The problem isn't COBOL, it is actually not that difficult to learn, JCL on the other hand is a PAIN IN THE @$$. I ended up dropping the class.

What is COBOL?

Article is dead on target. Quality IT people are in short supply whether it is mainframe or distributed systems.

At least IBM and the IBM user group SHARE is doing something about it. Where are the other vendors?

(In the interest of full disclosure, I am involved in the SHARE user group and one of the sponsors of the zNextGen effort which is developing new mainframers.)

This whole panic over vanishing mainframe skills seems a bit ill considered. Prior to starting my own web applications consulting firm, I was a contract IT worker. Just 3 years ago, my resume clearly indicated almost 10 years of COBOL skills.

On most occasions when I applied for a mainframe job, I heard nothing back; the resume vanished into the void. The rest of the time I was passed over during initial communication because of the lack of some minor skill (such as flowcharting which is obsolete anyway).

You would think that if an IT hiring manager were that desperate for COBOL programmers, he/she would hire any experienced person he found and then train them in the minor peripheral skills. But then again, most IT positions are handled by HR people who have no knowledge of the field. I wonder if they would reject an applicant for the job of carpenter because the resume didn't explicitly say that the person had experience using a saw.

Christopher hit the head of the nail.
HR must stop age discrimination disguised by trivial detail.
Boomers are relatively old and often retired but they are not all dead yet.
Boomers love to travel. They are trained to work where they travel.
Boomers could use the extra cash. All Boomers could use the respect honestly earned over many years. Boomers make good mentors, teachers, leaders. Boomers may be tired, retired, retreads but not yet dead.

I feel we are at least 10 years away from a mainframe crisis. Lots of people learned BAL, COBOL and JCL during the 70's, putting them in their 50's now. Even if they have been doing other things, I think most programmers could bring themselves back up to speed quickly.

We also need to remember that COBOL was designed to be easily learned. With a good set of manuals or a self study course, people who currently program in a more modern language, should be able to learn everything they really need.

Finally, if there is money to be made, the market will produce skilled legacy programmers.

It was many years ago when I was a contract programmer I had an upcoming COBOL gig. Granted I had Mainframe experience with another language but I had 3 weeks to learn COBOL. I picked up a manual read through it and did some practice programs by translating some small applications from the language that I knew into Cobol to get the syntax right.

I dove right into the project and learned while I went. No one knew that I didnt have COBOL experience. Any smart developer can pick up COBOL.

I would think there are other even more talented people out there who would roll up their sleeves and do the job.

I learned COBOL, I liked COBOL, and seven years ago, I looked for a job using COBOL and RPG as well -- there were none because everyone was looking for JAVA, PERL, and C++ developers. While somewhat vague, for the most part I have forgotten COBOL, but it won't be that difficult to re-learn should the need arise.

> Finally, COBOL-skilled programmers...are available domestically at an inflated price.

Isn't an "inflated" price a consequence of supply and demand? It's funny how extraordinarily high CEO salaries are "necessary to attract talent" but (attempts at) higher worker salaries are "inflated".

If you pay them, they will come.

Jose - If you think JCL was hard don't come looking for a PERL job.........

I have to agree with the last 6 posters starting with Christopher through Devon. The language sets are't that hard to pick back up. I spent 30 years as a consultant and learned and dropped a bunch of peripheral skills. There is no reason why there is a shortage other than management doesn't want to pay a US living wage for COBOL skills. The last mainframe job I had disappeared when the company was 5% short of a "prophet" target. I got tired of being refused interviews because I didn't have some planning tool or CICS screen painter or a reporting language in my "Recent" skill set. Today I have my MCSD & A+, am working on my MCSE and am a Systems Administrator. I don't make as much as I used to (yet) but I'm not sitting on the sidelines waiting anymore either. I am perfectly willing to pick up a COBOL coding jobe on the side using my VPN connection and telecomuting. I like playing with my grandkids too though so I won't be traveling unlike cornstoves :-)

Take the blinders off. COBOL isn't the only mainframe language around. There are many Pick (or multi-value) shops and Pick vendors around, IBM being the biggest of these.

What we in the Pick community experienced in the 90s was an industry-wide panic among management over being left holding the 'outdated technology bag' (because a few vendors like Microdata, Prime, Sanyo Icon, and Ultimate either collapsed or shrank), usually compounded by some fresh-from-Harvard MBA or other such know-it-all (who actually only knew Oracle) began campaigning for Oracle to replace the dinosaur 'COBOL'.

Most of these firms then allowed attrition to thin the ranks of their mainframe staff. Most also went bankrupt because of this. Companies that actually survived such stupidity (like Dynix, now SirsiDynix), neglected to offer existing staff any path to keep their jobs, preferring to keep them on only to maintain existing systems while new staff were hired to develop 'next generation' products, after which mainframe staff would be laid off.

Worse, some firms used this upheaval as a smoke-screen to mask their efforts to outsource their work to India, and even to Siberia. (No, I am NOT kidding! And, yes, I have first-hand knowledge of this, and I can even name names of companies and managers engaged in the practice. Just check out the growing din over H1-B visas, and firms who coach clients in tricking the system to legally pass over locals in favor of foreigners.)

Such treacherous behavior by management destroyed the trust. Those who could, moved into management or support roles at former clients who were determined to shield themselves from their former vendors' insanity. Those who had the opportunity to acquire 'new' skills did so, but then moved on to other employers. Many left IT altogether. These never looked back, and couldn't be lured back even by the higher demand and wages the remainder, like me, now enjoy, albeit in a much narrower and smaller market.

There is much to be said about the mainframe (any of them). I've had several 'web' developers (like my own brother) bemoan the shifting sands their whole industry is built on, from the technologies themselves to the employers. They look longingly on mainframe staff who get to work on stable systems whose major evolution is behind them, and whose major clients have been around for decades. But they fear making the jump because of the fud the media, inflated by management, constantly fill their eyes and ears with. Reading the trade rags, one would never guess that there was anything out there other than perl, php, .net or AJAX.

So, just like foreign car makers vs. the big 3, the problem is one of management's own making. Fix management's fickle nature, and the problem disappears.