Slashdot Mirror
The Greying of the Mainframe Elite
bobcote writes "The Boston Globe is running a story about the maintainers of the mainframes getting older and facing retirement. One of the problems is that many computer science programs don't include mainframes in their curricula anymore. From the article: "Amid concerns that America doesn't produce enough technically trained young people, mainframe computer users and developers are especially concerned. Most computer science students concentrate on small-computer technology, such as Microsoft Corp.'s Windows operating systems, or the popular alternatives Unix and Linux. Few have been trained on zOS, the operating system that runs IBM Corp.'s massive mainframes."
After all, there's no such thing as digital. Just as all the old analog dinosaurs were retiring the high-speed digital crowd discovered that maybe everything wasn't all ones and zeros.
Same applies to mainframes: mainframe technology has been dissed as obsolete for decades. Just as the microprocessors that (mostly) displaced them finally get to where they can use some of that "ancient" mainframe technology, the people who know how to apply it are leaving.
I'm sure a few will be willing to stay on the job if they're asked nicely enough.
Karma is a bitch -- especially the "comes around" part.
Lacking <sarcasm> tags,
Keep in mind, this is everything for us... and most of us don't even know it.
When you go to the dr's office, guess what's running your insurance data (usually....) ibm.
A friend's dad is 1.6 yrs from retirement and one of the last of the people in his area that run the zOS machines. It is scarry. Truely scarry.
I can talk some hardware with this guy, and a little bit of "good comptuing practices" sort of stuff, but I can't touch him for his knowledge of the workings of the code and systems. And *forget* finding those little "google:howto+topic" miracles like I do daily for my linux admin stuff.
I'm sure most linux savvy ops who know a little about databases could fill in, but there's going to be some issues in the next 5 years or so.
It reminds me of the Cobol joke... about the bloke who earned so much money fixing peoples cobol systems to make the y2k switch that he was able to buy himself a deep freeze. Only to have the 9999 bug crop up. They unfreeze him, tell him all kind of good stuff that's gone on in the world, and then mention to him that since he had Cobol on his resume he was drafted to rewrite some code by the community. (hehe...)
-=fshalor
Yep. To put it in perspective, most of the mainframe people where I work came here from NASA after the Apollo program shut down.
No, I'm not one of them. At 36 I was a kid when most of them came to work here.
"An unarmed man can only flee from evil, and evil is not overcome by fleeing from it." Col. Jeff Cooper
I work with mainframes myself and I can whole heartedly agree with TFA.
:)
Mainframes may not be the fastest growing area in IT, but they will be around for decades to come.
Remember: All your savings and all your bank debts only exist on mainframes. They control your reality.
Unfortunately, I am not Wil Wheaton
No, but a lot of universities had classes in various mainframe-type things, "data processing" and the like. z/OS is just an extension of the systems they've been running for decades, renamed to look "cool." So you probably wouldn't have found, say, a System/390 class specifically at a college, but you would have found a lot of data processing and COBOL classes that would have prepared you to work in that environment.
the college I went to (mid-90's) was phasing those out and bringing in VB and Netware classes. Personally, I think the mainframe-oriented classes were a lot better preparation to work in the IT/IS field than learning how to add and delete users and write "Hello World" with a mouse and a GUI editor.
It's just the payback for the closed source mindset: Mainframes are the biggest players of the secret info game: Pay me $10K and I'll tell you the answer, otherwise your payroll system won't work. Since the keepers of the secrets and the insider priests are dying off, so is the religion they use to control their customers. Meanwhile open systems are growing by leaps and bounds - not with the lush riches of a captive paying customer base but at least it will be around for a LONG time and pay enough to earn a living.
try { do() || do_not(); } catch (JediException err) { yoda(err); }
Mainframe computers are designed around a specific purpose: large volumes of repetetive transactions. This is why they are very prevalent in the banking, credit card, and other financial arenas. They handle the bill processing, customer database, etc.
Sure, you could attempt to blame companies like Microsoft for this, and you would only be partially right. If you do that, you have to add Intel, AMD, Sun, HP, and a whole host of other companies to the mix too, since they all contribute to the "smaller, faster computers are where it's at" attitude. A big reason why this attitude prevails, however, has to do with the "single point-of-failure" issue. When your mainframe crashes, you can do absolutely nothing until the necessary repair work is done. This is where the distributed computing environment works very well.
