Re: Software Eng. in Oracle: 2 questions

As I previously posted, there are virtually no metrics used for SQL projects. 4GLs are not as easily measured as 3GLs: I can't estimate the number of lines per code per programmer. SQL is messier: there are elegant solutions and inelegant solutions; both ways work and may use equal amounts of developer time. I have an ancient paper by Ulka Rodgers who suggested using CUOCOMO for estimating forms development: unfortunately, it was written during V5 and Forms 2.3 era. I have devised my own methodology of applying function point analysis, but considering the + or - 35%, I consider it to be no more than "best guess" estimate.

As to CASE: as a designer I have used various CASE tools (ORACLE and Excelerator) and I have designed without formal CASE tools. In my experience, I always ended up developing my own tools that in essence were CASE. I don't believe the learning curve is that long on ORACLE CASE, rather, I believe forcing designers to learn and adhere to formal practices that CASE requires takes a lllloooonnnnggg time. For many years ORACLE developers used "quick and dirty" approaches (and I am as guilty as the next person). The idea was to get something out to the user as soon as possible, so you would think about the screens, create the tables, develop the forms, and maybe doing a little testing, if time permitted. It is very hard to break old habits, and begin to develop a conceptual, logical, and physical model. Many folks out there still adhere to the WISC philosophy: Why Isn't Sam Coding?? CASE requires forethought, analysis, design, and re-design of models before you ever get close to developing a schema. Its a different way of doing business. But, in my opinion, a better way of doing business. Hope this helps.

Dagmar Anne Bogan

Kopania & Komorovski
Richardson, TX 75081

dagmara_at_delphi.com Received on Mon Sep 26 1994 - 00:13:52 CET