Tim Rowe-IBM
Corby James-Clearblade
John Rhodes – CM First
Marty Acks – MKS

Our second round table will be moderated by Peter Vigliano of ADC Austin and will feature the following individuals:

Tim Rowe-IBM
Corby James-Clearblade
John Rhodes – CM First
Marty Acks – MKS

JOIN the round table:

• Please use the following number:
+1 (646) 396-6228
• Enter the participant PIN code: 49768244 followed by #

Please note:  This conference is being recorded for our podcast, iIDEAS for the Modern i.  By joining this meeting, you agree to be part of our podcast.

June 1 – Preconference Training

June 2,3 Conference

Peter Vigliano:
Ok, thank you Bill. Welcome to our first monthly Modernization Round Table. Our goal in hosting this event is two fold. First, we want to give you, the CA and i community the opportunity to speak directly with a panel of Modernization experts that we’ve assembled and secondly, as a company that specializes in Legacy Application Modernization, we’re very interested in what you’re thinking about. We’re interested in just hearing what the buzz is out there within the companies themselves, but I want to assure you that nobody will be contacted about your participation in today’s event, unless you specifically ask for a follow-up communication. So, we want these to continue and effectively to be somewhat anonymous. The purpose is really to try to help each other out on this. We only have scheduled an hour so I’d like to move right into the introductions. We have four panelists this month. The first one is Chris Smith. He is a Technology Architect. Chris has about 25 years experience in Application Development as well as Systems Application and Technology Architecture. He’s worked with the IBMi for 23 years, as a Developer, System Administrator and Architect. He has about 22 years of 2E and 16 years of Plex and 11 years of Enterprise Java Architecture. The second panelist is Daniel Cruikshank. You’ll see his name up on the screen and he’s an IBM Application Performance Specialist with DB2 and since 1999, Dan’s primary focus has been on DB2 and SQL Optimization and Modernization. Besides continued Performance Analysis, he also teaches the IBM DB2 for i, SQL and Query Optimization Workshop held in Rochester, Minnesota and locations throughout the world. He’s considered by many to be a trusted Advisor for the IBMi Community. The third panelist is Eric Jacobs. He’s a Technology Architect and is an employee of ADC. Eric has extensive experience in Code Generation and Transformation. His experience has been a major factor in ADC building the tools needed to transform 2E models into Plex models. Beyond that he also has a great deal of experience in Parsing, Symantic Analysis, Object Model, Processing Data Modeling and JBM instrumentation. We also have with us today Marty Acks. Marty Acks joined MKS in 1998 and plays a strategic role in defining the companies product direction for MKS’ i Series products. Mr. Acks is an authority in the area of change and Software Configuration Management. So these are our four panelists. What I wanted to mention is that I have on the screen my email and my cell number if you want to text in or email in a question because I’ve already received several questions via email. So apparently that may be a route that some folks would like to go and also, there’s Q&A on the screen as well. I believe everybody has the availability to use that. So I’ll be monitoring by email and the screen for those. At this point, you know, this is designed to be an interactive session, so we hope that you’ll ask a lot of questions and I think it would be great, if you have a question, to introduce yourself and let us know if there’s a certain individual that you’d like to address your question to. So the floor is open and I see we have a number of participants. So the floor is open if you’d like to ask a question. Anybody have a question that you’d like to start with or would you like me to read one that came in over email? The first question I have, it looks like this would be for Daniel. The question is, “How do we bridge an existing application to SQL using Rational Open Access?” Daniel, can you spend a few minutes on that?

