I'm going to organize a 'Legacy Code' Birds of a Feather session at OopslaOhFour
. The emphasis is on discussion about systems that are large, untested, and difficult to change; systems we have to live with but can't really replace. When we look at those criteria, it seems that there are as many, if not more, systems like that written in modern OO languages as there are in older "legacy" languages (COBOL, RPG, etc). I've just finished writing a book about this topic, but I'd like to continue to compare and contrast experiences on this topic.
I don't know which evening this will be at OopslaOhFour
but I hope to settle that soon and sign up for a room as the conference starts. If you'd like to express interest, please sign on here, or leave me a note on my page. -- MichaelFeathers
So how did this turn out? -- AnAspirant, eagerly reading the aforementioned book
Anybody listening here? I was at this session. It was a bit disappointing in that there were only a handful of folks there, four or five I think. The main problem was that this BOF was immediately after Alan Kay's scintillating Turing Award lecture, and I suspect that many people wanted to hang around and talk and debrief about that instead of rushing off to a BOF.
The BOF was encouraging in that I got to hear other people wrestling with problems similar to mine, some technical and others not technical (that is, people problems). I later picked up a copy of MichaelFeathers
, and I also found it quite helpful.
19 July 2005
What's frustrating is the people who are writing legacy code -- as we speak.
By this I think you mean it's frustrating that people continue to write code that has bad CodeSmells, which requires extra work to be refactored and to be UnitTested. Yes, this can be frustrating. But my job responsibility isn't merely to improve code quality, it's also to educate and improve the skills of people on the team. So for me it's just part of the job. What I really find frustrating is people who refuse to recognize that their skills need improvement.