Re: Subject: Re: [GENERAL] A book for PgSQL? A need? yes? no? (fwd)

> Good point, but for any argument there are counter-arguments, and it is
> sometimes so tempting to throw in a few good ones!
> [snip litany of reasons that books are a pain]

>good points also.  maybe we don't need/want O'Reilly.  that's OK.
>certainly having something that can be zapped to kinkos and bound to user
>specs is wonderful.  that way the only people who really want booklike
>stuff will have it and no publisher gets stuck with inventory of
>unsaleable material.  but there is the mindshare factor.  an O'Reilly book
>on the shelves would lend a lot of credibility to postgres (as if it
>needed any).  and that would probably be worthwhile.  maybe someone is
>interested in writing/editing and a target would be to publish when
>postgres has pk/fk support, left and right joins and a few other features.
>
>$0.02,
>steve
----- End of forwarded message from Stephan Doliov -----

O'Reilly have some pretty tight requirements for books, which is a
good thing. One of those, common to all publishers, is that a product
is reasonably stable. All software gets updated, but no one wants
major work that gets outdated before the book even reaches the
bookshop. So it comes to Steve's point above. When postgres has
stabilised - or is it already? - then is the time to put together
a book that meets O'Reilly's requirements.
So, apart from pk/fk and joins, and no doubt a few other little
exclusions, how ready for having a book written around it is
postgres (PgSQL).
A good example is Sendmail. O'Reilly's book on Sendmail is a
masterpiece, and not at all a reprint of the docs. Sendmail has
been updated a number of times since the first book was released,
and there have been subsequent books released. But the fundamental
principles remain the same, so the re-write is not a "start from scratch"
work.
It comes down to this. To do a manuscript to O'Reilly specs is a
major undertaking. Is Postgres in its stable incantation, or will
there be a future release that is as radical a change as some of the
earlier ones? I'm not talking upgrades and additions here, but
basic principles. The fact that later editions may in fact have
such things as L/R joins is fine. But major rewrites of the whole
system is not fine.

The bottom line. I'm already working on a "Guide To PgSQL" that will
be published along T.Nelson/Addison Wesley/etc lines. This will cover
the material in Text Book style. Step by step bringing a new user
up to speed on the Postgres version of SQL as we know it now. Largely
non-technical. Not an installation manual. A User Guide/Tutorial.

I'm also putting the design together for a  book that will fit
the O'Reilly mould. In depth, technical discussion of just how
to make PgSQL hum. Then I see if they are interested in publishing.

Likely publishing time frame? The first will be out
later this year, hopefully in time for second semester in colleges
and universities.  The second... we'll see if there is interest in
publishing.

I love the comment by someone else on the list.
"Why a book, when we have a website full of docs?"
Answer: "Because I can put a book on the shelf in front of me"

great issue. I follow the discussions with interest.
--
Support Whirled Peas.
http://www.chalmers.com.au          robert@chalmers.com.au
Location: P.O. Box 2003. Mackay. 4740. 21'7" S, 149'14" E.
China House - Mackay & The Whitsundays - Robert's Bookshop