Re: database of maintainers

[Rob Browning <rlb@cs.utexas.edu>]

Manoj Srivastava <srivasta@datasync.com> writes:

> 	The data collected by the ongoing effort to collect maintainer
>  information is gathered an kept in a | delimited flat text file,
>  which is working quite well, at the moment.

Well, I'd advocate a scheme/lisp style database, but I'm biased.  It
has all the Pros you list above except one.  It's pretty hard to get
perl to do anything useful with it, but it's easy to parse from C, or
scheme, or emacs-lisp, etc, and it's tremendously flexible (if you can
stomach parentheses).

It has the added advantage of being able to handle new fields with no
effort, and being able to to handle heirarchical data.  For example,
consider (the silly example) that later we want to add a list of
maintainer food preferences:

((name "Rob Browning")
 (email (("Rob Browning <rlb@cs.utexas.edu>" preferred)
         ("Rob Browning <some-other@some.where>")))
 (location "Austin")
 (food ("Green Eggs" "Ham")))

Optional fields, field properties, etc are no problem.

Anyway, I imagine this is probably not likely to happen, but I've been
fighting with a system using a flat record (/etc/passwd) database that
should have used a heirarchical representation, so I've been thinking
a lot about this lately.

I ended up having to go with a less flexible representation because we
had to use perl.  In such a case, I'd say that David Welton's
suggestion is the best one (see later in the thread).

I'd really be wary of using anything like SQL.  In this case, I think
that's probably sledgehammering the mosquito.

-- 
Rob


--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word "unsubscribe" to
debian-devel-request@lists.debian.org . 
Trouble?  e-mail to templin@bucknell.edu .