VXTK evaluation:
This is an evaluation of I|Nova, VXTK's proprietary software product. It is similar in nature to other software product evaluations I have posted on SI in the past.
One of the major investors in VXTK became aware of the other evaluations I had done, and requested that I perform the same service here. David Hite, CEO of Veronex, in response to the insistence of the major investor, was kind enough to invite me to a demo of their product. After a couple of misfires with regard to scheduling issues, I finally got an up-close-and-personal view of I|Nova at Veronex's development headquarters in Santa Ana, CA. The following text documents my experience, and my opinions on their product and its applicability to the software market.
I received no compensation for doing this. Veronex paid my air fare and bought my lunch. It cost me a tank of gas for a roundtrip drive from Sedona, a parking fee at the Phoenix airport, and a couple of days' time. I spent one day on the actual demo trip, and another day's time spread out over a couple of days to write the evaluation. In the interest of full disclosure, I did receive 1,500 frequent flyer miles on America West.
I'll be as objective as I can, and as honest. I have no hidden agenda here.
Background: I was a career software developer for 25 years (now retired) and helped visualize, architect, develop, implement and support software that is running IBM mainframe data centers around the world. I know what I'm talking about, and I have a successful track record to back up my claims.
Disclaimer: These are my opinions. Not facts. If you use this information in your investment decisions, you do so at your own risk. That should satisfy any legal issues.
I did not sign a non-disclosure agreement with Veronex. I was told I could write whatever I wanted to. In fairness to them, I will withhold certain information that I think might give their competitors an unfair advantage if disclosed.
I own no shares of VXTK now, have not owned any in the past, and don't intend to purchase any in the near future.
OK, enough of the background info ...
My visit started out with a quickie tour of the Veronex development headquarters. Relatively small, but they have a clean and professional appearance.
I was introduced to Tom Price, the brains behind the I|Nova concept. We exchanged the customary greetings, and then he gave me a synopsis of his life story. This was, I'm sure, to let me know that he has relevant experience in the IT industry.
He then gave me a PowerPoint slide show presentation that they typically give to management types. It explained the problems faced in IT departments, and how I|Nova approached those problems. It also discussed the Y2K glitch and claimed to solve it.
I asked Tom how many I|Nova systems had been successfully implemented, and he deftly avoided the question. The response was something like "our teaming partners are finding a lot of interest".
After the "management overview", three of Veronex's techies came in to do the actual demo process. Included in that group was Tony Speed, their VP of Development. I got the impression that Tom Price doesn't keep up to date on the day-to-day changes being made to his baby. It was good to have the low level techies around so I could ask the nitty-gritty questions that came up.
I also got the distinct impression that I|Nova is a conglomeration of a few different toolettes. The main conceptual piece is what Tom Price had cooked up over a period of several years while servicing clients, while the analytical pieces for actual migration processing were acquired and made to fit the overall game plan.
VXTK's product has been much maligned in the past on this SI thread, and poked fun at on others. This was a result of some overly enthusiastic claims made in press releases and on the Veronex website. Tom was aware that their marketing approach was incorrect, and mentioned that they're working hard to correct it.
I|Nova's underlying concept is that all computer applications consist of basically the same functionality. Thus, there should be a way to share those common functions.
This concept is absolutely not new, although Veronex tends to promote the idea as something totally revolutionary. The concept was in vogue several years back when the "R" word was popular .... Re-engineer. Everybody jumped on the band wagon to produce "Re-engineering" products. The idea was to analyze existing code, determine what it does, then re-engineer it to be more maintainable and extendable. Tom insists that I|Nova is the only product that does this magic at the enterprise level.
Basically, what the I|Nova process does is what all the other guys do ... create an inventory of software portfolio entities, analyze them all, and determine which entities have which relationships with which other entities. After a humanoid determines what to do to enhance those relationships, there is an automated process to recreate an entire set of code to support the updated applications. The Y2K enhancement is just a specific instance of this whole process ... just as the Y2K support from several software vendors is nothing more than a specific instance of their capabilities.
The above process is done by several of the "language and/or platform transformation" tools, produced by companies such as Alydaar (ALYD), Consygen (CSGI), Forecross (FRX.U), Peritus (PTUS), etc. Interestingly enough, none of those companies were mentioned as competitors. They think their main competition is a product named "Opal" from Computer Associates (CA). Personally, I think they're way off base with that assertion. I found it surprising that they didn't even recognize a few of the company names listed above.
