Skip to main content

Not Your Typical Mainframer: Dr. Magie Hall on COBOL Training

Reg Harbeck: Hi, I’m Reg Harbeck and today I’m here with Dr. Magie Hall, who is assistant professor of strategic business analytics at WU Vienna. Magie, welcome. You’re not a typical mainframer as such, but your role in the mainframe and legacy and business computing ecosystem is tremendously important, so we’re really thankful for this opportunity to get to know you. Maybe if you could start by figuring what’s your background? How did you end up dealing with computing issues, given that I understand your degree isn’t even in computing?

Magie Hall: Yeah, that’s my secret shame, I suppose. I was originally trained in the social sciences. My bachelors and masters were both in political science, and I transitioned into technology for my PhD at Karlsruhe. It was great because all of these things I had always done as a social scientist you could do with a couple of lines of code, and there went years of my life that I had spent hand-coding things with a just a couple of sentences functionally. From there I moved to the University of Nebraska and I got a position in IT innovation, was kind of looking around Omaha. I don’t know if you know Omaha. It’s this great environment because it’s kind of alone. There’s not a lot going on. Kansas City is probably the next biggest city, and from there it’s eight hours to Denver, eight hours to Chicago. So it’s very unique, and it’s got a great business and technology ecosystem. I was looking around. I have this sort of a hunch about online learning is that it doesn’t do its job. We thought about online systems even before Covid. We thought about online courses that are going to be for the masses. We’re going to democratize education and the people who graduated from online courses, the people who are successful in online courses, they were you and me, functionally. They’re people who are already successful at it, already well-educated, already had multiple advanced degrees. So I had the sort of suspicion that we could probably do online education better, and we could use it to help people who actually needed it. For me that was looking at adults experiencing homelessness. I went around Omaha—again, a great community—and just sort of knocked on everyone’s door who would let me in their door for about 15 minutes. I asked them what would it take for you to hire someone who is currently homeless into an IT role? What kind of job skills do they need? What kind of background do they need? What would it take for you to overlook this problem? It was uncanny. I talked to so many more people than I needed to because I didn’t believe it. Almost to a person, they all said we need the mainframe—specifically, we need COBOL. There is no one teaching it in this region. We don’t have enough people by spades. If you had somebody who had sort of, Python—one of the new languages, so that they’re not isolated—and then COBOL, specifically COBOL, then we’ll hire them in heartbeat. Oh, okay. And I’ve been working on that project ever since.
Reg: So now I’m going to dig into your interesting background. Your accent suggests to me that you grew up in the United States. Is that a good guess?

Magie: Yes so, I’m originally from the Pittsburgh region, sort of northwest by about an hour—there’s this weird little spot between Pennsylvania, West Virginia, and Ohio where they are all basically connected, and I’m from that.
Reg: Oh cool, I went there. My kids and I went on a big long North America wide road trip, and we wanted to go Steubenville. Sunday morning, Steubenville, Ohio—and so I was just using the app and just grabbing the nearest hotel and somehow ended up in West Virginia, which boggled my mind because I didn’t know West Virginia goes that far west—
Magie: It goes all the way up.
Reg: Yeah. It was an interesting experience which probably is too much for this podcast, but let me just say that I’ve never been in a restaurant where the nonsmoking area was an area where you get stared at, like asked to leave as quickly as possible, and it takes up only a tiny fraction of the restaurant. It was fascinating [laughs].
Magie: Yes. Welcome to the northern Appalachians.
Reg: So that said, obviously you moved from there to not just academia, but European academia.
Magie: Yup, absolutely. I kind of took a circuitous path. I graduated from my bachelors in Pittsburgh, so I call myself as someone from Pittsburgh even though I’m from the Appalachians. From there I had this sort of international life. I was in Lebanon and Northern Africa, and Germany—
Reg: Wow.
Magie: And just sort of traveled all around. It was great. I totally recommend it. If you’re in your 20s and can manage to do it, do it, even if you have to take on a little debt. So the student loan debt, oh my God—but at some point I was living in Switzerland and working for the UN, and decided that this was just not my life, I guess. I wanted something a little bit more independent. I wanted something that I could own more, and academia is always going to be that route. I got an offer from my professor—they call him the Doktorvater here, the father of your doctorate. I got an offer from my Doktorvater Gustav Weinhardt and moved to Karlsruhe, where Karlsruhe Institute of Technology is one of the premier Germany or German technical universities. I couldn’t have done any of this—we wouldn’t even be having this conversation without that chance.
Reg: Wow. Now this whole time though—I mean you’re going up the social sciences route and was it in Omaha or in Germany or somewhere else that you suddenly started making that connection to technology?

