Re: The Art of SQL

ghaverla(at)shaw(dot)ca (Gordon Haverland) wrote:
> On Thursday 03 August 2006 12:46, Brad Nicholson wrote:
>> On Thu, 2006-08-03 at 08:59 -0600, Gordon Haverland wrote:
>
>> > I'm a little more than half way through this book (The Art of
>> > SQL), doing a book review of it for my local LUG.
>> >
>> > All in all, it looks like a very good book to me. I do think
>> > that it should be looked at by novice SQL people, if for no
>> > other reason that to try and keep certain bad habits from
>> > forming.
>> >
>> > For the novice, I think some other document which talks more
>> > about indexes and foreign keys is needed as well.
>>
>> That's not the books target market. If you read the section at
>> the start "Assumptions made by this book" - it assumes that
>> you're already pretty darn familiar with databases, SQL,
>> indexing, ect. I'd imagine that there are books target at the
>> Novice market that are more appropriate.
>
> I looked at that. I still think that anyone who is writing SQL
> will still benefit from reading it. Or, at least reading the
> stuff they understand. And with time, I'm sure they will have
> moments when they realise, "... that's what he meant when he
> wrote ...".

My preference would be to have some sort of "starting reference" under
my belt first... TAOS assumes you already know what a SELECT is, what
a JOIN is, what, in general, GROUP BY does, and such.

But I'm not sure it's needful to get to the "grizzled, wizened old
master" point for the book to become of value. It does quite a good
job of describing strategies for generating good queries rather than
horrific ones.

And while there is certainly a "knack" for writing SQL, I don't think
it's the same as some of the languages where there needs to be a
tutelage of fighting through writing ghastly stuff before reaching an
epiphany after which it starts to make sense.

For instance, in Lisp, beginners tend to write ghastly, awful stuff
consisting of fighting to pull apart lists, and accomplishing pretty
much complete destruction of them in the process. Forth beginners do
the same sorts of things with code where you can't see the logic
beside all the "stack smashing." _Good_ code in these languages has
very little of the "struggle to shove data structures to and fro;" the
experts know how to accomplish things without fighting with the
language.

I suppose an argument can be made that a "TheirSQL" model of the world
involves a naive view where "data smashing" takes place in the front
end application, but I'm not sure how far it would be good to take
that.

It seems to me that TAOS would be worth being exposed to fairly early.
I always preferred seeing *good* references to things, and competent
code samples. I've never been all that keen on C, but was never
disappointed to have learned it from K&R. They were very skilled with
it, and their examples demonstrate reasonably skilful use.

I think seeing some examplary code early is *way* better than plodding
through crud before seeing anything good.
--
wm(X,Y):-write(X),write('@'),write(Y). wm('cbbrowne','gmail.com').
http://cbbrowne.com/info/languages.html
Black holes are where God divided by zero.