<< Previous in series     Next in series >>

Paolo is back with some more teasers about ABS (ATI) Framework. But I am sure of one thing after reading his explanation that such framework exists and somebody out there is using it.

Thankfully Paolo is not done and there is another article that he is going to post giving more details about it. Here is what he has to say about framework

---

Wow, I never realized a small discussion can spur a great interest about ABS. Thanks Neel for opening this thread. ~grazie~ now I felt obligated to say something…

I have been running on support management and enhancement since my last Siebel development 3 years ago. So I am probably less technical and less qualified to talk about ABS from the technical aspect. However I could share the outcome and management experience. A year ago, we had two of Siebel enhancements starting around the same time. Both are equivalent scales but targets for business audiences.

To create competitive environment, one was given to a boutique consulting house (Team A) other than the current big consulting firm (Team B). Appointing two system integrators did not create conflicts as we have two Siebel instances. Since this forum is created for knowledge sharing, it’ll be very improper to spread marketing message and I’ll never reveal the consulting firms here.

Here are my findings:

Size of Team
—
The Team A employed 2 developers (config, integration, workflow, scripting) and 1 of our in-house engineers. The Team B mobilized a football team of 11 developers (config, integration, workflow, scripting).

Since both enhancements were fixed-cost projects, we didn’t really care much about the head counts anyway. However as we learned from the truth at the end, less head counts doesn’t mean less productive and less efficient.

Testing
—
The Team A started the test phase ahead of 1 week and finished all targets ahead of time. In total we gained extra two weeks before the planned schedule.

The Team B reached the test phase on time but failed terribly. The test phase was extended for another month. From the Mercury tracking, both teams produced the equivalent number of bugs. However it was the Team A which was able to produce the quick turnaround. The moment the bug was reported, the corresponding fix was patched and ready for retest almost with no waiting time. Sometimes the errors were fixed even before they were lodged. How did they do that? Who knows!

Without the tremendous results from Team A, the performance of Team B would be regarded “business as usual”. Obviously Team B was under a lot pressure and their fixes were rushed to compromise quality. Sometimes they fixed something by breaking another. There isn’t any shortcut in quality, thus the test phase was delayed even further and the deployment was delayed by 2 months (this was considered not bad already in typical Siebel project if anyone out there was laughing ).

Overall the UAT team was pretty happy about Team A and high marks were given to them in survey form. When we asked them why they performed so well in development and testing with such a small team, the answer was the ABS Framework which enables a new perspective of development. I didn’t ask more questions as I thought it was just like other Siebel modules such as Data Validation or Haleys.

Documentation
—
Initially, we thought Team B produced better documentation than Team A because they simply produced more. When you see a lot of Word documents and PowerPoint slides, you felt the deliverables were tangible. However, we quickly discovered one problem from the documentation package of Team B – it was not reflecting the current running application. What is the point if the documents were not updated with the actual current Siebel system! At one point, few documents were so different from the system and not usable at all.

Team B produced only 2 documents: Designs & Technical Spec and Usage Manual. A lot of graphical illustrations and most importantly, they are 100% accurate upon rolling to production environment. Since only 2 documents need to be maintained, it didn’t take much time to update. Well, how do we keep the documents updated in the subsequent post-production fix and subsequent development? This has something to do with the self-explanatory logging in ABS Framework. After running support for few months, I began to understand why the logging becomes a critical source of information for accuracy.

Now I dare to say that my documents are as current as the running application. Are they important? Without a doubt! I have less problems of training new support engineer who are now more well-informed. I have more confident in estimating the subsequent enhancement without wasting time of letting developers figured out in the blind. The application knowledge becomes transparent and transferable.

Money, money, money
—
I never heard of one major Siebel implementation which was done below the initial budget. Money was never enough for Siebel and it was like a slow poison. Is it Siebel’s fault? No, it is the matter of how to do it right. If a project employed less head counts and spent less time, of course you used less money. Any savings produce profits nowadays. If you read here by now you should know which team is favorable. In the time of financial crisis, getting things done with less money is not only a golden rule, but a universal law.

From what I learned Siebel in the past, Siebel is generally not difficult to learn, but it is the experience that counts, especially proven experience. However if there is something beyond good experience, I would say creativity. Creativity brings change like what Obama always says. We need change in Siebel community. Whoever out there created ABS Framework, should be encouraged to do more. I am NOT ready to give endorsement to ABS, not until I witness the results of next enhancement starting by next week. Two victories in a row will pretty much tell how ABS performs consistently in large implementation. Four months later, we will see.

At last, the late Buon Natale to everyone!

I will express few of my concerns and complaints to ABS in Part 2. Continue…

<< Previous in series     Next in series >>