Re: I think that relational DBs are dead. See link to my article inside
Date: Thu, 06 Jul 2006 20:12:43 GMT
Message-ID: <%8erg.60479$Lm5.18587_at_newssvr12.news.prodigy.com>
Dmitry;
Save yourself some time (may be too late) - take a look at
www.intersystems.com the "cache" DB is based on the old MUMPS DBMS which was
supposedly a 'node-based' data storage and retrieval system first developed
for medical applications (like neural networks?)
As Mr. Badour has said elsewhere in this thread (more than once) - you may
be "plowing old ground." Cache already supports everything you have listed
that you do, but it also has multiple users, multiple threads, transaction
triggers, SQL query support, and lots more; all without tables - but not
compliant with strict RM rules (but what is?)
BFaux
"Dmitry Shuklin" <shuklin_at_bk.ru> wrote in message
news:1152128550.212098.266230_at_p79g2000cwp.googlegroups.com...
> Hi Ed,
>
>> Joe Celko has an approach for handling trees in SQL. It is more
>> difficult than in a network model since elements in a tree form an
>> ordered set, while the Relational model deals with unordered sets. But
>> if that's your only flaw for Relational model, that's a pretty weak
>> arguement.
>
> Yea, this is just ONE known example of flaw.
> In my research trees doesn't have ordering by default. Ordering should
> be implemented by developer. It is very easy because each link from a
> node can be colored by identefier.
>
>> I've never done Nueral nets, but I was told once it is implemented with
>> a set of weighting tables. Maybe that is an old approach.
>
> This is normal but not single possible approach to ANNs modeling. I am
> using different approach - object oriented.
>
>> Good luck with your research.
>
> Thank You Ed
>
> WBR,
> Dmitry
>
Received on Thu Jul 06 2006 - 22:12:43 CEST