Magie: My love has always been words and that in the social sciences as well, and it carried through to the technology. I eventually did something with sentiment analysis, NLP sort of stuff. I love words. I love understanding meanings behind words, understanding their uses. The word latency is the thing that I love the most. I use it all the time when I’m talking or teaching classes to students. Latency the way the computer scientists mean it and the way the rest of us mean it are not the same thing, and it matters that there is a difference there. So because I love words, it was fairly easy to transition into sort of an NLP space sentiment analysis space and—
Reg: NLP, just for our audience. That’s natural language programming.
Magie: Sorry, yes. Natural language processing.
Reg: Processing! Sorry. Keep going.
Magie: Yeah, so methodologically I used NLP or sentiment analysis, or a little bit of both, and built out my skill set. I do also some work with image analysis, social analytics. I use social media data and of course none of this is a part of the mainframe.
Reg: Of course everything touches on the mainframe, and right now—
Magie: Yeah, a lot of it.
Reg: We all need to be lifting.
Magie: Absolutely. More things run on the mainframe that you’d be surprised, I guess.
Reg: Oh yeah.
Magie: Or you wouldn’t be surprised, but many would be.
Reg: Well what’s interesting here is that so far, your journey has been an intellectual and academic journey. You and I are on the COBOL committee. We just had a call recently and were talking about how the COBOL and the business computing and mainframe world has something of a distinction from the academic technology world that is really important as we take a look at how to make sure it continues to be staffed. And to see that you’ve got this wonderful linguistic orientation and word orientation, which is something my family cares very much about. My brother Dr. James Harbeck actually writes a nearly daily column called Word Tasting Notes, [[ LINK: ]] just about the taste of words in language. So I really get what you’re talking about here, but it’s interesting because your reason for your current activity comes from a completely different but related part of yourself, and that’s your care for humanity, which is so important… Then at some point you took this connection from social sciences and language and all of these, and then suddenly—bam! You saw an opportunity to bring it all together in humanity. How the heck did that all come together?

Magie: If I’m going to be overly honest, I’m going to say it came out at the bottom of the third bottle of wine one day [laughs].
Reg: Oh cool.
Magie: Yes, exactly. I have a great colleague who is at Google, and he shares this same sort of sense that all of this is just not doing what it’s supposed to do. We shouldn’t have some of the problems that we have if technology is really performant the way that we originally intended. If you look back at sort of the original foundational documents of creating the internet, it’s supposed to be for education and it’s supposed to equalize use and it’s supposed to be this great tool, and it’s not. That’s not fair. It’s not fair because we all put this in together, but we’re not getting equal use out of it. So he shares this really deep conviction. It’s Marcus Kreuzer. He’s over in California and at the bottom of way too many bottles of wine, we took a piece of plastic and stuck it to my wall and started mapping out the sort of, how can we fix this problem using education as our cornerstone. So that’s really how it started. That’s the overly honest version, though.
Reg: I like it! That’s great. It’s so human, once again. So as sort of a person who is really carrying this burden on your shoulders—just by choice, as somebody seeing a need that is out there and everybody else is too busy earning an income to realize that the well is going to run dry if they don’t prime it. Somehow having had this opportunity to see the need—I gather in some ways in North America but clearly, you’re working on it, literally working on it—but in Austria now. How did that all happen?