Having worked on mainframes in the early part of my career, I know that they were useful then, and still are. They excel at what they were designed to do... large volumes of repetetive transactions.
It wouldn't hurt for computer science students to learn about mainframes, or even limited resource embedded systems. It would make them better, more well-rounded IT folk.
OCO is Loco
The one that I like best involved backing up to tape. Apparently tape backup started not as tape, but as thin steel ribbon. This was some heavy stuff, so they employed 3-5 horsepower motors to spin it. Of course, if the motors weren't calibrated right, the steel tape would often snap. One guy even lost his arm to this tape.
How's that for nuts? Computer maintainers don't get these kind of injuries anymore I'd assume. What with steel tape being phased out.
I guess the company I work for works differently. We have our own proprietary operating system and you are expected to know how it works. Luckily we've got a library full of manuals and a test system you can log into.
As for the engineers, we've got a tiered mentoring and peer review process. Yeah, we have a couple of senior engineers leave a year, but by the time they've left, they've also mentored and cultivated the younger enginners.
The training perdiciment is the same all around. Nobody wants to pay for training, so the alternative is reading manuals instead of playing Wow...
Yes Francis, the world has gone crazy.
There were several reasons for this. One was that during IBM's "dark days" in the early 1990's all the young people took the severance packages and fled the mainframe groups. They knew they could learn other technologies and the packages were too good to resist. The older people stuck with what the knew. Then as IBM slowly recovered the recovery didn't focus on mainframe technologies, so new people didn't get hired into those groups. When they finally realized that they did need to hire new people it had been nearly a decade since those old people had trained anybody and they really didn't know how to do it.
I came in with a CS degree from Stanford and was told by one manager that if I worked in his group I would spend two years debugging other people's code. That wasn't attractive to me at all. Bright people want to go somewhere where they can have an impact, but the older guys saw us as a threat and were very reticent to teach us anything. All four of the people I was hired with left for different either different groups in IBM or other companies. The mainframe world couldn't compete with the glamour of the internet boom.
Honestly, I spent four months trying my best to learn this stuff but nobody wanted to teach me. I could see that it was going nowhere. There is going to have to be a real culture change if a hand-off of this stuff is going to happen.
Lasers Controlled Games!
Where I work we have a relatively young staff because only about 1/2 will retire in the next 10 years. At 33 years old, I'm the youngest by about 10 years. One of my co-workers told me that I'll be chained to my desk when I'm as old as she is but those chains will probably be made of gold. Whenever any vendor or customer comes on site the first thing they say is "I never see anyone your age doing mainframe work."
It is a pitty because given a fair chance I bet people would like being an admin once they got past the initial learning curve. The monitoring and automation tools are nothing short of incredible. I can tell what each program is waiting on, what data it is reading, who has higher priority, how long it has been running, how much IO it has done, and lots of other things. I can even alter the memory of the program as it is running (although I'm too chicken to do it). I can also go back in time and get this information from days ago so when I get the "it was slow yesterday" problem I can easily investigate.
I didn't learn a thing from college regarding the mainframe. College was for general logic, problem solving, and overall data structure. Everything I learned was on the job training. When I started one of the older guys said it takes at least 5 years to make a good systems programmer. Anything less and you have a dangerous person who only thinks they understand what is happening. I would have to agree.
The mainframe is really nice in some areas. It is an ego rush to fix a problem that is keeping a multi-billion doller company from shipping any new products (I did that yesterday) and the people I work with are great because they are always willing to share experience and historical knowledge. When they retire I'll miss them.
The price you pay is that many systems have 30+ years of customization in them. They are incredibly complex and very tailored so no two are exactly alike and as a systems programmer I'm expected to be the "final expert" on any problem the users can't solve. This includes finding out why a program that was written when I was three years old no longer reads a PDS properly or why a job that hasn't changed in 5 years suddenly stopped working. It can be lots of fun but it can be frustrating too especially because the bosses really don't want to hear "I don't know" for an answer and "just reboot" isn't even in their vocabulary.
the ability for companies to teach got decimated by the endless rounds of cost cutting.
