Another nasty holiday season on RBT

AMuzi wrote:
On 1/24/2019 7:32 PM, wrote:
On Thursday, January 24, 2019 at 5:00:50 PM UTC-8, Frank Krygowski wrote:
On 1/24/2019 6:30 PM, John B. Slocomb wrote:
On Thu, 24 Jan 2019 17:38:39 -0500, Frank Krygowski
wrote:

On 1/24/2019 5:13 PM, wrote:


You with your Cobol who doesn't seem to know that all of those cobol
systems were replaced by full systems written by Oracle and the like, ...
Cobol was used to write business systems and those were almost
entirely replaced with large company systems written by major corporations.

I'm not an expert on computer languages, but I remember reading this not
long ago in some liberal rag or another: "Indeed, despite its advanced
age, Cobol is still the most prevalent programming language in the
financial-services industry world-wide."

https://www.wsj.com/articles/do-you-...you-1537550913

What does WSJ stand for again?

I don't know whether Cobol is still the most prevalent language in the
financial services but a few years ago my boat was parked next to a
guy's who worked for IBM on a project to replace the computers and
design a totally new software system for a major bank here.

The testing of each portion of the new system took in the neighborhood
of 6 months and IBM was committed to have several of their people
assigned at the bank for a number of years after the system was
completed in case of any problems.

Given the time lost and problems that might be encountered I would
think that changing an operational Cobol system over to a different
language might not be exactly what a bank or large financial house
might want to do.

Like so many things, it probably comes down to benefits vs. detriments.
You've described some detriments of migrating from a working COBOL
system to something more modern. What would be the benefits of investing
that time?

Or flipping the coin: What are the detriments of continuing to run a
COBOL system that's performing well?

In case anyone takes this wrong (hey, it's Usenet), I'm asking serious
questions. I really want to know.

(I do know one guy who, after retiring from his normal job, got several
years of good freelance work because he was good at COBOL.)

--
- Frank Krygowski

I bought a new 64 bit 4 core HP desktop. I got a light line instead of a
phone connection. I went from being in the bottom 5% of computer speed
to the top 10% in the entire world. My system boots so fast you don't
even notice it. The screen takes more time to come up than the computer
to boot and stabilize.

With large scale systems this is a relatively minor upgrade. Do you
suppose a large banking system would worry about changing from an old
system if they could get a performance upgrade that substantial?


Large firms generally, especially banks and financial firms,
still do, in fact, run vintage mainframes with Cobol.

http://cobolcowboys.com/cobol-today/

Being 'old' is not a critical factor for things which work
well, for example my Corvair or the meat motor on my 1953
bicycle.


If you have a small clean system with well documented interfaces and
functions and you want to ditch COBOL, it's a piece of cake. However, if
you have a 50 year old accretion of code, string and barnacles that nobody
knows what it does (although they all agree it currently works), there may
be better things to do with your money until the day that you absolutely
HAVE to replace it.

