Having programmed in Pascal, Delphi, C, C++ and of course Basic, but never been in touch with COBOL.
So I wonder is COBOL hard to learn?
So I wonder is COBOL hard to learn?
This link should give you a pretty good idea.
https://stackoverflow.blog/2020/04/...is-a-60-year-old-language-suddenly-in-demand/
https://stackoverflow.blog/2020/04/...is-a-60-year-old-language-suddenly-in-demand/
I took a COBOL class around 1991. The fun part was it was on some old machine where it may take 45 minutes to compile, link and run a relatively simple program.
COBOL might hurt your soul if you look at it. On the other hand, the paycheck might be good if you do it.
I happen to have a COBOL programming book that I need to show the kids. One has completed C++ and another has some Java and python experience.
I just want a good picture of their faces as they try to read the COBOL book. They twisted, confused, poor faces.
I'd characterize COBOL as a Domain Specific Language. Technically it is turing complete, but I would not want to write say IC design software with it. C on the other hand along with Pascal, PL/1, ALGOL, ... I'd characterize as general purpose. A really really domain specific one I studied in college is SNOBOL. Interesting language.
History is interesting in this area. I'm old enough to remember having to fill out a form saying why we needed a .com domain name and you did not get it until it was approved. Now they hand them out like cookies, and much less expensive. I think we paid 100/yr for the domain in the 80's. I'm old enough to remember when networking was basically in the building/campus and did not flow outside. ARPANET until the 1980 was basically research, and then few got to access it. The company I worked at out of college was a mil contractor and the project manager used milnet to communicate with DC. I remember ISDN as pt-pt networking, which again was much more secure. As the net grew up, security really did not follow it. the nets allure was it was completely open. Much like MS was a single user OS and suffered more security issues than the *NIX counterparts which were built ground up for hostile environments. NT was a total redesign, but still not as good as the *NIX due to backwards compat needs. Conversely, I'd say *NIX(Linux more than BSD) has gotten sloppier in response to single user use cases. To me sudo was a mistake and really a response to one person machines as an example. Don't get me started on systemd...
But back to network, it is a bit odd that you can connect to essentially any machine in the world and transfer any data you care to. It is a bit odd that someone in NK or Russia or China or Iran can probe my Dallas machine looking for exploits with impunity. (And they do). I still believe the day is coming (maybe IPv6 full adoption) when countries wall off the external packets and they have some sort of "port of entry" approach like goods into a country have. China already does this inward.
History is interesting in this area. I'm old enough to remember having to fill out a form saying why we needed a .com domain name and you did not get it until it was approved. Now they hand them out like cookies, and much less expensive. I think we paid 100/yr for the domain in the 80's. I'm old enough to remember when networking was basically in the building/campus and did not flow outside. ARPANET until the 1980 was basically research, and then few got to access it. The company I worked at out of college was a mil contractor and the project manager used milnet to communicate with DC. I remember ISDN as pt-pt networking, which again was much more secure. As the net grew up, security really did not follow it. the nets allure was it was completely open. Much like MS was a single user OS and suffered more security issues than the *NIX counterparts which were built ground up for hostile environments. NT was a total redesign, but still not as good as the *NIX due to backwards compat needs. Conversely, I'd say *NIX(Linux more than BSD) has gotten sloppier in response to single user use cases. To me sudo was a mistake and really a response to one person machines as an example. Don't get me started on systemd...
But back to network, it is a bit odd that you can connect to essentially any machine in the world and transfer any data you care to. It is a bit odd that someone in NK or Russia or China or Iran can probe my Dallas machine looking for exploits with impunity. (And they do). I still believe the day is coming (maybe IPv6 full adoption) when countries wall off the external packets and they have some sort of "port of entry" approach like goods into a country have. China already does this inward.
I'm not an expert, but in my opinion issues start with corporate networks managed by incompetent system administrators and employees (please note that I'm not saying everyone is, just many) who don't give a damn if the world falls apart.
So an internal threat that accesses external threats.
And external threats that attempt to access corporate networks.
So they had to think of a remedy to prevent anyone in a company from opening any type of file without worrying about the consequences.
The slowdown of corporate systems is also due to the fact that any executable file (including script and batch) could be malicious.
And to incompetent system administrators.
There is no way to know before it executes a hostile code and huge databases are created and accessed simultaneously by thousands and thousands of requests that are inhibited first and then analyzed and then maybe executed.
And all this takes time.
Someone thought of reassuring lazy and incompetent system administrators by offering them a fabulous software that in a short time was installed on millions of machines.
That company in a short time reached a turnover of 4 billion dollars.
Thanks to fearful, lazy and incompetent system administrators (please note that I'm not saying everyone is, just many) who thought that all their problems were solved by software that uses artificial intelligence.
This.
So an internal threat that accesses external threats.
And external threats that attempt to access corporate networks.
So they had to think of a remedy to prevent anyone in a company from opening any type of file without worrying about the consequences.
The slowdown of corporate systems is also due to the fact that any executable file (including script and batch) could be malicious.
