RE: agility programming and DBA's?

I'm coming a little late to this thread, so I'll possibly summarize some stuff that has been said. But I have some personal experience, in that I managed a project where we used some aspects of Extreme Programming.

First off, as people have said, there are various versions of this type of technology, the two best known being Agile Programming and Extreme Programming. They share some principles but have some differences. Frankly, in the beginning, I thought that both were probably invented as an apologia for doing a project which didn't have any formal specs or requirements to begin with (both coming from the java community), and to do some damage control and actually get working software. There is a measure of that, but, when I put some of the things into practice, I found that they really worked. =20

However, and this is another bug-a-boo (?) in the java community (if I can over-generalize): they fail to adequately differentiate between analysis (analysis? What's that? You mean I actually have to spend some time upfront before I start churning out code?) and detailed design/implementation. David Hay has written extensively on this as regards to data modeling and it's also applicable to application design. Before I started my detailed design/implementation phase, we had been through an analysis and requirements gathering phase so we had a baseline for what we were doing and approved functional requirements.=20

The things that I found worked best from Extreme Programming were: some form of pair programming; living with the customer; and frequent customer deliveries and turnaround. Pair programming is quite problematic in terms of who gets paired with whom. (Mentor and student? Two of equal abilities?) But living with the customer and frequent prototypes and iterative development worked very well. HOWEVER, the process has to be carefully controlled; if you can't eliminate scope creep, at least you can note it and possibly get paid for it (either with more $$ or more time). To do detailed unit testing (as in "Write the test first") you really need a very strong management commitment to allow you to do that. Steven Feuerstein (sp?) has written extensively on doing unit testing in a PL/SQL environment, and Ken Atkins had a wonderful presentation at last year's ODTUG conference.

BTW, this year there is a presentation on Agile Programming in a Data Warehouse environment slated for ODTUG. Looks to be a good paper, practical experience in terms of what works and what doesn't. The author says that one of the things that works best is, in fact, daily "huddles", where everyone reviews their days, but keeping it short, less than 5 minutes per person.

But, as you can see, this was all in a development shop and it's a methodology for "programming", so it's unclear to me precisely what the new manager may have in mind. The daily huddles? Pair responsibility for various projects? I'm curious.

Leslie
Leslie Tierstein
Senior Consultant
Vision Chain, Inc.
The first software to power the demand data network phone: 202-459-4129

-----Original Message-----

From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Paula_Stankus_at_doh.state.fl.us
Sent: Thursday, May 26, 2005 7:24 AM
To: mark.teehan_at_csfb.com; oracle-l_at_freelists.org Subject: RE: agility programming and DBA's?

Overheard a rumour that a new manager whom I think has not managed before is considering using Agility Programming as his management style. What is Agility Programming - does anyone know and how might this work in a DBA shop or not?

--

http://www.freelists.org/webpage/oracle-l
--

http://www.freelists.org/webpage/oracle-l Received on Thu May 26 2005 - 11:25:59 CDT