sqlite file format, anyone?

[Derek Atkins]

Matthew Vanecek <mevanecek at yahoo.com> writes:

> MySQL is what Derek has been *researching* for an embedded DB, as well
> as libdbi for a generic SQL interface (which would open us up to things
> like SQLLite).  My *main* objection to MySQL and SQLLite is the lack of
> transactional security and foreign keys; and thus data integrity (Yes,
> MySQL has transactions and foreign keys, but you have to know to compile
> MySQL to support InnoDB and you have to explicitly create tables in the
> InnoDB format).  If nobody cares about the integrity of their data (or
> if we can force the InnoDB format for MySQL), then my objections
> disappear, for what it's worth (probably not much).

FWIW, I don't think "transactional security" is important for SQLLite
-- it's a single-user process, so there is no concurrency issues; the
"database" is "locked" (much like the XML file is "locked").  The only
thing you need to worry about is crashing in the middle of an update,
but frankly you need to worry about that with XML too so it's no worse
than where we are.

> I like the libdbi idea, to a point.  Postgresql and other capable
> databases do provide event notification APIs that are not available in
> lesser- or barely-DBMS systems such as MySQL.  Those APIs are not
> available through libdbi, and therefore a way to work around the lack
> would be required.  Not an impossible task--just something to keep in
> mind.  There are other multi-user considerations, of course, but those
> can be enumerated at another time.

Yes, let's please put this off...

-derek
-- 
       Derek Atkins, SB '93 MIT EE, SM '95 MIT Media Laboratory
       Member, MIT Student Information Processing Board  (SIPB)
       URL: http://web.mit.edu/warlord/    PP-ASEL-IA     N1NWH
       warlord at MIT.EDU                        PGP key available