrote:
> I remember in the '70s and '80s running IBMs VM operating system.  We h=
ad
> a source code license.  I even took Amdahl classes about it, where we d=
id go
> through the source code to understand EXACTLY how things worked.  It wa=
sn't
> open source, but the source was available for customers as part of
> their rental/lease
> price.  This included all their OS related stuff, including code that
> wrote channel
> programs, emulated hardware (whether it was there or not), etc etc.
> Mainly in assembler, rather than PL/S or something else.
>
> Some of the code was ugly, but pretty readable.  You also learned to be=
lieve the
> code, not the comments. ... I did especially like the full page "This
> space for rent."
> signs in the code. ;)
>
> In those days for interactive use we used CMS as a single user OS to
> run under VM
> in a virtual machine.  Initially it was written by Columbia
> University, but has been re-named
> since. (from Columbia Monitor System to Conversational Monitor System w=
hen IBM
> started marketing it).
>
> At one time IBM was trying to reduce the amount of source available,
> and there was
> an almost universal uprising from VM systems programmers.  Not the
> fact that the source
> was gone, but without it doing low level debugging was almost
> impossible.  We had some
> great systems geeks that could read core dumps better than the Sunday
> comics, but not
> everywhere was so fortunate.  IBM relented for a while.  I moved on to
> UNIX where OS
> source availability has been generally available at least enough to
> learn the basic
> architectures, even if your specific OS was not easily available
> (HP-UX, AIX, SunOS, but
> old AT&T, Minix, Linux, BSD, etc have been as time goes on).
>
> The MFT, and MVS (the big batch systems) were not in my area, but I
> think my friends in those OS support
> areas said they had the OS code available too (mainly PL/S if I remembe=
r right).
>
> Once cheaper hardware became more prevalent, source code availability
> tightened up,
> so by the time M$oft came around the generic policy was Object Code
> Only for OS's.
>
> Enough of reminiscing.... FOr now anyway.
>
> On Sat, Feb 28, 2009 at 8:29 PM, Jerry Feldman <gaf-mNDKBlG2WHs at public.gmane.org> wrote:
>  =20
>> On 02/28/2009 08:22 PM, David Kramer wrote:
>>    =20
>>> I don't buy that for a second.  Microsoft thrives on being just
>>> incompatible enough with every standard that you can't easily run
>>> someone else's software in its place.  Going open source would also m=
ean
>>> exposing all their file formats, as well as the code they put in to
>>> intentionally break compatibility.
>>>
>>>      =20
>> This is typical of industry leaders. Look at IBM in 60s and 70s.
>>
>>    =20
>
>  =20


--=20
Jerry Feldman <gaf-mNDKBlG2WHs at public.gmane.org>
Boston Linux and Unix
PGP key id: 537C5846
PGP Key fingerprint: 3D1B 8377 A3C0 A5F2 ECBB  CA3B 4607 4319 537C 5846








BLU is a member of BostonUserGroups
BLU is a member of BostonUserGroups
We also thank MIT for the use of their facilities.

Valid HTML 4.01! Valid CSS!



Boston Linux & Unix / webmaster@blu.org