----- Original Message -----
From: "Charlie Caro" <ccaro@...>
To: <IB-Architect@egroups.com>
Sent: Thursday, September 07, 2000 12:55 PM
Subject: Re: [IB-Architect] Considering changing dsql/dsql.c Opinions
please.
> Changing code is ALWAYS the last step of software engineering. If you skip
the
> due dilegence steps before that, it will be difficult to maintain a stable
code
> base.
I really wish I had the skill and experience to work with the Interbase
source. But unfortunately, I have basically no knowledge of C, or C+, or
even C- for that matter. Im great with Basic, COBOL(YUK!), and primarily
Object Pascal(Delphi).
For what my opinion is worth, and its probably not a lot, I think that the
top priority, and basically the ONLY priority at this time is getting the
build stable, and to compile easily by anyone with a good knowledge of the
language. Going through the code and fixing spelling errors in comments, and
fixing indentations is fine. It can't hurt, and it might even help one to
understand the code while doing it. But, I don't think that ANY functional
changes should be made to the code until the primary goal has been
accomplished. Once that has happened, then the code will be accessable to
many more developers, and thus there will be many more eyes looking at it to
find potential pitfalls of any changes.
I really do applaud any individual who wants to make it better, but I think
that making it better can wait a month or two, or whatever it takes to get
the build 100% stable. I also think that waiting a little will allow more
developers to become familiar with the inner workings of the code. Certainly
anyone who wants to experiment with the code should be welcome to do it. The
experimental code can be posted to Firebird, and marked "Experimental", but
it should not be committed to the "official" build until it is absolutely,
positively, 100% clear that it is not breaking anything. Guessing, or "I
think it's ok" is not good enough.
We should REALLY listen to guys like Charlie Caro, he has lived and breathed
this thing for years, and understands it far better than any of the rest of
us(outsiders). The code probably has a thousand flaws in it that have been
cause from years of different teams of developers working on it. We have all
been there. You add a chunk of code to fix a flaw in another chunk of code,
and so on and so on.... My guess is that these flaws are held together like
a house of cards.
Im not trying to start a fight, Im just posting my opinion, for what it's
worth. And of course, my opinions are based on the fact that I have not
studied the Interbase source, so I could be completely wrong on my
evaluation of it.