Re: Database Application with Obj-C
Re: Database Application with Obj-C
- Subject: Re: Database Application with Obj-C
- From: Ruslan Zasukhin <email@hidden>
- Date: Tue, 23 Dec 2003 18:05:38 +0200
on 12/23/03 5:36 PM, Kim Friesen at email@hidden wrote:
>
This topic seems to come up consistently. I'm curious as to why EOF or
>
SQL Libraries are the first things to be discussed.
EOF -- because this is Apple and OS X.
SQL because this is 90% of database market I think.
OO DBMS are not so popular. They grow, but still have many issues.
>
Is it because of Legacy databases that people want to connect to?
>
External reporting tools ? Interoperability ? I'm not sure that for a
>
new standalone or relatively small multiuser application that I would
>
like to have an unnecessary binding between my model objects and a
>
relational schema unless it buys me something significant.
>
>
Are there other viable options ? i.e. Object databases ? XML
>
databases? What about something like ZODB ?
IMHO OO DBs is not very right idea.
You want ask me why? In short...
1) OO DBMS can be reached ONLY and ONLY from few OO languages as
C++, Java, Smalltalk.
They by definition cannot be reached from procedural languages as
C, Pascal, Lingo of Macromedia, ...
2) OO DBMS try __absolutely__ to hide database access and give you
impression that you work with C++ objects.
BAD IS: that such idea make them to add a lots of overhead to implement such
feature.
BAD IS: now developer is not able see when it works with RAM objects and
when its algorithm will touch HDD. In contrast, If developer see that now it
touch HDD he is able write algorithm in other way to get faster speed.
3) OO DBMS store objects not as records of table. They claim that this way
allow them be faster especially if this is IMAGE/Picture databases...
I VERY NOT AGREE WITH THIS!
In Valentina, for example, BLOBs and Picture fields NOT slow down your work
with tables. Absolutely!.
And I can prove mathematically, that storing of objects not in records of
table, will cause SLOWER access to big sets of data and their iterations,
What is required for Search, Grouping and aggregative calculations.
-----------------------------
XML Database
-----------------------------
Another strange beast :-)
All data are stored on HDD in the XML format.
-- Just image how much more will be files. Because they are text.
because each record will have big overhead on tags comparing
to super-compact RDBMS way.
-- so files are bigger -> more time to load them from disk.
they are text -> so we need spend time on parsing of text.
IMHO this is juts a modern stream...
Only advantage of it -> data can have NON - Table structure.
But modern Object-Relational DBMS that have new features such as Arrays,
List of objects, ... Can offer the same and in more effective way.
--
Best regards,
Ruslan Zasukhin [ I feel the need...the need for speed ]
-------------------------------------------------------------
e-mail: email@hidden
web:
http://www.paradigmasoft.com
To subscribe to the Valentina mail list go to:
http://lists.macserve.net/mailman/listinfo/valentina
-------------------------------------------------------------
_______________________________________________
cocoa-dev mailing list | email@hidden
Help/Unsubscribe/Archives:
http://www.lists.apple.com/mailman/listinfo/cocoa-dev
Do not post admin requests to the list. They will be ignored.