/V 286, /V Win, /V PM, /V Mac & VisualWorks and VisualAge) all without ever getting an appraisal from one of these HR 'survivors' because they wouldn't know an object if they tripped over one.
:-)
HR people are supposed to be part of the solution, increasing the talets of the pool with 'on the job' training, but they are part of the problem because they are driving the need to increasingly specific 'skill sets' for entry positions.
Entry no longer means, 'getting in, figuring out which way is up, and fitting in making yourself helpful.'
Entry is now a list of requirements being administered by somebody who doesn't know, or want to know, what a job 'might' entail.
They went through the same cost cutting (some might say 'throat-slitting',) as the rest of the organizatin and the HR positions are now staffed by the survivors, the once eigteen-year-olds who managed to hang on because they didn't cost enough to get rid of.
'Knowing' is now everything and 'being able to figure it out' is now worth nothing because it can't be 'measured scientifically' by people who administer the tests.
I am now an old techie and I am just now getting a bachelor's degree in a non-techie field because I couldn't ever get another job doing what I'm doing right now.
I was into object-orientation and Smalltalk since 1985 (Methods) and I am closing my career in 2005 with VSE (after having worked with
I am also aware of the limitations of objects (without relationships, they aren't enough) but I don't care enough anymore to 'fight' the good fight.
The machines that I've worked on (Wang 2200, IBM 360s, DEC PDP/11s, IBM 370s, Z80, x86s, PowerPCs), the languages I've used (BASICs, Cs, Pascals, ProLOg, Lisps, APL, PL/I, Smalltalk's, PHP), the operating systems I've used (Wang BOSS, RSTS/E, OS/360, CPM, Microsoft pre&post Windows, Mac Linux,), the database systems (VSAM, ISAM, IDMS DB, MDBS III, MySQL, PostGreSQL,) didn't really matter worth a damn.
They were just means to an end. I just kept the 'end in sight' and the solution was as simple as following a line.
After 20 years, I figure I deserve a break.
MSBPodcast.com The opinions expressed here are my own. If you don't like 'em... Think up your own stuff.
I agree with you. Unfortunately in the tech world, especially with the fast turnaround employment rate, HR does not want to spend money on training anybody for obscure things, even if one is fully capable of learning the ropes in a matter of weeks and already has a general understanding of it. What companies generally want is people that can do things Right Now The First Time. It really sucks for recent grads. And it's really great for veteran in the field.
Basically what you are left with is 10% of all tech people that are Googleworthy(companies go after them), 30% of all tech people that are trying to get in the field (this includes people that are genuinely interested and people that are in it for the money, although the latter group is shinking very quickly) and 60% that are absolutely mediocre that just happened to be very very lucky and advanced high enough in the corporate world before the bubble burst where they are considered invaluable resources and have no trouble looking for a job. The problem for the 30% trying to get in, is that the 60% mediocre group has set the standard for the industry's performance/level of expected intelligence, and unfortunately, has been set so low that your biggest asset in the hiring phase is proof you've "been there, done that," not your "potential to do it all."
As a 20-something who works with mainframes (and who works for IBM, in the interest of full-disclosure), I must say that learing Solaris or Linux in college does not mean that you naturally have the ability to be skilled in z/OS immediately.
I was a Windows and Linux guy in college, and was hired by IBM to be a mainframe guy right out of college. It took me at least a year, and more like 2, to feel comfortable with the mainframe OS and the concepts associated with the mainframe (like a shared-everything architecture vs a shared-nothing architecture on *nix and Windows) vs. the distributed world.
Most employers don't think far enough in advance (and don't want to shell out the $$) to hire someone to be a "shadow" to the expert for a year (or two) so they can become more than just a blind novice on the platform... they want someone who can contribute now. And don't believe the hype... learning z/OS is not nearly as simple as knowing Unix and applying a few extra concepts to the mainframe side.
As for the guy who said all his friends were concerned about their mainframe jobs and that being a mainframe person was "limiting their options". . . are you serious? There's not a major company in the entire world that's not using an IBM mainframe (with the possible exception of Microsoft, HP, and Sun). Of course, you'll usually be constrained to working in whatever location a company's datacenter is located, but isn't that a contraint you face as a Unix admin, too?
- Proofs of Sturgeon's Law Delivered Daily -