WebSphere Application Server & Liberty

WebSphere Application Server & Liberty

Join this online group to communicate across IBM product users and experts by sharing advice and best practices with peers and staying up to date regarding product enhancements.

 View Only

Don't look back… or should I?

By Laurent Barthelemy posted Wed May 11, 2016 06:06 AM

  

While thinking about my first blog note, I thought it would be appropriate to have a kind of introduction.

I entered the world of Data Integration in the heydays of EAI, some 20 years ago. At this time, I was a C programmer in a pharmaceutical company, and I shared my time between writing GUIs for the manufacturing systems, and writing interfaces to exchange the little information we had to share between the mainframe, where the main systems were, and the VAX-VMS system, where the manufacturing systems were hosted.

When I was maintaining an interface, most of my time was spent counting bytes to make sure the data array I received was properly split at the right places to extract meaningful information. This was a tedious job, and when I found a bug, I had to get back to the initial message, compare it to the information in the master database, in order to determine if the issue was in the message I received or in my interpretation of it.

So, when some strategic decisions established that the mainframe was to go and be replaced by a new ERP system, we had to rethink the architecture, and we soon discovered that integrating SAP R/3 using some custom C code may not be the way to go. This was a few years before 2000, and this year was indeed the reason for this deep change, as the Cobol programs were not Y2K compliant, and the effort to make them reach this level was deemed worthless.

So we had to choose a software to help us achieve our goals in a more efficient manner, and this is when I discovered Mercator. The company was just entering the French market at this time, so they had little references to show in our field, and we decided that the best way to confirm the software was what we needed was to get a training, combined with some technical workshops where we would validate the main aspects of the architecture in our own landscape. I was really impressed that these guys were so confident that they accepted to install their software in our own systems, and to build the first interfaces in a few hours, without even having had a look at our specifications.

Fast forward to 2001, when I had become knowledgeable enough on Mercator to join as a consultant when I decided to change jobs. What I did not really expect was that I was already so hooked in the software and in this particular field of IT that I would still be a consultant in the same field, mainly working with the same piece of software, 15 years later!

It still puzzles me to see that the concepts at work in the early versions of Mercator would still be valid 20 years later, in what is now called IBM Transformation Extender. Of course, there has been many additions, improvements, new standards taken into account, but the basic objects are still built on the same concepts and most of the time you can even open a very old source file in the latest version, and get a working map.

So I'd like to tip my hat to these people who designed the software that makes me wake up every day, knowing that I'll have an interesting day, and that I'll most likely learn something new in a field I discovered 20 years ago.

I hope to be a regular writer, but don't be surprised if the main subjects are around Data Integration, and more especially ITX. I'll be delighted to exchange with you on these subjects.

0 comments
7 views

Permalink