Is it just me, or is Sqlite3 goofy?

[Mike Owens]

On 12 Sep 2006 08:29:34 -0700, Paul Rubin
<"http://phr.cx"@nospam.invalid> wrote:

> > > But no one appreciates my finding those faults.
> >
> > No one appreciates the tone in which you report these alleged faults,
>
> Your tone is not so great either.

And what would you expect after someone who has take little to no effort to

1. read the documentation

2. seek help in the appropriate places, or

3. raise concerns to the appropriate people

unjustly and publicly vilifies an otherwise wonderful volunteer effort
to provide a free, high-quality, open source software package?

> > necessarily agrees with the faults that you find, nor elected you
> > system test engineer of the SQLite project.
>
> It's an open source project, as you like to say.  Everyone is a test
> engineer.

Only real test engineers provide constructive feedback to the
appropriate places. Posting ill-informed flames to the Python mailing
list is hardly constructive feedback, or worthy of being called test
engineering.

> > > > standard as well. What's your position on that? Do some Googling and
> > > > you can easily find 18 ways that Oracle's PL/SQL deviates from the
> > > > standard. And T-SQL is plainly nowhere close.
> > >
> > > And how many of those systems use dynamic typing?
> >
> > And how many conform to the standard?
>
> How many of those deviations are justified in their documentation by
> the responsible parties claiming, in effect, that they're smarter than
> the standard's designers?

No one said they are smarter than anyone else in effect or otherwise,
but rather the SQLite project stated that it disagreed with static
typing. When you write free software, you have the freedom to write it
however you want, and that is the justification. This is the approach
that SQLite took, they are free to do so, and in no way have they been
deceptive about it.

And because SQLite deviates in this way does not make it any more
guilty of deviating or not conforming to the SQL standard than any
other SQL database.

> It seems obvious to me that there should, at minimum, be an option to
> turn this particular nonstandard behavior on and off.

Then the obvious means by which to accomplish this are like any other
open source project: At a minimum, post this suggestion to the mailing
list, write an RFE, or best of all, write the missing code yourself
and supply a patch to the bug list.