Daniel Cruikshank:
Yeah, I certainly can. I could probably spend the whole meeting. Thanks Peter, that’s a good question. As you may or may not be aware of, Rational Open Access RPG addition was introduced last year. It’s a product that is an add-on to the RPG Language Compiler. It basically provides a key word that you would put on the Fspec and once you’ve applied those key words to the Fspec, basically the IO associated with that file is now going to be handled by your program. You are going to intercept the IO, whether the IO is to a display device or whether it’s to the data base and now you’ve become responsible for handling that IO. Where we see this being a big player in database, is the fact that many of our customers, who have grown up on this platform, started out on system 38 or system 3X have to moved to the 400 are now running on power, which we call the IBM i. We’re seeing that the traditional IO, that’s been around since the system 38 days, is starting to reach some limits. Especially as customers have grown in the last 30 years and we’re starting to process sets of data that are in the hundreds of millions if not billions of rolls of data. Frankly, traditional IO was never intended to handle this type of volume of data. The product that is most suitable for handling large amounts of data is the SQL product. The SQL is a set based product that is designed to exploit especially the architecture of IBM i., the single level storage, the built in relational database and to process large amounts of data. Now, one approach to take would be to rewrite all your applications in SQL. That doesn’t seem feasible, so on the other hand, what you could do… let’s say you’re running through a performance issue where you’re not scaling and you need to take advantage of some very large blocking capabilities, you could write an SQL score procedure that does a blocked fetch, for example, against the database and returns a large volume of data… and then using the handler you could intercept the first IO operation and call the handler. The handler goes out and gets a large block of data and starts returning that data one record at a time, back to your RPG program. That data is now sitting in memory and you’re running at processor speeds basically. You are running at today’s power-based architecture. You’re running at as much as 4.6 GHZ on these new processors. So you’ve eliminated the IO bottleneck that we’re starting to see with traditional IO’s. At the same time you’re taking advantage of the SQL engine, which around the turn of the century we totally rewrote this thing so that it’s integrated in the database itself. To give you an idea of what this means, is that if you had a chance to see Watson on Jeopardy and to understand what Watson was doing. Watson was fed an answer and came back with a question and basically blew the humans out of the water. The optimizer on IBM i uses that same form of technology. The difference is that the optimizer we call Watson Junior if you want, is given the question in the form of an SQL statement. Because it’s integrated into the hardware it can get the answer very quickly and bring that back to you. So that’s one of the reasons why we are very excited about this handler, this open access capability and we see it as a way for many of our customers to continue to scale on this platform as they start running into some of these ceilings.

Peter Vigliano:
Thanks Dan. Does anybody out there have a question to follow up on that with Dan? Okay, well Chris does. He’s one of our panelists. Chris is going to ask a question to you Dan.

Chris Smith:
Dan, does this product allow you to design any kind of backend program to the RPG IO, sort of like triggers allow you to attach any style of AS 400 program to the trigger?

Daniel Cruikshank:
Yeah, very good question Chris. Absolutely, you could write the handler in any available language that is supported on the i today. So it’s not required that you write the handler, let’s say in RPG. The interface is actually a parameter in the form of a data structure; as long as you can map that data structure you can write that program in your language of choice.

Peter Vigliano:
Okay, thanks Dan. We’re ready for another question. If you don’t have one I’ve got one that came in on e-mail. I’ve got several of them here. I think this one is probably going to be for Marty. Marty, the question is:
“What new techniques and trends are you seeing in the ways that companies manage development and how are these companies doing less with more?”