AMuzi wrote:
:On 1/24/2019 7:32 PM, wrote:
: On Thursday, January 24, 2019 at 5:00:50 PM UTC-8, Frank Krygowski wrote:
: On 1/24/2019 6:30 PM, John B. Slocomb wrote:
: On Thu, 24 Jan 2019 17:38:39 -0500, Frank Krygowski
: wrote:
:
: On 1/24/2019 5:13 PM, wrote:
:
:
: You with your Cobol who doesn't seem to know that all of those cobol systems were replaced by full systems written by Oracle and the like, ...
: Cobol was used to write business systems and those were almost entirely replaced with large company systems written by major corporations.
:
: I'm not an expert on computer languages, but I remember reading this not
: long ago in some liberal rag or another: "Indeed, despite its advanced
: age, Cobol is still the most prevalent programming language in the
: financial-services industry world-wide."
:
: https://www.wsj.com/articles/do-you-...you-1537550913
:
: What does WSJ stand for again?
:
: I don't know whether Cobol is still the most prevalent language in the
: financial services but a few years ago my boat was parked next to a
: guy's who worked for IBM on a project to replace the computers and
: design a totally new software system for a major bank here.
:
: The testing of each portion of the new system took in the neighborhood
: of 6 months and IBM was committed to have several of their people
: assigned at the bank for a number of years after the system was
: completed in case of any problems.
:
: Given the time lost and problems that might be encountered I would
: think that changing an operational Cobol system over to a different
: language might not be exactly what a bank or large financial house
: might want to do.
:
: Like so many things, it probably comes down to benefits vs. detriments.
: You've described some detriments of migrating from a working COBOL
: system to something more modern. What would be the benefits of investing
: that time?
:
: Or flipping the coin: What are the detriments of continuing to run a
: COBOL system that's performing well?
:
: In case anyone takes this wrong (hey, it's Usenet), I'm asking serious
: questions. I really want to know.
:
: (I do know one guy who, after retiring from his normal job, got several
: years of good freelance work because he was good at COBOL.)
:
: --
: - Frank Krygowski
:
: I bought a new 64 bit 4 core HP desktop. I got a light line instead of a phone connection. I went from being in the bottom 5% of computer speed to the top 10% in the entire world. My system boots so fast you don't even notice it. The screen takes more time to come up than the computer to boot and stabilize.
:
: With large scale systems this is a relatively minor upgrade. Do you suppose a large banking system would worry about changing from an old system if they could get a performance upgrade that substantial?
:

:Large firms generally, especially banks and financial firms,
:still do, in fact, run vintage mainframes with Cobol.

Oh, the hardware is all amazingly modern. The code, not so much.

I know some former cobol programmers, and one current one. One of them
makes a **** ton of money at it, but he's been doing it forever, is very
good at it. The other people who have done it all make more money, in
better working enviornments, using modern languages.

(I work in lisp, so I get the obscure and ancient points.)


--
sig 111

Frank Krygowski wrote:

:Like so many things, it probably comes down to benefits vs. detriments.
:You've described some detriments of migrating from a working COBOL
:system to something more modern. What would be the benefits of investing
:that time?

Better understanding of how it all works. It's easier to add new
features, or change existing ones. It's easier to be sure it won't
all stop wroking because some greybeard retires, and no one else knows
that every fifth tuesday of the month, you need to jiggle a lever.

:Or flipping the coin: What are the detriments of continuing to run a
:COBOL system that's performing well?

:In case anyone takes this wrong (hey, it's Usenet), I'm asking serious
:questions. I really want to know.

It costs money. Lots of money. You have to figure out what the
current system does -- what it actually does, not what you think it
does, not what the documentation says it does, but what it actually
does -- and turn that into specifications for the people doing the new
work. Then they have to write it, it has to be tested, etc. Then you
have to make the changes that the production system got (because a
regulatory requirement changed, for instance) and test them.

the common way of dealing with it is to do new work in something else,
but leave the core stuff alone. So if you add a mobile phone payment
application, that's done in shiney new stuff, but it still talks to
the cobol system that does the actual debiting of checking accounts.


--
sig 85

On Fri, 25 Jan 2019 02:15:17 +0000 (UTC), Ralph Barone
wrote:

John B. Slocomb wrote:
On Thu, 24 Jan 2019 17:38:39 -0500, Frank Krygowski
wrote:

On 1/24/2019 5:13 PM, wrote:


You with your Cobol who doesn't seem to know that all of those cobol
systems were replaced by full systems written by Oracle and the like, ...
Cobol was used to write business systems and those were almost entirely
replaced with large company systems written by major corporations.

I'm not an expert on computer languages, but I remember reading this not
long ago in some liberal rag or another: "Indeed, despite its advanced
age, Cobol is still the most prevalent programming language in the
financial-services industry world-wide."

https://www.wsj.com/articles/do-you-...you-1537550913

What does WSJ stand for again?

I don't know whether Cobol is still the most prevalent language in the
financial services but a few years ago my boat was parked next to a
guy's who worked for IBM on a project to replace the computers and
design a totally new software system for a major bank here.

