Can you say Oracle? :-) Along this same line, and based on a few years of using SQLite and following the messages on this list, here's my two cents worth.
First a disclaimer: I have no knowledge of the detailed implementation and architecture of SQLite, other than by inference from discussions on this list. I did look at the source once long, long ago, and quickly decided, life is too short. With the background of the never ending drumbeat of "feature, feature, feature..." on this list as a reason, I wonder if the structure of SQLite could be "enhanced" to better support the "plug-in" concept, aka: as with Firefox, IGoogle, and the like for instance. This would allow SQLite (the module) to remain true to its roots (i.e. the "three choices") but better support those who desire to enhance SQLite through the use of both plug-ins and wrappers. This could better address a broader range of applications and allow the developer to selectively "bloat" the final result with only those features required for the project at hand. One thing that would greatly enhance concurrency (the most heard feature request?) would be to implement a finer grained locking structure in the base module. Currently you can either lock or unlock the database, period. All modern database implementations lock to, at a minimum, the the "record" (row) level. The implications of record level locking is pretty obvious related to concurrency. With finer grained locking implmenting multi user "server" wrappers and plug-ins would be better facilitated and allow the base SQLite module to remain true to the three choices. Fred -----Original Message----- From: sqlite-users-boun...@sqlite.org [mailto:sqlite-users-boun...@sqlite.org]on Behalf Of sub sk79 Sent: Sunday, September 20, 2009 9:16 AM To: General Discussion of SQLite Database Subject: Re: [sqlite] Most wanted features of SQLite ? >> 1. Generate a loadable SQLite extension and distribute it with DB. > > We can write C extension without any wrapper. Why we need to learn > your extension for getting the same result and possible new bugs? PL/SQL has a very wide user-base and a huge repository of existing code-base in the world. Using StepSqlite PL/SQL compiler this huge base can use SQLite by reusing their code as well as reusing their skills - no learning curve for this set of users. For those who do need to learn PL/SQL, it is still worthwhile to utilize StepSqlite because: PL/SQL is specially designed for and very suitable for handling database operations, while other general purpose languages like TCL, lua, C , C++ etc are very suitable indeed for application programming. StepSqlite enables combining the respective power of these two sets of languages in writing apps with SQLite back-ends by using the concept of Stored Procedures/Functions and Packages. All big databases operate this way as well: for ex, Oracle lets apps written in C++ to call stored procedures written in PL/SQL. StepSqlite is committed to bringing this power and convenience from the big database world to the small database world while preserving the small, fast and reliable. More benefits of StepSqlite described here: http://www.metatranz.com/stepsqlite/benefits.html http://www.metatranz.com/stepsqlite regards, -sk _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