Magie: I have my Austrian position. I was at University of Nebraska and my husband is German, actually. And the pandemic was very, very hard, having family all over the world and everybody is under different lockdown constraints. It became sort of a family imperative to try to get back to Europe to be a little bit closer to my in-laws. So there’s nothing magic there. I’m still working on the project. I’m still pushing through this conversation. Our major site is Omaha. That’s where the project is based for now, but we’re looking always at ways to expand it into different cities. Dallas is going to be probably a big site for us. We’d like to go into California because the problem is so acute there, and of course bringing it back here to me in the EU.
Reg: Now I’m going to guess that the homelessness experience in the EU is—although people are still basically people—but there are probably different nuances to it, one of which is that you guys are a lot further north than most people realize. As somebody who is based in Canada, I’m startled by the fact that so much of the EU is actually parallel to Canada and not the United States. So you have the same day light and sometimes some of the same weather, although we’re all kind of cooking right now. But I’m curious. As you talk about these different geographies, both within the United States and then external to the US, how the homeless situation and the ability to incorporate that into the future of IT vary, but have things in common?

Magie: I would say the more interesting case here is how do we use the same systems that are working with homeless adults in the States, and use it for current social problems in Europe. And for us, that’s going to much more be things like linguistic minorities, people who come in from let’s say Eastern Europe and then move to any of the western European EU countries. That’s going to be asylum seekers. You saw just a few years ago where we had this just millions of people, trains of walking migrants, coming out of Afghanistan and Pakistan, for example, and integrating these people who are just very much on the margins of society with that same mechanism though. So take these critical undertaught languages that run everything we do and make sure that they have the skills that they need to be so well trained that nothing else matters.
Reg: Hmm cool. I love the faith in humanity and individual humans given an opportunity that you have here, and I support it. That said, do you have any practical examples of what particularly works well in creating that connection in any of your locations yet?

Magie: Oh, that’s such a big question, I almost need to have you shrink it down. So are you talking at the curricula level, or are you talking about at the partnership level? There are so many elements.
Reg: Well let’s talk about the goal, which is to have the mainframe staffed up. As I told you in our previous conversation before this interview, back in 2004 I wrote a white paper about the need to get a new generation on the mainframe, and still 18 years later it’s people like yourselves—who hadn’t read that white paper but just saw a need—who are actually doing something about it. The people in charge of the various organizations with mainframes and running COBOL are too busy on day-to-day business to really look up above the horizon, and so as we take a look at the fact that we really desperately need a new generation on the mainframe. The average mainframer’s age would still be going up if it weren’t for the fact that there is now a few new mainframers. But if you can think of some specific examples, maybe a bunch of different unrelated examples, where people now have jobs in the organizations that run the world economy, writing COBOL or doing systems work on the mainframe or stuff like that that is a direct outgrowth of your creating that on ramp from being homeless?

Magie: Okay, perfect. We are partnered with a selection of organizations, mainframe houses all over the midwest functionally, a couple of organizations in the south as well, and what we do with our program is we talk to these partners. We talk to these partners and we ask them to give to us something that’s wrong with their organization—so, buggy code or QA tasks or something that for them represents a common thing that their staff would be dealing with, their IT staff would be dealing with either in the mainframe or in app development—and we take these problems and actually have our learners create solutions for them. So it’s like a portfolio of experience that the learners have, and give these solutions back to our partners. From that, it creates an assessment center function. You have shown what it is that you would do if you were attacking this buggy code or if you were doing some sort of testing, and that gives them, the hiring manager and the HR management, sort of a baseline level of being able to understand. So that’s sort of our magic trick there. It’s really heavily aligned on partnerships. In terms of hiring, we see a lot of interest in testing, QA and testing. I’m not sure why that artifact is happening but we get a lot of interest in our learners mainly because of the degree mismatch, right? Getting people in the door when they don’t have a degree or they don’t have a fitting degree can be tough, so we do get a lot of orientation towards QA and testing.
Reg: Hmm interesting.
Magie: Like to change that but—
Reg: Yeah. Testing is a really important part of it, and I have to say testing does require a set of perspectives that are non-conformant with the perspectives of those writing the products. I always figure that a CIO who had never actually worked with the technology or a 2-year-old might be among the best testers in terms of just discovering stuff, especially because I mean we have so codified all the standard tests using automated testing that it’s the ability to discover unexpected things that are such a big part of that. So it’s obviously somebody with an unexpected background can do that, but it’s also a foot in the door of course, not only for them but for other folks who are homeless. Now I assume that there is a very substantial—not just educational but certificate or other kind of qualification—granting part of what you are doing so that employers can have the comfort that they’re dealing with somebody who really does have the smarts and the dedication. How does that all work?

