-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 05/04/2010 07:30 PM, Doug wrote:
> Sure, if the book is just a copy-paste of the website,
> that's not too helpful. But hopefully things are stated differently, or
> examples are given which can be valuable.
I picked some random spots to check:
"GROUP BY" - several paragraphs in the book and 3 sentences in the SQLite
doc. I think they could both do with a lot of improvement especially in
showing how it differs from "ORDER BY", using and not using aggregate
functions etc. In both cases there is no way I could read the description
and then explain what it actually did given some examples unless I knew SQL
already.
Date/Time - A "chapter" in both and both are pretty good, but with a little
more content and warnings in the book, but for example the book omits to
mention how Vista will screw up timezone calculations. It also advises on
the textual representation being less efficient but doesn't give useful
advice on how much. For example if it takes 5% longer to sort a million
dates stored in text vs number then that may be acceptable but if it is 500
times as much then not.
FTS3 - The book is virtually useless despite having a page or two. You'd
still have to look online to work out how it actually works, how to add
data, query formats, stemming etc.
Virtual tables - The online doc gives better information on the distinction
between xCreate and xConnect. The book pretty much avoids any detail on
xBestIndex. This is by far the most complex method and very hard to
understand and get right until you have worked with it a few times.
SQLITE_DEFAULT_FILE_FORMAT - Book names the SQLite version that introduced
the new format (3.3.0) and adds a caveat about SQL parsing on open. A
little more useful than web site but then the web site does have a more
detailed page about backwards compatibility and for example mentions which
versions can't cope with a database where ALTER TABLE ADD COLUMN has been used.
ALTER TABLE - both have a page of doc with the web site mentioning the
version issue and the book not. The wording is a little different between
them in some places but neither is clearer. Neither gives an actual example.
abs() - One page in the book and one paragraph on the web site. The web
site is better as it explains what happens if you pass in a string that
cannot be converted to a number as well as Abs(-MAXINT64)
glob() - Tersely described on both, both referring to the GLOB SQL keyword.
Of course you can click in the web site but not a book. Book is better for
GLOB keyword actually mentioning what the match syntax is, but excludes any
examples. Neither mentioned anything about the international issues (eg
does [a-z] also include é?
max() - Book separates scalar vs aggregate into separate pages, neglects to
mention the collation functionality that is described on the web site.
pragma index_info - Two sentences, one diagram, no example in the book and
one sentence gibberish on the web site.
Neither the book nor the site go into usable detail about Unicode, locales,
internationalization, localization, the right way to deal with it in SQLite etc.
So on the whole the reference stuff in book doesn't really add anything over
the web and can be wrong or misleading. It could be better by giving
examples. It could show what happens when irregular or out of bounds values
are supplied. It could give performance considerations. And if it was
removed the book would be a lot shorter. (Presumably computer books are
bought on thickness no matter how much of that repeats what you can see in
your browser.)
Although I may seem negative, this is a constant glass half full thing that
really annoys me. Sure it is a lot of effort to author a book to the point
that it is somewhat equivalent to the web site, but so many times things
stop there. To me it is only what happens beyond that that makes a book
valuable and I'm hoping that will happen with this one.
Roger
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
iEYEARECAAYFAkvjQvIACgkQmOOfHg372QQP5wCePiqnhJS/XFKTEyoJBrKiqqPb
P34AoI5OTGlLZe/Dmp9yniNVxJ8GKCyy
=MvMI
-END PGP SIGNATURE-
___
sqlite-users mailing list
sqlite-users@sqlite.org
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
