Sam Ruby wrote:
>
> Geir Magnusson Jr. wrote:
> >
> >> Smart data structures and dumb code works a lot better than the other way
> >> around.[1]
> >
> > As a general rule, that's all well and good, but the point is to provide
> > as part of the API an easy to use class to allow simple access to
> > repository information abstracted from the data delivery and
> > representation.
>
> Read what you wrote. In other words, you disagree with Fred Brooks, Eric
> Raymond, and myself on this topic. That's OK - we can agree to disagree.
<sigh>
Just to try and put a bullet in this, note a few things :
1) In reality, I would be disagreeing with Fred Brooks, if I was
diagreeing. You are just quoting Eric Raymond, who quotes Fred Brooks.
The actual quote is :
"Show me your flowcharts and conceal your tables, and I shall continue
to be mystified. Show me your tables, and I won't ususally need your
flowcharts; they'll be obvious."
An interesting twist would be if this changes and how in light of OO
domain analysis, which I guess is a better representation of 'tables'.
The more relevant quote to the problem, from the same chapter, is :
"The programmer at wit's end for lack of space can often do best by
disentangling himself from his code, rearing back, and contemplating his
data."
The quixoitic result of the XML suggestion in the context of the above
is that the space problem (code, not data in my case) gets worse :) The
properties rendition of the repository data and the XML rendition are
equivalent in content - you can transform from one to another and back.
The difference is that the properties version allowed limiting my
external codebase to the JDK, which I wanted to try to do.
And an ironic finish to this is that the section in Brooks' book with
the above two quotes starts with the sentence :
"Beyond craftmanship lies invention, and it is here that lean, spare,
fast programs are born."
lean... spare.... fast....
2) This started because you didn't seem to notice that I was *agreeing*
with scott that XML would be better, but was just noting what I was
focusing on with what I hoped was a humorous remark. I guess it wasn't.
3) My focus was on getting the Repository API right, abstracting away
the underlying delivery mechanism, which is irrelevant to the user.
Thirdhand quotes about smart data structures nonwithstanding, the right
idea was figuring out what information is needed to represent a
repository, and abstracting the repository information from it's
transport. XML is simply a [great] means to an end but not the only.
I chose a Properties because it's supported by the JDK, and that was a
design contraint I was working under.
geir
P.S. I do believe I will be able to use an XML after all with a very
lightweight SAX parser written with embedded systems in mind. small...
--
Geir Magnusson Jr. [EMAIL PROTECTED]
System and Software Consulting
Developing for the web? See http://jakarta.apache.org/velocity/
"still climbing up to the shoulders..."
