Home > Archive > ASM370 > March 2004 > new to mainframe asm
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
new to mainframe asm
|
|
| v3ct0r99 2004-03-19, 8:28 pm |
| Hi all.
I'm currently taking a course in s/370 assembler at local college.
The school believes s/370 will provide us knowledge to get work
as a systems analyst\programmer.
I already had some exposure with low-level x86asm.
So far I like HLASM alot...
It's odd how this has all evolved for myself.
3 years ago it was "I wanna hack" and I got many virus and
pissed off my friends and family. A year later it was c/c++;
another year and i realized being 1337 meant learning
asm and architecture. Now I'm over all that shit and thinking
about getting real work. So there appears to be more work in
mainframe realm and many "real" coders find intel chips to
be inherently flawed(beginning to see this)...
Learning architecture is intense stuff so I need to choose
wisely. I'm 25 and I dont think I can do both at the same level
of expertise, at least not as a 5 year goal. I predict I
can become a 'guru' of either art in 5 years of intense study...but again
will my efforts be wasted? Five years could mean a whole new
technology; I'd rather go with a classic that will never die,
like a Brooks Bros. suit which is always in style ;)
regards,
Jason
| |
| Peter H. 2004-03-19, 8:28 pm |
| [color=darkred]
I'm 25 and I dont think I can do both at the same level of expertise, at least
not as a 5 year goal. I predict I can become a 'guru' of either art in 5 years
of intense study...but again will my efforts be wasted?[color=darkred]
Eight to ten years of intense study will most likely be necessary.
Although I started out programming assembly language on BUNCH machines, I
quickly went to the then-new S/360, and I spent 30 years programming these
machines.
Most systems programmers can master at most one or two components (these being
Task Management, I/O Supervisor, Job Management, Data Management, etcetera,
etcetera, etcetera) of the S/360 OS or its successors.
It is quite rare to find someone, such as myself, who has successfully mastered
all components.
| |
| Anne & Lynn Wheeler 2004-03-19, 8:28 pm |
| peterh5322@aol.comminch (Peter H.) writes:
> Most systems programmers can master at most one or two components
> (these being Task Management, I/O Supervisor, Job Management, Data
> Management, etcetera, etcetera, etcetera) of the S/360 OS or its
> successors.
>
> It is quite rare to find someone, such as myself, who has
> successfully mastered all components.
from presentation i gave at boston share meeting while undergradudate:
http://www.garlic.com/~lynn/94.html#18 CP/67 & OS MFT14
besides the pathlength stuff mentioned in the above ... one of the
changes was the genesis of fair share scheduling .... which was
shipped in standard cp/67 release, dropped in the initial conversion
to vm/370 and then re-introduced with the resource manager:
http://www.garlic.com/~lynn/subtopic.html#fairshare
another undergraduate activity was the original global/clock
replacement algorithm. at the time, the working set paper had just
been published in ACM ... but this work differed significantly
.... including the use of global replacement instead of local
replacement.
http://www.garlic.com/~lynn/subtopic.html#wsclock
over then years later when somebody had done their phd thesis work at
stanford on clock ... there was big issue made about whether global
replacement was better than local replacement (and whether or not the
person should even get their phd). part of the resolution was early '70s
comparison on same hardware & software base of both local and global
replacement implementations (with global replacement having upwards of
300 percent better thruput compared to local replacement).
various postings about getting blaimed for helping create ibm PCM
controller market ... when I was undergraduate, I added tty/ascii
support to the cp/67 kernel and was trying to play games with 2702 to
do automatic terminal type recognition. it turns out while you could
change the line-scanner with the SAD commands, they had took shortcut
in 2702 and hardwired the oscillator ... so in testing, I sort of got
TTYs to work on "2741" baud line. as a result, a project was started
at the university that reverse engineered the ibm channel interface
and built a channel board for an Interdata3 minicomputer. the
interdata3 was then programmed to simulate 2702 functions ... but line
baud rate was software programmable (instead of hardwired oscillator).
http://www.garlic.com/~lynn/subtopic.html#360pcm
--
Anne & Lynn Wheeler | http://www.garlic.com/~lynn/
| |
| Peter H. 2004-03-19, 8:28 pm |
| [color=darkred]
various postings about getting blaimed for helping create ibm PCM controller
market ...[color=darkred]
Yeah ... I was employed for nearly 20 years by Amdahl Corporation.
Provided MVS software support and technical support of sales for the 470 and
580 series.
Team leader for Amdahl's first PCM introduction, the very successful 4705
series of 3705 replacements.
Specified and architected the 4725 and 4745, both being 3725 and 3745
replacements.
Designed the 4745's "Very High Speed Scanner" feature ... T-1/E-1 to/from the
FEP to a BOC CSU.
Developed compatbility code for MVS (all then-current versions) to support the
6100 DASD controller as a 3990 in an SMS environment.
| |
| RFCOMMSYS 2004-03-19, 8:28 pm |
| v3ct0r99@hotmail.com wrote:
>
> Now I'm over all that shit and thinking
>about getting real work. So there appears to be more work in
>mainframe realm and many "real" coders find intel chips to
>be inherently flawed(beginning to see this)...
>Learning architecture is intense stuff so I need to choose
>wisely. I'm 25 and I dont think I can do both at the same level
>of expertise, at least not as a 5 year goal. I predict I
>can become a 'guru' of either art in 5 years of intense study...but again
>will my efforts be wasted? Five years could mean a whole new
>technology; I'd rather go with a classic that will never die,
>like a Brooks Bros. suit which is always in style ;)
>regards,
>Jason
>
>
Best of luck to you, but don't be so sure about IBM mainframe architecture
"never dying". I've been an IBM mainframe assembler programmer for 25 years,
and I must tell you that everywhere I look, I see signs of the mainframe's
demise. For example, the place I work at recently had a management coup, taken
over by a bunch of idiots who think that hundreds of servers running unsecure
Microslop "software" is better than that stodgy old IBM mainframe.
Now I face the prospect of checking my brain at the door and becoming a
point-and-click "programmer". How degrading. Mainframe jobs have virtually
dried up in my town and state, so I'll have to move out of state if I get
desperate enough. Check out the various employment search websites for
"mainframe assembler", and you'll see what I mean.
Mind you, I'm not saying that mainframe architecture SHOULD BE obsolete, or IS
in any way inferior to this Microslop shit. But that's the general impression
management has nowadays. After all, the Microslop machines display pretty
colors and animated smiley faces, so they MUST be better, no?
| |
|
|
|
|
|