On Fri, Apr 23, 1999 at 08:46:01PM -0700, R Garth Wood wrote:
> On Sat, 24 Apr 1999, Craig Sanders wrote:
> > "proof by assertion" is no proof at all.
> Then I shall invoke my favourite: proof left to the reader.
That's usually "proof left as an exercise to the reader", btw.
And, as my physics examiner once wrote on an experiment report of mine,
"The reader doesn't need the exercise."
We already have lots of tools for manipulating text files. grep, sed, cut,
cat, tac, wc, more, sort, uniq, tr, col, diff, column, patch, and that's
not even mentioning awk or perl. vi and emacs and stuff all cope with text
files *wonderfully*.
And yes, you can rewrite these, or write replacements, or whatever to do
the same with a database. Sure. Great. But it doesn't count until you've
actually done it. Not when the people who like text files can point
straight at a couple dozen tools that cope with it all easily.
If you're willing to put some code where your mouth is, come back to us
with all those tools written.
Next thing. SQL just doesn't cut it. SQL is a databasey thing --
most admins just plain don't know it. That's fine, maybe they should
learn. Databases can do some pretty gee darn nice stuff. But they don't
now, and, personally, I think vi/emacs and grep/friends are much more
effective anyway.
Finally, I, personally, am not going to help you do this. I quite like
the way things work now, and I have some suspicions that doing it as a
real true database will just end up not working. The reasons have already
been given by others and rejected by you, so I'm not going to bother
explaining them, but suffice to say, I'm not convinced.
As such, you're not going to be changing netbase for a while. I suspect
there are quite a few others who are in a similar position, and you won't
be changing most .deb's to do databasey configs, at least for a while.
That said, maybe I'm wrong. And I am, or at least, I like to think I am,
willing to change my opinion. But in this case at least, you are *not*
going to do it by arguing on -devel about it. You're going to do it by
writing some code, and demonstrating that, hey look, you're all a bunch
of idjits, it *does* work, just like I said.
You can call it "Debian ITYS/Linux" [0] or something, if you like. Then,
when you can actually use it to, say, serve a web page, we can think about
merging it into Debian proper, or just giving up on this whole Unix thing
and going straight to DBix, or something.
But until then, well, *sheesh*. Show me the code or get out of my way.
Or something.
But can we at least drop the eternal "a registry/database is good! text
files are lame!" debate. Please?
Cheers,
aj
[0] Debian I Told You So Linux. I reckon that'd make a good name for
a proof-of-concept fork. :)
--
Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. PGP encrypted mail preferred.
``Smart, sexy, single. Pick any two (you can't have all three).''
-- RFC 1925, paraphrased: a guide to networking in the '90s
Attachment:
pgpdPTPc8pgzD.pgp
Description: PGP signature