Marty Acks:
It’s kind of an interesting, broad topic. I suppose, to reiterate what Dan said, I probably could talk for a while on that. I’ll try to keep it a few minutes. A couple things that we’ve seen trend wise, you know, starting in the early 90s. Sarbanes-Oxley caused a lot of people to look at the way in which they do development. In kind of a forced matter if you will, that they had improved general processes and the approaches those folks took are all across the board versus manual techniques and they tend to focus in two areas. One is to have some formalized changed management processes. Some of that involved purchasing tools like we offer in the market. Some included, to be honest, just getting better documented procedures and doing what they say they are going to doing terms of process. So we also found as IT auditors came in, they tend to be all across the board in terms of saying “Yes, manual procedures are okay or you need to have this type of tooling.” So that started a bit of a trend about getting the whole Sarbanes-Oxley and IT audit. It continues on to the day that it’s not a servant. When we talk to customers they are not asking Sarbanes-Oxley questions anymore, they are asking… there is always an IT audit, “How do we do things better?” tone to what they are looking at, so we’ve seen a change in perspective in terms of it use to be in the 90s, people were concerned about checking code out and moving it into production and now they are interested in other types of questions. Why did I make a change? What was the reason for this, what was the justification? Should I have done this versus that? We are also finding that across both IT customers and people that do engineering type work which has pretty similar challenges. We are also seeing companies adopting techniques and technologies to get to the market more rapidly. So, for example, Agile is kind of one that has a certain buzz to it today. So we are seeing a lot of companies about that. We’re, as Bill Hunt had mentioned, the Agile Development…we have certainly implemented Agile and customers see that in what we’re doing and then we have compressed also not just things we build things internally, but how we deliver to market. So I think the days of monolithic large releases are kind of going by the wayside with a lot of vendors. If you’ve noticed Firefox just announced that they are going to have major releases. It’s all going to be Front releases and you’re going to see more and more of that where people are delivering software on demand rather than you know, wait a year, 18 months, two years… and you get this big monolithic thing. One of the challenges with that for a lot of the companies is that if you do that internally you have classic problems of by the time this big monolithic thing got delivered it’s actually not what you need anymore. It might have been close to what you wanted or needed two years ago but businesses change too rapidly. So we are seeing that and the final thing that we’re seeing is that people are looking at things in a less silo fashion. I came out of the ERP space…worked a lot with ERP tools back in the 80s and the core to all the ERP products at the time is that they were all tightly developed, tightly integrated, typically built by one development team for all the apps work. People are looking at things like that so they don’t look at their bug tracking as a separate problem from change management, from a separate problem as testing, from a separate problem for deploying the code. They are looking for something that is rather seamless and works in a very, very tight fashion rather than throwing things over a big wall as you go from system to system. So, those are some of the trends that we are seeing over the last few years.

Peter Vigliano:
Good… good answer. Would anybody like to follow that up with a question or want to hear anything more about that? Okay, time for another question. I’ve got one unless someone else out there has one. Feel free please to jump in and ask anything that you want to ask. Okay, the next question that I’ve got and I guess this would be for someone owned the ADC side. We have a large 2E model and we are interested in integrating the database with external applications. How could one approach this? Chris would you or Eric like to take that one on?

Chris Smith:
Certainly…you have external applications and more than likely these applications don’t necessarily live on your IBM i. So you have to start defining what the needs of your integration are. I mean, is it one directional, is it by-directional? Do you want to have all your data living in just one repository and how can you have your integrated application to get at this data at an acceptable performance level? As soon as you start dealing with copying data then you run into all the classic concurrency issues and if you’re sharing the data you have, you know, who’s going to be responsible for the business rules? You have various SOA type of solutions that you can bring to this with your 2E application. Out-of-the-box, 2E offers the ability to expose your 2E functionality as a Web service call, so that you can have your external applications consume these Web services directly and then the application integrity and data store is managed by your 2E application. Using things like the rational RPG open access you can get your 2E application to see control of some data to your external applications so that, for instance, your 2E application could be reading and writing data that was originally in an IBM I table. Well, it could be managed by some other application now and the rational open access then provides a traditional IO interface for your 2E application to the external data which can be through…for instance, if you’re trying to go SOA through a Web service call to your external application or to some sort of service bus.

Peter Vigliano:
Eric, do you have anything to add to that?

Doreen from CDW:
Can you hear me?

Peter Vigliano:
Yes, of course we can.

Doreen from CDW:
This is Doreen from CDW. We are currently doing exactly that. We have a monolithic, huge, Synon 2E model. Our database is stored on the i series and we have this major order entry type application written in 2E. So, what we’re doing is, we are taking the green screen and the action diagram and we’re taking the validation routines and making those external functions so that all the business validation rules are in one external function that can be called. A stored procedure can wrap it. The updates are a separate or an external program function that can be wrapped as well. So then we’re taking right now .net to do our GUI and then they are calling these external functions to do validation and updates. Then, you know, that could be any language. We’re doing .net but you could be…your GUI or your screen could be anything but your business rules, your validation, your updates are all in 2E and they are in the model. Then you can run parallel which is what we do. We keep the green screen, we plug-in these new validation and update routines so that we can run parallel with the GUI and make sure that we are getting the same results and it’s working quite well.

