DISCOVER – Part 2

R.B. – Welcome to IBM i DevOps TechTalk, where we discuss key topics and questions with ARCAD experts. I’m Ray Bernardi and I’ll be your host today. Joining me is Alan Ashley, a senior solutions architect here at ARCAD. In today’s TechTalk, Alan and I will be talking about ARCAD discover. Discover is ARCAD’s AI powered analysis tool. It’s designed to provide deep insights into your applications. It allows your teams to easily visualize and document those applications. Whether you’re a developer working to modernize your code, a business analyst aiming to understand functional logic, or an IT manager focused on governance and compliance. Discover has something for you. Let’s get started. So, Alan, since our last talk on discover, quite a bit has been done to the product.
Could you go over some of the highlights and some of the features and some of the things that have been added?

A.A. – Oh, sure. So since our last episode, I think it was episode 20 that we talked about DISCOVER and how you can look at your repository and start to gain some of that application knowledge from that. Well, since then, we’ve added several new features and enhanced a couple of features that were already there. One of those was kind of a code description area where you can look at the code, you can look at the subroutines and you can see how it is being put together.
Well, now with the help of AI, we can actually break down some of those descriptions from just a summary of what it’s doing to a full technical analysis. And what is also nice is you now have the option to compare subroutines between programs, and you can pull them up on the screen graphically, have a compare and look at the technical descriptions of those side by side, going through and seeing some of those similarity.
Maybe that subroutine is something that could be a candidate for a service program, that could be used by all the programs, if it’s that similar. That’s just one of the ways you can start to look at that application, not just from a code explanation standpoint. And another point on the kind of the code descriptions that come out of this with the AI, since we do have different levels from summary to generalize to true technical, a developer can look at it, a Java developer can look at it who may not understand RPG and see how that function is working.
But you may also have a business analyst or an auditor that can look at it, and they just need a summary of what it’s doing. Will you have that? And what’s nice now is you can actually save off stage descriptions and include that in your documentation model that you build later, which is really one of those things that we know developers don’t like to write documentation.
This can help facilitate that function going forward. Now the other area is kind of a view of how you look at your application. Back in the day we called them functional trees or macroscopic views of your application. So you can see all the linkages between components. Because we had standard naming conventions. In today’s world there’s no naming conventions for the most part, and you don’t can’t tell visually which programs go with what. You have to actually analyze the programs, dig into and see what they are. Well, again, with the help of AI and the metadata repository that we have access to that defines your application. We can look at those programs and we can see the interactions.
And sometimes these programs have weird names. There’s no commonality in the name. You can’t look at it and see what it does. Well now we can take those and we can see what the linkages are between the programs. So if you had a bunch of 15, 20, 30 programs that were dealing with order entry, but none of them had “ord” in the program name or the file name, we can figure out that those are part of the order entry through the some of the descriptions.
And then we can label that functional tree as maybe order entry or something to that effect. Now you can go in and tweak it after the fact. You can rename it to maybe something more pertinent to you, but that is something that we’re building automatically from the repository based off of the information we have about the application. And to me, from a high level view of looking at an application, what a great way to look at it.
You can just kind of zoom in on your different components. You can start to see the interactions between those and some of the impacts. You can actually right click on some of these and say what’s the impact to the shown components. Maybe not all the components being shown yet. And you can see those interactions and how the impacts are going to start going through there.
So when you start having changes to your database, changes to programs, you can start to see what some of those impacts are going to look like.

R.B. – But that’s pretty neat, Alan. So you’re basically telling me that this thing can look at things very detailed, like a microscopic view, but you can also come up to a macroscopic level. Now, one of the things that I’ve heard about with this new releases is called data lineage. Now, does that have anything to do with everything that you’re talking about?
How does that play into this?

A.A. – Okay. So Data Lineage is really looking at how fields and programs are going to relate to each other and how things are being calculated. So if you had, for example, a 5250 screen, or maybe you had a table, and in that table you have a calculation field. That is the summary of attacks, a quantity, the price of the component, maybe an insurance program that’s doing some kind of adjustments as well.
These are all calculations and inputs going into that field. Well, you want to know how does that work. Well with data lineage you can actually go to that field and expand it out. And you can see all the inputs that are going into that field. So for example, if it was an order entry and at the end you had a total, when you look at that total it is being calculated through quantity price tags all getting summarized.
But it’s more than that. As you expand this out, you get this visual representation of the flow showing the program where in the program and what the actual line of code is. That is calculating that step in the overall calculations.

R.B. – Now, correct me if I’m wrong, but what you’re saying here is this thing is literally tracing data as it moves through the application, showing you where and how it’s transformed in that application.

