Re: How are OO databases doing

Clinging to sanity, "Arturo Hernandez" <arthernan_at_hotmail.com> mumbled into her beard:
>>OO "databases" are just network model databases in new clothes. The data
>>model is clumsy, inefficient and difficult to work with.
>
>>It's no surprise the market rejected them.
>
>
> I share your opinion on network databases. If we look at the problem
> from a broader perpective. Development on the client is still done with
>
> OO langages. And for developers like me it is still a pain not to have
> an integrated technology. So why the disconect? Why did OO databases
> fail in the market?

The trouble with the OO databases was that they generally were clumsy; possibly more clumsy than the ways they attached to SQL databases.

Furthermore, portability immediately goes out the window, from several perspectives:

1. You're commonly tied to one language, either C++ or Java
2. Modifying object types is liable to provide versioning problems (CORBA always had plenty o trouble with this), so that data structure changes will be deeply tied to the applications
3. Standards more talked about than observed...

It strikes me that of those, #2 is likely the really troublesome part. There's a return to the "brittleness" that IMS and ISAM applications always suffered from.

-- 
"cbbrowne","@","gmail.com"
http://linuxfinances.info/info/slony.html
It is interesting to note that before the advent of Microsoft Windows,
`GPF' was better known for  its usage in plumbing: "Gallons Per Flush"
-- dedmonds_at_aw.sgi.com (Dean Edmonds)