Peter Vigliano:
Good.

Chris Smith:
have you tried to get your 2E application to consume outside services yet?

Doreen from CDW:
Yes, we’re having no issues right now. We’re doing all kinds of different SQL functionality. We’re really kind of breaking apart the 2E green screen and making it more object oriented so that the screen is just for input and the business rules and everything are getting isolated…but still using the 2E model and action diagramming, you know, if a rule changes and you’ve got these two systems running parallel. You only have to change the code in one place and everybody absorbs that business rule change. So it’s kind of nice.

Chris Smith:
Yes, that does illustrate the point of that you… do get to re-factoring your legacy application so that your business rules and validations are separate from your user interface. Then, like you said, you can start putting different faces on your user application and they both execute the same rules and validations.

Doreen from CDW:
Right…it’s working good.

Peter Vigliano:
Good, Doreen thanks so much for your input. By the way, while you’re on the line here, do you have a question for anybody?

Doreen from CDW:
I was hoping…in that I think it’s still too young…I was hoping for GUI green screens within 2E not having to go to Plex.

Peter Vigliano:
What’s wrong with going to Plex? We need to have that discussion off-line maybe.

Doreen from CDW:
Okay

Chris Smith:
I remember those days when Synon was still a standalone company. They did have a product for they were building in GUI applications and the technology wasn’t ready yet…so it kind of got retired.

Peter Vigliano:
Doreen, are you coming to the conference in June?

Doreen from CDW:
Absolutely, it’s in my backyard. I live about 3 miles from the hotel.

Peter Vigliano:
Okay, well we’ll see you there. You’re going to hear some information that I think is going to be interesting to you.

Doreen from CDW:
Good

Peter Vigliano:
All right, thank you so much Doreen.

Dan Cruikshank:
Doreen, this is Dan Cruikshank with IBM. First of all I just want to say that you’ve made my day. Basically what you just described is exactly the direction that we are proposing to all of our customers…and to hear somebody say that “this is actually working” is always good to hear. One question I have for you is, what release are you on the IBM i?

Doreen from CDW:
Not the most current one. We are still working on upgrading. I think that they are having trouble with the WDSC or whatever…the Java part of it.

Dan Cruikshank:
Okay, well I just wanted to say that you are using SQL and I think you are going to be very pleased with some of the enhancements that you will find in the later releases. Especially in 6.1 and 7.1 and on the same topic of going to multiple different databases and whatnot, in 7.1 we have added three part qualification. So basically what that means is you can specify database, schema objects. Typically it’s a short cut form of DRDA but it’s also highly strategic. I applaud that you are using SQL and that you are going in that direction. So thank you very much.

Peter Vigliano:
Well good, thanks for speaking up Doreen. We’ve got 30 more minutes, which means we can answer a whole lot more questions. I’m going to open the floor again to see if someone has a question. Okay, well I’ve got one more. I’ve actually got three more but I’m going to ask them one at a time. Somebody wants to know about Agile and I think Marty this is probably directed towards you. It says “Agile development seems to have a lot of buzz…what is this really all about?”