A.A. – Exactly. So if you were to change one field in a program, let’s say quantity, you can literally look and see how that’s going to impact everything else. Because quantity may be set up in a program to default to zero. Maybe it’s defaulting to five for whatever case. But you can see how those are put together. But even more than that, between the fields in the program that are being interacted on, we have a little hover button. In that button you can actually see the calculation that is taking place and all the information about the program that is handling that.
And if you’ve modernized some and you actually have procedures defined in this program, you can actually see the entire procedure that is impacting that field. And you can open it up and you can actually view the source code for that to see how it is going to go through. So let’s say you had a project and you needed to expand a field.
You could use Transformer field for this. That’s a great service that ARCAD offers that can help you go through and expand a field, because that’s not easy. But maybe you’re curious about what that scope is going to look like as you’re beginning this. Well, once you look here and you can look at one field in a cell, what’s the impact here all of a sudden you’re looking at 15 things that need to be changed, because those are all the driving inputs into this.
And you start to look through those and maybe you can start building a scope before you reach out to ARCAD and say, okay, we need Transformer field in the services to help us with this, because this is going to be a monster. But you start to understand how this is all put together and what some of the real impacts are for this.
And again, this diagram, this visual representation of this can be saved off and included in any documentation around any of the changes. Maybe you’re building a use case now you’ve got a description and a visual representation of it saying, okay, we need to change this. This is where it’s going to be, because ask any developer to go find a field in all the pieces of code where it’s being referenced.
They’re going to miss something. It’s going to be hard.

R.B. – This has to be a huge time saver. What I mean, what you’re saying to me basically is a developer can start with a field in a table and instantly see how that field gets calculated by the information system.

A.A. – Exactly. And going through that, that can maybe then starts to layout the scope. And right there you maybe saved an hour of digging through all the pieces of code to build this out. Now you can look and say, okay, that’d take me about 20 minutes to actually make these code changes. Done. versus an hour of looking through all the pieces of code that need to be changed.
Now, it’s not going to tell you exactly which line you need to change. You still have to be a developer, but you can look through and say, okay, it’s going to be on this line. Oh, but it’s also referenced on this line. In this line that way you don’t miss anything.

R.B. – All right. Correct me if I’m wrong though, but this can also help with compliance and regulations like GDPR and HIPAA. I mean, those things are focused on the traceability of data. And that’s exactly what data lineage is doing, isn’t it?

A.A. – That’s a great point. Because I wasn’t even thinking from that perspective. I’m cutting down in the technical weeds of things from an audit perspective. Audit says, hey, how are you calculating this? There you go. Now you have your visual representation showing you all the pieces of information that go into it. Health calculated.
You hand that to an auditor and they’re happy because now you’re saying this field is being handled this way and this is where all the inputs are. You can look at those programs and say, yes, we have; look at your software and say, oh, yet there’s safeguards in here, but it can show you all the programs that are involved through this.
It’s a little bit more, it’s not as truly in-depth as maybe diving into the cross-reference and going through those methodologies, but sometimes the developer is not around to do this. You need the business analyst or the program manager to be able to look at this because you want your developers to develop. A business analyst can pull this up for an auditor and go, here you go.

R.B. – So we can trace how data is computed in an information system. We can work with retro documentation. We can build architectural diagrams up that application. I mean all of this is getting you ready for an audit. So this is obviously a tool that you can use for audits.

A.A. – Yes. And this all comes back to the producing and showing the return on investment: that ROI. And this starts to lower that cost because you can produce information faster and more effectively.

R.B. – So what I’m hearing here is that this tool can be used by your developers during a modernization effort. It can be used by your business analysts so that they can quickly, at a glance, understand how things are being worked on and so on. It could also be used by CEOs within the organization to make sure that compliance and things, regulations are all adhere to.

A.A. – Right. Because it’s right there in front of them. They have access to it. And, we’ve talked now for 10 or 15 minutes on this, and we haven’t even touched on some of the AI functions that are really going into this. And so I think we’ll need to save that for another day, another discussion down the road when we start getting into how we’re breaking that down and how AI is coming in to not only Discover but other ARCAD products as well.

R.B. – All right. We’ll definitely do that. And thank you for your time today, Alan. I really appreciate it.

A.A. – Thanks.

R.B. – So I’ll summarize just a little bit. ARCAD discover is an AI assistant for visualizing, analyzing and documenting applications. During this discussion, we talked a little bit about functional mapping and reverse documentation. We talked about visualization of architecture, data flows and dependencies. We also talked about how it can deal with governance and compliance issues. All of this can cut your analysis time from weeks to minutes.
It allows you to make smarter decisions about your application and allows you to do it with more confidence. We just started to scratch the surface on some of its AI features, so we will be doing another techtalk on this and covering those AI components. But for now, thanks for listening.