The State of the Mainframe? Better Than Ever
Software Developer Chris Craddock explains what's great about the open, modern mainframe
This transcript is edited for clarity.
Reg Harbeck: Hi, I’m Reg Harbeck and today I’m here with my friend and colleague Chris Craddock, who I have known for a good part of my career. Rather than giving a current title for you, maybe if I can get you to introduce yourself. First of all, what is it you do right now, and then we’ll kind of dig deeper?
Chris Craddock: Okay, sure, yes. Right now, I do z/OS software development for some commercial z/OS software providers, I do AWS web/cloud development for some other clients and I do technical due diligence evaluations for software acquisitions for a software company that’s acquiring other software companies to build a portfolio.
Reg: Cool. Now of course to have such a wide range of abilities, you sort of had to get there from somewhere, and I happen to know that although you are talking about AWS, you also have a very deep mainframe pedigree. So maybe if I can get you to sort of say how did you end up on the mainframe back in the beginning or whatever part of your career that was?
Chris: I went to university. I grew up in Australia. I went to university to study mathematics and physics, and of course I was mad keen on that. Then somewhere partway through my university career, I discovered programming. That was the end of that.
Reg: So what language was that?
Chris: I started out with FORTRAN. One of the professors challenged me to produce a program for him that would translate prime numbers palindromically and then check if the reversed digits were still prime.
Reg: Oh, how fun. So, are there very many palindromic prime numbers out there?
Chris: There’s an alarmingly large number of them, yes.
Reg: Wow. That is cool.
Chris: Yes. There’s an infinite number of prime numbers, so there may well be an infinite number of palindromically prime numbers.
Reg: So, there’s probably more palindromic prime numbers than Mersenne prime numbers. There are only about [51] of those known to exist.
Chris: Maybe, maybe, I’m unsure. But in any case, you know, I was probably 19 years old and I thought, how hard can it be?
Reg: Yeah, well that is cool.
Chris: So I picked up a book and I wrote my first FORTRAN program.
Reg: You said—so now this was in your last year of university, second to the last year?
Chris: Second to the last year of university—and by being successful at that as an undergraduate, I got an account in the computer center, which undergrads never got, and I learned UNIX. This is all like around 1977-1978.
Reg: So UNIX was less than a decade old at this point?
Chris: Oh yeah, it was dripping wet. Then I went to work and I started working on MVS mainframes in my first job.
Reg: This was the original MVS we’re talking about because this would be right at the time that it became MVS.
Chris: It was right at the split between 3.8 and SP, so I actually started on MVS/SP in 1980.
Reg: So, you didn’t get to see the source code for the stuff you were working on [laughs].
Chris: No, I did not. In fact, I’ve seen a bunch of fiche over the years but no, I never got to see source code, and I worked for a government entity. Then I was doing software development in an insurance group, and I was doing a lot of APL software development. APL had this kind of cool extension capability that would allow you to build a program that you could interact with the outside world from within the APL world. So, I built a set of tools that allowed me to do system programming tasks—
Reg: In APL?
Chris: In APL, yes.
Reg: You like pain, don’t you?
Chris: You know, yeah, I kind of do like pain, I guess. Then fast forward a couple of years, and I got recruited into an IBM federal assistance division project. That’s where I got to do some software magic that caught the attention of some people in IBM, and that ultimately led to me being recruited to a startup in California in Silicon Valley. I went and started there in 1990—well late 1990, early 1991—where I was in pain again. I was doing UNIX kernel development and hardware architecture for a really exotic multiprocessing UNIX system.
Reg: Now was this entirely in C or did you have to go deeper in some cases?
Chris: Interestingly, because a lot of this—UNIX at the time didn’t really support multiprocessor machines. It did not until SVR 4.2, I think. So, before that if you wanted to do multiprocessing, you kind of had to do something yourself. That was my introduction to C++, and so I built the kernel extensions in C++ because why not?
Reg: Yeah.
Chris: I wasn’t there all that very much longer. Then I went to Boole & Babbage and then I was in heavy duty core system stuff.
Reg: Boole & Babbage: I’m going to guess a lot of their stuff was written in high level Assembler.
Chris: Yes, most of the stuff was in Assembler. Some of their stuff was written in C. There were things like the views and so on in the monitoring products, but I was responsible for the low-level infrastructure. I took over the low-level infrastructure from Ron Higgin, our mutual friend and colleague. That was acquired by BMC shortly after and I had a little brief stint in the middle at Candle with a project that was started and then canceled. And so I—
Reg: Moved on again?
Chris: It was almost like I was never there. So, I went to BMC—or I was asked to go back to BMC, and I was at BMC for quite a long time. I built the new mainframe infrastructure for essentially all z/OS products out—
Reg: So that includes the Control suite.
Chris: Yeah, most of the products. Even now, 20 years later, most of the BMC product line either runs directly on that infrastructure or uses parts of that infrastructure.
Reg: Now does that infrastructure have a name or a tag? All the other mainframe vendors seem to like to group their infrastructure stuff with a particular name. Was there a name you called—?
Chris: It is RTCS. It’s the Run Time Component System: RTCS.
