[Israel.pm] Class::DBI Talk

Shlomo Yona shlomo at cs.haifa.ac.il
Mon Jun 14 09:39:18 PDT 2004


On Mon, 14 Jun 2004, Yuval Yaari wrote:

> 1) CDBI in general, how it saved my life, things in theory
>    (relationships, etc). To make a long story short: less code.

i'd love to read it on the mailing list.

> 2) Hands-on: how do I make a CDBI project? How do I define my DB/tables?
>    How do I create relationships? More code, less talk.

This is more interesting to see in the lecture.

> Having 0 (z-e-r-o) SQL is great.
> Database changes that doesn't necessarily "hurt" your code is devine.
> Not writing SQL makes development time so much shorter.

If you can demonstrate these issues on a hands-on lecture it
will be very useful.

> Anyway, I guess Gaal will make you understand the "why use CDBI" part :)

Gaal? Can you explain why?

> Tell me of everything that you would like to know about it (I am no CDBI
> guru, but I guess I could dig stuff up if I don't know them) or forever
> hold your peace.

1. What is the roadmap to follow when you have data
structures that you'd like to query or maintain through a
database? How would you map them to database tables? Are
there things to keep in mind when you do that?
2. Given a mapping between your data structures and your
database tables, what sort of interface to the database is
preferable?
3. How to maintain (read: survive) changes in your data
structures? What can you do in advance in order to reduce
the amount of code to be changed in such cases?
4. What is the roadmap of dumping a database full with data
back into your original data structure(s)?
5. How do you hide your SQL? Where do you end up writing
your SQL after all?
6. When doing a project which involves a database, what
things hace to be known up front? What things are best to
workour later?
7. Are there any portability issues? How one can handle
them?

I think that one can relate to these in a lecture in a
theoretical way or in a practical way. I'd be more
interested in seeing the practical side of things, having
theory only mentioned briefly, so that we end up seeing real
engineering decisions and implementations rather than
theorms.



-- 
Shlomo Yona
shlomo at cs.haifa.ac.il
http://cs.haifa.ac.il/~shlomo/



More information about the Perl mailing list