I|Nova works in two "modes". One mode is as described above ... code in, code out. This is the way most Y2K fixes would be done.
The other mode is the "codeless" environment. All the data element information is extracted, existing screen definitions and file formats are captured, and then they're all defined to the product through a series of panels. They have an "Application Manager" which uses all that information to control the application representation to the end user. In this mode, the user manages the process, not the code. Terrific concept!
Unfortunately, the logic from the code isn't captured during the analysis phase, so all the business rules need to be redefined to the system. This is a major shortcoming, and is why the various "Business Process Reengineering" tools from way back fell short of their desired goals.
Also, the "codeless" phrase is a bit misleading. Every process needs conditional logic. For instance, "if payment is late, add a penalty". The way I|Nova implements conditional logic is by providing their own "language" to write the "if" statement in. That language is nothing more than their own version of "code". The syntax bears a resemblance to Fortran. Anybody who has seen Fortran knows that life for an I|Nova maintainer isn't much fun.
The claims being made by Veronex press releases that the tool is "Fully automated" are outrageous. Those are some of the claims that got Veronex in trouble with the techie types on the internet discussion boards.
The tool does do analysis of existing code. It finds data elements and where they're used. But, a human has to sit down and decide which changes to make, and then make them happen. Certainly, the tool is far better than performing the same activities manually. At best, I|Nova is a tool that automates some of the processes necessary to perform maintenance functions. Fully automated? No way. Note that there are no tools in the marketplace that do that. And there never will be.
Various observations made during the demo:
Tom Price said the tool automates 60-70% of the work that is required to migrate from a mainframe to a client/server environment.
According to the FASB accounting rules, maintenance of existing code is an expense. According to Veronex, since new code is produced, the cost is considered an investment.
Instead of using 5 average programmers to maintain an application consisting of code, the I|Nova "codeless" environment only requires one senior person with knowledge of the application's business process flow.
Tom Price said that a strong point for the tool is the ability to standardize and renovate COBOL code. This was an interesting comment, and is in direct conflict to some of the earlier claims made. The ONLY language supported is COBOL. Earlier claims made it sound like ANY language could be converted into the "codeless" environment. I suppose that's a true statement, as long as you didn't mind having to write the analysis piece that strips out the data definitions. But, if you have to do that, why buy their product?
The tool reportedly does a good job of standardizing data definitions and identifying common code segments. This theoretically reduces the amount of time it takes to maintain the code each year. All identical code snippets are made into copy books. The same is done with data definitions. An attempt is made to determine which similar file definitions are referring to the same file, and a copy book is made of them, too. The source code is then updated to refer to the new copy book's definition.
Source library managers are not supported. All the code must be placed into a standard MVS PDS. This is an operational nightmare. Other competing products have the same problem, but some of the more mature products have provided native library manager support within their product. This alleviates many problems.
The code analysis and code manipulation processes are all done on an NT workstation, so the user's code portfolio has to be downloaded to the workstation from the mainframe PDS. Various utilities have been provided to ease the pain of that effort, including the ability to FTP the files.
The Y2K dataname identification facility is somewhat limited. Only basic search criteria is allowed. No dataname masking is provided. This needs to be beefed up to compete with other products.
I would expect the product to run into a performance degradation issue as they run into larger applications. The demo I saw ran against about a dozen programs. Many applications have hundreds of programs in them. That would make some of the dataname lists extremely long ... and building the scrollable Windows would take a lot of time.
They seem very proud of one particular site where they converted a Wang application supporting lots of end users to the "codeless" mode. When pushed for details, it turned out that it was both a product and services deal. You can cover up a lot of product shortcomings with manpower.
There have been several comments made in press releases and on the discussion boards that the I|Nova system can be used to migrate an application that is missing source code. This one confuses me. If the tool analyzes source code to find the data definitions and data usage points, how will the tool help if the code is missing? I think what they mean is that the tool can be used to define the business process that is being performed by the missing code.
Everyone I spoke with had high regard for their product and its capabilities. They really believe they have the world by the tail. From my perspective, they're playing in an ivory tower. They need to get out and see what the rest of the world is doing.
If you have any questions, please feel free to ask.
Regards,
TED |