Reg: RTCS, right. Okay. So now by this time it was about mid 1990s when you were at BMC. Is that about right?
Chris: No. I went to BMC between ‘98 and ’99. I did all of that work, and we delivered the infrastructure and a couple of products up to like around 2002, and then I did a lot of other sorts of infrastructure-y tasks and—
Reg: This is around the time that you and I met, when you moved on from BMC?
Chris: Yes. I moved on from BMC to what was then Computer Associates, which is where I met you. I left BMC not for any particular personal reasons with BMC, but I had a need for health insurance that BMC did not provide. I had no choice but to go somewhere else, and so I ended up at CA and I enjoyed my time at CA.
Reg: Fair enough. Well, you certainly made waves while you were there—you and Ron Higgin, of course. You and he were just a spectacular team.
Chris: Thank you. Thank you.
Reg: So that still brings us another two decades of stuff. What have you been up to since then?
Chris: I went to Hewlett-Packard. I was in the CTO office at Hewlett-Packard for like 4 1/2, 5 years, and in my last year and a half of Hewlett-Packard I was the CTO for product development in the services business—so, what used to be EVAs.
Reg: Oh wow.
Chris: So kind of overarching responsibility for things like the airline management systems and insurance managing systems for customers.
Reg: Now this substantially was this just on a UNIX platform? How many other platforms were you also using?
Chris: Well ,the airline stuff was mainframes, but a lot of that was TPF and I was the architect for a redevelopment of the HP.com e-commerce site, which at that time was about the sixth biggest site in the world.
Reg: Wow.
Chris: And that was on HP-UNIX machines and you know, usual work technology. Then I had a succession of Silicon Valley things, Salesforce and Intuit and anyway—
Reg: So, what would make you go to SHARE? On the one hand, obviously, you’ve got a depth of mainframe context, that might be a reason, but I gather that you’re not currently focused on the mainframe so much as that’s part of your whole list of different things that you know about. What brought you to SHARE?
Chris: Well, I’d been doing cloud stuff—predominantly AWS—for most of the last 16 years but also Google Cloud and the Microsoft Azure Cloud and a number of other things. I built a cancer research cloud for Oracle on the new Oracle cloud, but a couple of years ago, I guess right around the COVID time, I was approached by some old cronies to see if I would be interested in doing some Z software development, product development for a company that was doing some automation stuff. I kind of thought well, I don’t know. Sure, why not? I wonder if I can still do that? It turns out I could.
Reg: Like riding a bike.
Chris: Like riding a bike. So I did two products and then maybe oh probably about a year ago Dave Cole—again, mutual friend—asked me if I would come build some software for him.
Reg: Cool.
Chris: So I’m building software for Dave Cole on Z obviously and that’s probably 50-60% of my time, and I’m doing technical due diligence evaluations for another company that’s doing software acquisitions, company acquisitions, and then I’m still doing cloud development stuff for exotic things like processing medical records in PDF form and extracting all the unstructured data from the PDF and transforming it into structured data that can be stored in the database for medical research. Things like that. So that’s the sort of stuff that keeps me busy.
Reg: So I get the distinct impression that you have no intention of slowing down.
Chris: No.
Reg: I support that. I also do not intend to retire. I’m going to be at it as many decades as I can. So we’re probably the same that way as well. That said, that means that you’re going to be here for a good part of the future of the mainframe. So tell me [about], you know, the near, medium- and longer-term future of the mainframe. You’re talking about your future, so maybe you can give us some insights. What do you think is going to happen?
Chris: Even though I’ve been a Champion of Z and been down in the guts of Z for decades, pretty much outside of IBM development I’m probably one of the handful of the most knowledgeable people in the world about how that beast works.
Reg: Right.
Chris: Throughout the industry we’ve been talking about modernization for going on 30 years, and what’s most exciting to me is it is finally happening. What I find the most exciting is all of the things like Zowe and z/OSMF and newer languages coming onto the platform and opening it up to our kids—and our grandkids, frankly. So I’m diving headlong into helping out Frank DeGilio with z/OS automation with Python and I use the VS Code/Assembler editor every single day.
Reg: Cool.
Chris: I use all of that technology and it’s fabulous, because for the first time ever this platform is on a level playing ground. From a development technology point of view, it is on a level playing ground with the entire open systems world because the open systems world has been frankly light years ahead for such a long time. Now all of a sudden in a big rush, just over the last few years, it’s like oh my goodness, I can use all this stuff and deploy software on Z. How good is that?
Reg: That is awesome. Okay, any last thoughts before we finish up?
Chris: It’s been a pleasure seeing you again.
Reg: Yes, and you, Chris.
Chris: And I’m going to keep coming back to SHARE.
Reg: Excellent. Excellent.
Chris: It’s like seeing old friends every time. And I figure right now I will probably keep building Z software for the rest of my life, however long that is.
Reg: Excellent. Thank you so much, Chris. This has been outstanding and very insightful. So tune in next month and we’ll have another one of these. In the meantime, do check out TechChannel.com for all kinds of other content. I’m Reg Harbeck.