The testing of each portion of the new system took in the neighborhood
of 6 months and IBM was committed to have several of their people
assigned at the bank for a number of years after the system was
completed in case of any problems.

Given the time lost and problems that might be encountered I would
think that changing an operational Cobol system over to a different
language might not be exactly what a bank or large financial house
might want to do.


Cheers,
John B.

It's like road disc brakes or 12 speed cassettes. You know that eventually
you're going to have to convert, but it may be a good business decision to
postpone the conversion as long as possible.

I'm not sure that I will ever need a 12 speed cassette :-) Or disk
brakes :-)


Cheers,
John B.

On Fri, 25 Jan 2019 02:15:18 +0000 (UTC), Ralph Barone
wrote:

AMuzi wrote:
On 1/24/2019 7:32 PM, wrote:
On Thursday, January 24, 2019 at 5:00:50 PM UTC-8, Frank Krygowski wrote:
On 1/24/2019 6:30 PM, John B. Slocomb wrote:
On Thu, 24 Jan 2019 17:38:39 -0500, Frank Krygowski
wrote:

On 1/24/2019 5:13 PM, wrote:


You with your Cobol who doesn't seem to know that all of those cobol
systems were replaced by full systems written by Oracle and the like, ...
Cobol was used to write business systems and those were almost
entirely replaced with large company systems written by major corporations.

I'm not an expert on computer languages, but I remember reading this not
long ago in some liberal rag or another: "Indeed, despite its advanced
age, Cobol is still the most prevalent programming language in the
financial-services industry world-wide."

https://www.wsj.com/articles/do-you-...you-1537550913

What does WSJ stand for again?

I don't know whether Cobol is still the most prevalent language in the
financial services but a few years ago my boat was parked next to a
guy's who worked for IBM on a project to replace the computers and
design a totally new software system for a major bank here.

The testing of each portion of the new system took in the neighborhood
of 6 months and IBM was committed to have several of their people
assigned at the bank for a number of years after the system was
completed in case of any problems.

Given the time lost and problems that might be encountered I would
think that changing an operational Cobol system over to a different
language might not be exactly what a bank or large financial house
might want to do.

Like so many things, it probably comes down to benefits vs. detriments.
You've described some detriments of migrating from a working COBOL
system to something more modern. What would be the benefits of investing
that time?

Or flipping the coin: What are the detriments of continuing to run a
COBOL system that's performing well?

In case anyone takes this wrong (hey, it's Usenet), I'm asking serious
questions. I really want to know.

(I do know one guy who, after retiring from his normal job, got several
years of good freelance work because he was good at COBOL.)

--
- Frank Krygowski

I bought a new 64 bit 4 core HP desktop. I got a light line instead of a
phone connection. I went from being in the bottom 5% of computer speed
to the top 10% in the entire world. My system boots so fast you don't
even notice it. The screen takes more time to come up than the computer
to boot and stabilize.

With large scale systems this is a relatively minor upgrade. Do you
suppose a large banking system would worry about changing from an old
system if they could get a performance upgrade that substantial?


Large firms generally, especially banks and financial firms,
still do, in fact, run vintage mainframes with Cobol.

http://cobolcowboys.com/cobol-today/

Being 'old' is not a critical factor for things which work
well, for example my Corvair or the meat motor on my 1953
bicycle.


If you have a small clean system with well documented interfaces and
functions and you want to ditch COBOL, it's a piece of cake. However, if
you have a 50 year old accretion of code, string and barnacles that nobody
knows what it does (although they all agree it currently works), there may
be better things to do with your money until the day that you absolutely
HAVE to replace it.

I mentioned the guy in the next boat that worked for IBM. The system
that they were installing for a major bank was expected to take more
then a year to install and test. Not a minor investment in either time
or money.


Cheers,
John B.

On Thu, 24 Jan 2019 14:13:58 -0800, sltom992 wrote:

On Thursday, January 24, 2019 at 11:36:14 AM UTC-8, news18 wrote:

Going back, there were plenty variations of memory chips. At one stage,
that was how hardware received "code" updates. Using such shouldn't
have been an issue for Tommies Million dollar plus project.

