Re: Navgation and/or query approach (was Re: Flamewar object databases vs. relational databases)
Date: Sat, 21 Jul 2001 23:29:52 GMT
Message-ID: <3B1E11B6.D137F652_at_myview.de>
hi there!
Marten Feldtmann schrieb:
> In the first sight this may seem to be not nice, but actually you
> just have the same problems within your application as one would
> have between applications on different computers. If I solve my
> internal problem I also solve the network concurrency problem.
on a conceptual level you are right: on that level there is no difference between several representations of a persistent instance in the same process and different representations of a persistent instance in different processes. however, from a technical point of view this is very much different: the communication means you can use for synchronizing are very different (method call vs. interprocess communication). this can be harmonized to a certain degree by using some kind of communication layer (like CORBA, RMI), which has of course some performance implications.
> In general I go via queries nearest to the objects I want to show,
> after reaching this state I switch to navigation.
>
> On the other hand I *per definition* decide NOT to navigate via
> special (speak important) relations between objects but only via
> query.
>
> But all this requires one to think about the usage of "="
> (object identity) or just "=" (value equality) or to test objects
> against their OID.
"=" is an assignment. you possibly meant ".equals()", did you?
regards
robert
-- Robert Klemme Software Engineer ------------------------------------------------------------- myview technologies GmbH & Co. KG Riemekestra? 160 ~ D-33106 Paderborn ~ Germany E-Mail: mailto:robert.klemme_at_myview.de Telefon: +49/5251/69090-321 ~ Fax: +49/5251/69090-399 -------------------------------------------------------------Received on Sun Jul 22 2001 - 01:29:52 CEST