NJ needs COBOL programmers

[raz-0 1,259]

On 4/8/2020 at 1:39 PM, MartyZ said:

The funny thing is that corporations and governments have been trying to get rid of cobol for over 40 years now, and guess what, it ain't going nowhere. Mainframes can handle more stress, and they can do it faster then any server farm out there. Look at google, they have over 500k servers worldwide to be able to handle their volume, while the largest corporations, that handle 100s of millions of transaction daily, only need about 10 mainframes. Contrary to popular belief, mainframes are still more powerful then most servers out there.

Now consider the programming language, how many mainstream mainframe languages are there, maybe 3 or 4. Now how many distributed languages are there, 100s? with new ones popping up every day. What's going to happen in another 40 years when we have a shortage of Java, C#, C++, Angular, JS, etc... programmers? 

BAHAHAHAHAH!!

100s of millions of transactions a day!!!!! Oh my god sooo many transactions. 

This is the funniest shit I have seen all day. You do realize that it isn't even rounding error on google's volume right? 

If you don't have coherent work to perform that exceeds the capacity of a server then something bigger is stupid. You can go mainframe, or you can go clustered. One will be IO limited though. Which is why most new mainframes/supercomputers look a heck of a lot like server clusters when you cut them open and look at the guts. 

As an employee of a business that STILL buys mainframe crap from IBM after they said please stop buying mainframes, I will tell you that it sure as shit isn't outpacing our non mainframes. We've been bottlenecked by them for the last 20 years. 

Nursing them along has just been cheaper than doing the right thing and replacing them. Until now. Not because replacing them got cheap, but because the cost of the things they could not do well got to expensive. 

Really big computers don't have much purpose other than trying to undermine encryption or doing really massive simulations. 

Also, the argument of "what's going to happen when we need language X programmer in 40 years" is the worlds shittiest argument in favor of mainframes and cobol. Whatever happens, it literally can't be worse than what has happened with that combo. Which is having massive amounts of neglected systems running it with nothing but a few boutique shops of legacy developers essentially doing forensic maintenance. 

At least with current stuff, if people aren't fucktards about data preservation, you will have version control left, and with cheap massive storage, the odds the digital devleoper docs and such won't be tossed in the dumpster. But even if both fail? Right where we are with cobol. 

 

[MartyZ 692]

4 minutes ago, raz-0 said:

BAHAHAHAHAH!!

100s of millions of transactions a day!!!!! Oh my god sooo many transactions. 

This is the funniest shit I have seen all day. You do realize that it isn't even rounding error on google's volume right? 

If you don't have coherent work to perform that exceeds the capacity of a server then something bigger is stupid. You can go mainframe, or you can go clustered. One will be IO limited though. Which is why most new mainframes/supercomputers look a heck of a lot like server clusters when you cut them open and look at the guts. 

As an employee of a business that STILL buys mainframe crap from IBM after they said please stop buying mainframes, I will tell you that it sure as shit isn't outpacing our non mainframes. We've been bottlenecked by them for the last 20 years. 

Nursing them along has just been cheaper than doing the right thing and replacing them. Until now. Not because replacing them got cheap, but because the cost of the things they could not do well got to expensive. 

Really big computers don't have much purpose other than trying to undermine encryption or doing really massive simulations. 

Also, the argument of "what's going to happen when we need language X programmer in 40 years" is the worlds shittiest argument in favor of mainframes and cobol. Whatever happens, it literally can't be worse than what has happened with that combo. Which is having massive amounts of neglected systems running it with nothing but a few boutique shops of legacy developers essentially doing forensic maintenance. 

At least with current stuff, if people aren't fucktards about data preservation, you will have version control left, and with cheap massive storage, the odds the digital devleoper docs and such won't be tossed in the dumpster. But even if both fail? Right where we are with cobol. 

 

Really? IBM is telling companies to stop buyin mainframes? Is that why they are pushing zOS connect as if it's the best thing since sliced bread? IBM is pushing zOS connect to allow shops to expose CICS code as APIs and large compnies are buying into it. Why, because it's easier, and the old addage holds true, "if it ain't broke, don't fix it"

What I don't understand is why there has to be a new language popping up on a regular basis? What is the driving force other then to create chaos in the industry to have hundreds of different languages that do the same thing? Cobol could have been adopted to run in a distributed environment, and it was tried, but there was no followthru, why? Because there was no money to be made?

For those who believe that cobol will die, and believe in all the BS arguments of why it should die. Just keep in mind that those arguments and that belief are older then some people on this board. And will still be around long after most of us retire.

[45Doll 5,873]

Well if I have nothing better to do maybe I'll take the CICS Tutorial. There's a lot of buzzwords I'd almost mentally erased.

The last company I worked for directly in the 90's wanted to replace their mainframe based customer service system. It was supposed to be a 'client/server system based on Unix, Oracle and OO language capability'. After the contract was signed...

The vendor: "We can't guarantee the C/S system will work."

Bwuhahahaha! It ended up as legacy RPG on an AS400 with a DB2 backend. Definitely ready for the turn of the millennium!

 

[raz-0 1,259]

2 hours ago, MartyZ said:

Really? IBM is telling companies to stop buyin mainframes? Is that why they are pushing zOS connect as if it's the best thing since sliced bread? IBM is pushing zOS connect to allow shops to expose CICS code as APIs and large compnies are buying into it. Why, because it's easier, and the old addage holds true, "if it ain't broke, don't fix it"

What I don't understand is why there has to be a new language popping up on a regular basis? What is the driving force other then to create chaos in the industry to have hundreds of different languages that do the same thing? Cobol could have been adopted to run in a distributed environment, and it was tried, but there was no followthru, why? Because there was no money to be made?

For those who believe that cobol will die, and believe in all the BS arguments of why it should die. Just keep in mind that those arguments and that belief are older then some people on this board. And will still be around long after most of us retire.

IBM literally said we won't sell you more mainframes. Then customers kept on trying to buy them and at some point not upholding your fiduciary responsibility as a publicly traded company is bad for you as an individual.  

So now they sell them again and have spent time room to figure out how to integrate then into a modern environment better. 

Internally it seems to be converging on simply a specialized version of modern architecture being presented via clustering as a traditional mainframe. Not because it is better but because there is money in enabling the mindset of maximal neglect.