You with your Cobol who doesn't seem to know that all of those cobol
systems were replaced by full systems written by Oracle and the like,

I hate to bust your bubble, but I know about the rollout of DBMS and if I
was to follow your method I'd be squarking how now one else in a state
government department could make the CASE data dictionary work in a
distributed way so it worked throughout the state. I was just told it
needed to be installed and set up and went on my merry way to the shock
of the Oracle guy who later turned up, expecting to have to do it all and
who only had to do the fine tuning. Shrug, it was just par for the course
when I used my skills.

Have another go and before you do, brush your teeth, It might make you
appear a bit more intelligent;
https://www.newscientist.com/article...lly-know-what-
causes-alzheimers-and-how-to-stop-it/

On Fri, 25 Jan 2019 06:30:11 +0700, John B. Slocomb wrote:


Given the time lost and problems that might be encountered I would think
that changing an operational Cobol system over to a different language
might not be exactly what a bank or large financial house might want to
do.

Funnily, I've seen that in manufacturing when I sold a dozen ancient
network cards to a big US manufacturer of construction equipment. $US100/
card two decades ago was small beer compared to the millions they had
tied up in the sutomation of the production line.

Kicked my self that I'd landfilled half of them the week before. XT cards
out and hang onto the AT cards a bit longer.

On Thu, 24 Jan 2019 20:00:47 -0500, Frank Krygowski wrote:


Like so many things, it probably comes down to benefits vs. detriments.
You've described some detriments of migrating from a working COBOL
system to something more modern. What would be the benefits of investing
that time?

Or flipping the coin: What are the detriments of continuing to run a
COBOL system that's performing well?

In case anyone takes this wrong (hey, it's Usenet), I'm asking serious
questions. I really want to know.


It is really a matter of whether you can get ongoing support for the
hardware and can still get software updated.

And if you can stall the upgrade a little longer, then something far
better may come along, so your ownly doing one upgrade/changeover
compared to three.

Also, some companies/people are really hanging in there with their old
kit being fully depreciated and maintanable where it isn't worthwhile to
a competitor to invest in new equipment.




(I do know one guy who, after retiring from his normal job, got several
years of good freelance work because he was good at COBOL.)

On Thu, 24 Jan 2019 17:32:43 -0800, sltom992 wrote:

I bought a new 64 bit 4 core HP desktop. I got a light line instead of a
phone connection. I went from being in the bottom 5% of computer speed
to the top 10% in the entire world. My system boots so fast you don't
even notice it. The screen takes more time to come up than the computer
to boot and stabilize.

With large scale systems this is a relatively minor upgrade. Do you
suppose a large banking system would worry about changing from an old
system if they could get a performance upgrade that substantial?

"Light line"? like fibre optic compared to old copper.

Ooooh shinies, **** weak, real computers crunch data and do massive
amounts of it it rather than flashy lights for entertainment of children
and upgrading your little toy woukdn't even scratch their work load.

What do you actually do with it?

How do you know you went from the bottom 5% to top 10%?

FWIW, I avoided paying the HP tax and just went yumcha systems of which
we have 8 64bit boxen being 3x8 core, 3x6 core and 2x4 core. It is gross
over kill for a backwater website, mailer, ftp boxen, spinning rust
storage and two linux workstations and one MSWin boxen for games.

The real reason why so much grunt; it is more economic to buy them then
chase 4cores globally.

Aaah, you probably beat me in the NAS we purchased, it is only an ARMv8 2
core.

On Friday, January 25, 2019 at 1:45:18 AM UTC, AMuzi wrote:

Being 'old' is not a critical factor for things which work
well, for example my Corvair or the meat motor on my 1953
bicycle.

--
Andrew Muzi
www.yellowjersey.org/
Open every day since 1 April, 1971

"the meat motor on my 1953 bicycle"

Heh-heh. So modest!

Andre Jute