The consortium members ought to agree to adopt all changes that the SQLite architects may make to the SQLite core and ought to agree to maintain a basic degree of interoperability as defined by the architects. A SQLite database developed in one application ought to remain portable and queryable by another, at least to the extent that it makes no use of any proprietary features introduced by a consortium member.

A case in point: Adobe AIR treats INT PRIMARY KEYS as if they were INTEGER primary keys yet in the SQLite core these are treated very differently: INTEGER PRIMARY KEYS are treated as an alias for the rowid; but a column defined as INT PRIMARY KEY is treated as a normal column, not an alias. In order to get accurate query results in AIR with a SQLite database created outside of AIR, any table with an INT PRIMARY KEY must be recreated without a PK declaration, and a unique index must be placed on the INT column instead. Adobe has said it will document this difference in an upcoming revision of their docs.

Developers also should not have to guess why a query plan is not the same from one consortium member to the next. If SQLITE3.EXE EXPLAIN QUERY PLAN returns a plan for a query that is different from from the plan returned by a consortium member’s implementation, it should be possible for the developer to consult the consortium members documentation to find out why. If interoperability is sacrificed to some other end, the consortium member should be as open about that choice as possible.

There should be more to consortium membership than an obligation to fund core development.

Check out my page at http://www.squidoo.com/sqlitehammer for more discussion on SQLite. Mike Chirico’s page at souptonuts.sourceforge.net is also a great resource on programming SQLite.

http://www.firebirdsql.org

I’m not so sure I’d put HTML5 with Adobe Air vs. Mozilla Prism. vs. Microsoft Silverlight. My main reason is that they are all platforms rather than standards. Also Adobe Air uses WebKit, Prism uses Gecko, both likely HTML5 adopters. So Adobe Air is HTML5 proprietary. Mozilla Prism is XUL HTML5. HTML5 not end-user marketable since it’s not a product, it’s a standard. Air, Prism, Silverlight are. Google Gears if it succeeds would likely end up being both since I’m pretty sure Google will push for it and even put their developers on the task of implementing anywhere and everywhere to ensure their Apps work correctly anywhere.

@Chris: good point, I accidentally left them out when putting that list together. Added now.

In fact if anything I would say that HTML5 belongs in the “Adobe Air vs. Mozilla Prism vs. Microsoft Silverlight” line.