Marty Acks:
Yeah, I touched a little bit on it before. Let me maybe attack this one by talking a little bit in a high level about our experiences. We started on an Agile path maybe about six months ago and the key reason we went to it… it’s about being responsive to customers you know. Again, that’s a software vendor speaking. So if I put it in terms of being an IT organization, it’s really about being responsive to your user communities. I would tell you that there are some very basic articles…I forget the guy’s names…Ken Schwaber and Jeff Settleman. So if you Google these guys there are a couple resources that you can go out too. There’s a scrum.org and there’s a scrumalliance.org. So those are the two semi-competing bodies that talk about this and that guy that found them sort of went off into a another group. So there’s a little bit of disharmony in the scrum community in that there is a group of people lined up to make more money than those purists. You can sort through that yourself. Some of our initial experiences introduced a lot of terminology and ways to think about things. It also forces you to slice and dice your work up. So at a high level you have these things… it changes your development and tries to crunch things into smaller pieces… So you create these things called user storage for example, that are small slices of what you can do. So for example, developers like to say “okay, I’ll have this big large thing done in a month,” and you’re forced to decompose things. So it’s very hard, it’s extremely hard for some types of projects and it’s easier for others. So you’re forced into decomposition. You’re also forced into responsibilities. So you have a concept that is called a product owner, which is a role that I describe as “one guy to shoot.” So that’s the person that decides what goes into this business application or in the case of a software vendor it’s a product. In a software vendor, in a software product company, these people come from what are typically called Product Managers. So I’m a Product Owner and a Product Manager here. In the user communities they come from a variety of different roles. They come from business analyst, people that have been in each community and maybe have an infinity for IT. They also come from people in IT that don’t necessarily want to do the hands on development on all the data and have a very strong infinity for the end users. So, it changes the roles a lot and it’s also pretty disruptive because it causes you to kind of do things rather differently. It also forces you to work and think in terms of priorities, not in terms of sets of things. So, you have priority one, you have priority two, you have priority three on the types of things you want to do. One of the core principles is that you have to finish the first thing before you start the second one. You also group your work into things called Sprints that are anywhere from 2 to 4 weeks and at the end of the Sprint you have to be done. There is a hard and fast rule that you can’t just leave a lot of residue, which is a common IT problem. So those are some of the characteristics. You have a lot of discussions about “okay, am I supposed to be doing this? Who is supposed to be doing that?” You also have a lot of people that will do variations of this Scrum methodology. The Scrum methodology is laid out in very specific terms. In terms of what certain groups should do and what others should not. From my point of view, we are in the middle of it, it’s a little disruptive and you still scratch your head sometimes, but it’s already shown benefits from my point of view. So I’ve become, in a relatively short period of time, I’ve become a pretty big fan of it actually. So those are some observations about having implemented it. The other thing is that I am talking to the IBM I community folks…you see that it talks about across all the different platforms… you see a little bit of it in the IBM I press but probably less than the other platforms. There’s nothing terribly platform specific about it. In my mind it’s appropriate for any type of development, IT, engineering oriented. It really doesn’t have any sort of language aspect to it at all or platform aspect to it.

Chris Smith:
I’ve got a question for you…you know, I’m an old-time RPG programmer originally and I got pulled into the Java world kicking and screaming. One of my biggest pain points were all these words that had no meaning to me and it took a long time to gain the vocabulary. When I look at scrum, it’s déjà vu all over again. Where can you get something that will take an old dude like I am to be able to you know, “oh yeah, this is what this is” because as soon as you open up something that talks about scrum they are using a jargon that is ancient Aramaic.

Marty Acks:
A couple of observations. I didn’t get this link to Peter. There is one thing that I printed all that is written by Scwabber and Sutherland. It goes through the terminology and it is actually nice enough to bold all the new terms when they introduce them. When you print it out it’s a fairly large font and it’s about 10 or 12 pages. So it’s easily consumable from start to finish. The other thing that I would say is that it is a… how would I say it without being to condescending? It’s a pungent rich area, so you have lots of people writing articles on this. Then you have people that like to talk about doing it right. So they all want to have their own variation and their own differentiator. It’s a little bit like SQL, in that SQL had this vision of being a universal language but then every vendor that implemented it has to do things to make it different so that it’s their SQL. You have the same thing with Agile going on…is that people want to say that we are doing scrum, scrum plus or scrum better. So you have to be aware of those articles in one sense. On the other hand, they are good practitioner articles and there’s just some stuff that doesn’t fit well with some companies. There’s a thing called a scrum Master in Agile for example. We’re having a heck of a time trying to find experienced scrum Masters in certain geographic areas to hire them. So we are kind of using managers for scrum Masters but that’s not exactly the way you’re supposed to do it…but it’s been working okay for us. So you have to have a fine balance between learning the…staying with…you know, if you get too far off you’re not really doing scrum but you have to have some variations for organization. It’s what everyone that has done this in other places have told me. Anyway, to answer your question, I got a little bit off track…I would be happy to send you a copy of this article. I found it most useful.