Magie: Yup and we are, actually in some ways, the credential problem. We create our curricular materials on the basis of sort of the standard learning material, so the same thing that you or I were look at if we were trying to upscale ourselves on either Python or on COBOL or JCL, right? We take those same curricular materials and we work them through with much more—how do I say this without being mean to computer scientists [laughs]? With language that is much less exclusionary. We pay a lot more attention to context clues, to readability, things that aren’t necessarily standard in your intro level CS 101 class. I’ll say it like that. And so we use examples that work for their daily lives but also, we do so in language that’s not exclusionary but it’s the same learning material that everybody is doing whenever they’re doing their entry level, learn how to program, learn the concepts of computer science courses, and then they work through solutions. We actually pay our learners for providing solutions so they learn.
Reg: Oh neat.
Magie: They get to do the materials, and every time we give them a quiz, every time that they submit a little bit of code, we give them a little bit of money—sort of like an apprenticeship or almost like a crowd worker maybe is the right term—and they get a little bit of financial support, plus you’re learning a skill. You can’t argue anything against IT being a skill. And eventually—this is where we were talking about the challenges earlier—everything they do builds up into a skill space portfolio. Because what I can’t do is give some mom who lost her kids because she’s an addict and she’s in a recovery program, I can’t give her a four-year computer science degree in six months. It’s beyond me, right? It’s beyond anybody. But what I can do is say okay, I can guarantee that this person has learned the skills and the way that I can do that is show you what they have achieved. So we act as the clearing house there or the credential grantor, but everything we’re looking at is really showing ability as opposed to proving aptitude.
Reg: Now you’ve touched on a number of really important things to me, and in some ways what they all come down to is the nonacademic perspective because you’re talking about skills, not knowledge, and this is something I’ve found consistently in my computing career. If I want to teach somebody something about computing, I can’t sit at a chalkboard or even at a screen. I need to sit them down at the keyboard and get that muscle memory. That’s a skill thing, that’s not a knowledge thing—but then also choosing language that speaks to them. You know as somebody who comes from an academic family, I like to joke that my natural language is sesquipedalian pedantic English, which is great if you want to speak to university people but in the real world where business people are trying to get business results, that kind of language excludes you, even as it is language that’s intended sometimes to exclude people who aren’t academics. I think this is one of the big issues is in the world of business IT, which especially includes both COBOL and the mainframe, that we don’t need highfalutin academics. We need hard working, smart people who can just take stuff and make it work and don’t need to be using $5 words. And so I’m really enthusiastic about your discussion of using a language that speaks to the actual workforce that you want and not to some academic whose approval you win using big words. Maybe if you can just give a few additional insights into the cultural and linguistic approach to getting real people involved.
Magie: Sure. So, Reg, what is a variable?

