Karsten Hilbert wrote: >On Fri, Jan 20, 2006 at 09:42:29AM +0100, Bert Verhees wrote: > > > >>Cache it is an expensive database-engine. But so is Oracle or SQL-server. >>They >>are all expensive. >> >> >Well, there's always the OSS version by GT.M Sanchez which >is compatible enough to run Vista. > >Thanks for your take/insights on OO DB engines. Helps me >sort of understand. One fear I always have with them is that >they would tie me to one language accessing the objects >stored in them, namely the language which originally saved >the objects in there. Is that so ? > >Karsten > > This differs per DB. In case of Intersystems, there are interfaces for Java, dotnet, COM, ODBC, JDBC, COM, ADO,...., and more. ODBC, because the objects can also be queried by SQL. Interoperabilty is an important issue for Intersystems.
For other "OO"-database, I don't know. Intersystems called Cach? an OO-database, three years ago, but then the renamed it to Post-Relational Database. There are no ISO-defines for what an OO-DB really is, but it is all marketing, and I guess Intersystems received some complaints from wellknown people, and there fore changed the type-name of the database. One has also to be careful with the types. Different languages can understand different things in a certain type. f.e. 64 bits, 32 bits, 16 bits integer, 8 bits, or an Integer as object, signed, unsigned, little endian, big endian Type-names can have different meanings. Sometime you have to store data in raw-pointers because the type in which it is is not supported by your language. But these kind of problems are not especially for OO-databases, you can encounter them as well with RDB's But the chance to encounter these kind of problems is getting smaller and smaller. And the layers you use to access a database do mostly a lot of work for you. Hope this helps Bert