And to incompetent system administrators.
There is no way to know before it executes a hostile code and huge databases are created and accessed simultaneously by thousands and thousands of requests that are inhibited first and then analyzed and then maybe executed.
And all this takes time.
Someone thought of reassuring lazy and incompetent system administrators by offering them a fabulous software that in a short time was installed on millions of machines.
That company in a short time reached a turnover of 4 billion dollars.
Thanks to fearful, lazy and incompetent system administrators (please note that I'm not saying everyone is, just many) who thought that all their problems were solved by software that uses artificial intelligence.
This.
Not saying there aren't any, but having worked in highly managed environments, the complexity of dealing with hundreds or thousands of users is immense.
The drive to standardize started partly on cost, but also because of security concerns driven by the difficulty of applying software fixes in a timely fashion across a large organisation potentially on different continents and timezones.
So companies naturally looked for solutions that made this pigs breakfast manageable.
If you read the security articles in The Register you'll appreciate how complex the security world has become.
And TBH just how many mistakes can be made and how much unpatched and vulnerable kit is still out there.
The reason for its existence. But VERY clumsy.
FORTH is my favourite language of all time, but it is not really suited. I think a nice little 16 bit system on a chip that doe sFORTH natively could be a really useful thing.
dave
I thought it was cute. I had a FORTH ROM for my BBC micro in the 70s, I dabbled but really didn't have a project I needed it for and RPN hurt my head. All the action was in assembler or Basic.
Thought Modula2 had promise...
Mostly now I program in Java or C/C++. Keep thinking I should try Rust, dunno if there's a compiler for Arduino or not.
I can't disagree, but please note that I just said
😉
However, it is also software-monopolies that generate all this. IMO
I've done a bit of assembler (Z80, 8085, and some stuff on an 8-bit ST micro back in 1990-1) and then went across to C, C++ on the mBed platform after a long hiatus. That's closing down in July 2025, so I am just going to migrate across to the ST dev environment for GP ARM MCU's. ST have a really good selection of MCU's with loads of I/O options and a very good peripheral configurator.
Re the CrowdStrike thing - I can't imagine how difficult it must be to manage large organisations with multiple sites spread across the globe. I worked for a large corporation and the IT guys had their hands full.
Re the CrowdStrike thing - I can't imagine how difficult it must be to manage large organisations with multiple sites spread across the globe. I worked for a large corporation and the IT guys had their hands full.
Turbo Pascal - now that is OLD. Was that back in about 1985/6? I remember reading articles about it in Circuit Cellar (or was it Byte?) and thinking about learning it, but I never did.
In the early 80s i had a board with a dedicated FORTH processor, 16 kB of static RAM. It smoked the 286 we got when they first started appearing. It went nowhere.
I also wrote a FORTH implemntationfor the 6809 (best assembler i ever used). The Telesketch. Computer + telephone. The designer had some of the early optical computer patents. Telesketch went nowhere too.
Me too.
My first language was Algol W, my prof was one of the authors. Pascal is a frustratingly simple version specifically aimed at learning high level programming.
dave
Point taken.
I used to work for a relatively small (around 50 people) company, we struggled to get business because many of our competitors were much larger and or customers ran 24x7 businesses and wanted to be assured that we'd be around to support our stuff in London, NY, Tokyo etc.
Eventually we were purchased by a competitor 20 times our size that could offer that level of support and distribution.
True. But, to an extent, the monopolies themselves reflect complexity. There's a lot of time and effort that goes into making globally distributed, real time updating, multi-platform software.
I used to work for a relatively small (around 50 people) company, we struggled to get business because many of our competitors were much larger and or customers ran 24x7 businesses and wanted to be assured that we'd be around to support our stuff in London, NY, Tokyo etc.
Eventually we were purchased by a competitor 20 times our size that could offer that level of support and distribution.
Did a small amount of programming in Algol 68 as it was the system software language for ICL2900 series mainframes.
The assembler for those mainframes looked a bit little like 68000 assembler, but had higher level functions and actual honest to goodness strings.
On the system we worked on we did fixes to the COBOL code then used the assembler listing to create patches in assembler to be distributed to customers.
Those were the days...
My AlgolW experience was in the early 70s, and ran on an IBM 360. I was good at 360 assembler too. And that early experience was very useful when the summer course i took a decade later was 360 assembler. I wrote some very nice code. Taught the instructor a few things.
dave
dave
Those poor grad student teaching assistants are usually only a day or two ahead of the students in the course.
We did ECAP on the 360 in CS101, and later on SPICE.
This is a nice synopsis of circuit simulation history.
We did ECAP on the 360 in CS101, and later on SPICE.
This is a nice synopsis of circuit simulation history.
I'm not thrilled with your characterization of so many Sysadmin as lazy. More like, overworked, and half-exhausted
while frequently underpaid.
It's simple marketing: Make the target market feel like you sympathize with their pain, and then give them the
"solution". It's the oldest trick in the marketing book.