Peter Vigliano:
Thanks Marty… okay I’ve got another question that just came in and I’m not sure who we are going to address this one to. It’s either going to be Dan or Chris or Eric…it says “I’m interested in identifying and extracting existing business rules in whatever language it was written in… 2E, hand coded, COBOL or RPG. The initial purpose being to improve the quality of the code within the application… less duplication…more modular. What experience does the panel have in re-factoring code on the IBM I.?” there is a second part to the question and I’ll just read it also…the second part to the question, “what opportunities would I have to take those rules and push them to a new tool or language and what commercial tools could I look at to help automate this process?” Now, that’s a big question and I know we’ve got more than one answer for that, so who wants to start with that one? Okay, Chris.

Chris Smith:
Well, extracting rules from an existing application that has grown up over time is always a very, very large challenge. I have yet to find anything, that when you have that kind of mixed environment can give you 100% extraction of all your rules and migration. You know, that’s sort of the Holy Grail of modernization. The tools that give you various kinds of extraction but you’ve still got to roll up your sleeves and do some hardcore re-factoring before you can have a successful migration to your new platform. So, from an experience point of view, that’s been my experience in using design recovery tools.

Chris Smith-Technology Architect

Daniel Cruikshank-IBM i DB2 Application Performance Specialist

Eric Jacobs- Technology Architect

Marty Acks-ALM and Agile Specialist

The second episode of each month will be a one-on-one interview with an expert on the IBMi.

We look forward to having you join us for our round table discussions and hope you will enjoy our one-on-one interviews !

WEB CONFERENCE:

Go to: https://arkadinexpress.anywhereconference.com/, then use the following information to join the conference manually:
• Web login: 303000652
• Participant PIN code: 49768244

AUDIO CONFERENCE:

• If calling from within the US, please dial +1 (646) 396-6228
• Enter the participant PIN code: 49768244 followed by #

Click here for additional international access numbers.

Please note: This conference may be recorded. By joining this meeting, you agree that your communication may be recorded at any time during the meeting.

Date:  Wednesday, March 9, 2011

Time:  11:00am-12:00pm US Eastern Time

Our first round table will feature the following individuals:

Chris Smith-Technology Architect

Daniel Cruikshank-IBM i DB2 Application Performance Specialist

Eric Jacobs- Technology Architect

Marty Acks-ALM and Agile Specialist

Modernization is on everyone’s mind but we don’t always have an active user group to help us work through the maze of options.  That’s why we invite you to call in.  We hope you find the round table discussions to be useful and pertinent.  Our panel of experts will change from month to month so be sure to mark your calendar for the 2^nd Wednesday of each month.

Click Here To Join Live

https://www.livemeeting.com/cc/cai/join?id=KDWP56&role=attend&pw=ffx-%7E8M

Meeting time: Mar 9, 2011 11:00 AM (EST)   (Recurring)

Add to my Outlook Calendar:

https://www.livemeeting.com/cc/cai/meetingICS?id=KDWP56&role=attend&pw=ffx-%7E8M&i=i.ics

AUDIO INFORMATION

-Telephone conferencing

Use the information below to connect:

Toll-free:            +1 (866) 376-6162

Toll:                    +1 (660) 422-5140

Participant code:     3409730193

Add to my Outlook Calendar:

https://www.livemeeting.com/cc/cai/meetingICS?id=7ZJRNJ&role=attend&pw=N%28T6%5Et%3F%5BN&i=i.ics

FIRST-TIME USERS:

To save time before the meeting, check your system to make sure it is ready to use Microsoft Office Live Meeting.

http://go.microsoft.com/fwlink/?LinkId=90703

CA has announced that the 5th CA Plex/2E Worldwide Developer Conference will take place June 1-3, 2011 in the Chicago, IL USA area. Mark your calendars and stay tuned for further details.

For details on the conference, including agenda and presentations, please visit http://plex2e.com.