Reg: Hmm, there’s a good thing because as a computer person of course I’m going to say well, a variable is a memory location where we keep data that may be subject to change and you refer to in your program. But I’m going to guess you’ve got a different definition.
Magie: So you lost me at subject to changing, and I’m someone who deals with variables a lot. You can explain a variable. You could also give, let’s say, an analogy of variable. A variable for example could be the difference between a check and a unit of currency—a dollar, right? Both of them have some sort of representation of value, but neither of them is the value themselves.  They’re just a representation of it. So if I write you out a check for a dollar, if I give you a dollar, we have some sort of representation of that value. And a variable does the same thing. It represents some sort of value and you can have it in different formats, but also you can expand, shrink, contrast anything that you need to do, and those common different types of measurable units—we can talk about those measurable units in a longer conversation that’s not in the podcast, but this is that sort of just perspective change that you need to have. We can always use the real definition, or we can take just really basic examples that come from people’s lives. It’s an algorithm. Well do you look out of the window in the morning to check to see if it’s going to rain, and if it’s going to rain, you think about taking an umbrella with you, right? The sort of patterns that are ingrained in you, those are really just your brain’s algorithms. So now let’s break down what it is that these indicators are that you have something that’s in a pattern moving forward like that.
Reg: Hmm cool. So mapping it back to a more realistic experience vs. just pure concept—which is so funny because that’s really as our colleague Misty Decker likes to point out, the people the business technology was made for are business people—secretaries, accountants, administrators—people who are doing a job to get results, not people who are trying to show off their knowledge. And so to use real human concepts to understand this stuff I think is something that you can use for everybody else having developed it for people who really function at that level to discover we actually all function at that level. That’s when we’re actually being functional vs. kind of showing off our knowledge or something of that nature. And so I’m really optimistic that a lot of these findings that you’re making will map to a whole lot of other people. That said, if you were to put yourself in a position of somebody who could not merely predict but prodict, dictate the future of business computing, large enterprise business computing including organizations that have COBOL and have mainframes, what would that future look like and what would you have done or participated in to make it happen?
Magie: My sad perspective is I think it’s going to continue the way that it is going right now. There are some really great efforts out there, and Open Mainframe Project among them, of sort of saying hey, look. We’re walking into a burning fire. This is a slow-moving crisis, but like many slow-moving crises, we only really deal with it when it’s a little bit almost too late, and the mainframe has proven itself over and over and over again to be similar to a slow-moving crisis. The example that all of these mainframers had to come back from retirement at the beginning of Covid so that people could get their stimulus checks. That’s shocking and it should be terrifying to all of us, and yet we’ve made zero changes, right? The mainframe sector is just showing this over and over and over again. Slow-moving crisis, we’re going to deal with it when it’s way too late. What I would like to have an answer is that there are ways for some of these problems that we see, so integrating people at the margins of the digital era on the side of the digital divide, we can find a better way to almost leverage the two crises off of each other, right?

Reg: Hmm I like it.
Magie: We have two problems, but if we put them together we can find a solution. I don’t need these people who are in my program to take calculus, right? That’s one of the classic conversations in computer science education right now. How much math do they need? Calculus is just one of those hills we all like to die on. None of these people need calculus if they’re getting hired to do quality assurance level one. I’m sorry—they don’t, all right? So let’s meet people where they are at, meet the technology need where it’s at. It turns out that in some cases, we can really take two problems and it creates a better solution. That’s what I would love to see hopefully. You know I’m a pessimist and everything that I said first was wrong.
Reg: That’s awesome. Magie this has been an absolutely wonderful conversation. Is there anything else you’d want to share with us? I don’t want to miss the opportunity to hear it.
Magie: We are always looking for people to partner with. We are always looking for people to interview, to understand better what the needs are in the entire mainframe talent pipeline, not just on the COBOL programming. So reach out. We have a partner in Dallas, we have a partner in Omaha, we’ve got myself. And thank you, Reg. This is such an opportunity. Thank you.
Reg: Maybe one last thought then is can I get an email address or some other way people can reach you and contact you if they have questions or ideas or—
Magie: Certainly. Do you want me to say it out or would you rather type it?

Reg: Sure. Yeah, spell it out and I’ll put it in the transcript.
Magie: Okay so margeret dot hall at wu dot ac dot at.
Reg: Well thank you so much, Magie. I really appreciate this. I’ll be back with another podcast next month but in the meantime check out the other content on TechChannel. You can also subscribe to their weekly newsletters, webinars, e-books, Solutions Directory and more on the subscription page. I’m Reg Harbeck.