Mailing-List: contact tomcat-user-help@jakarta.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Users List" <tomcat-user@jakarta.apache.org>
Message-ID: <3E5AA577.4070701@cymulacrum.net>
Date: Tue, 25 Feb 2003 07:06:31 +0800
From: Chong Yu Meng <chongym@cymulacrum.net>
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.2.1) Gecko/20021130
MIME-Version: 1.0
To: Tomcat Users List <tomcat-user@jakarta.apache.org>
Subject: Re: JDBC & ORACLE implementation !
References: <20030224140339.16480.qmail@web20205.mail.yahoo.com>
In-Reply-To: <20030224140339.16480.qmail@web20205.mail.yahoo.com>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit

I think if you take Oracle installation, configuration and maintenance 
out of the picture, you definitely have a much more workable plan. I 
agree with Peter, in that designing the tables and application logic are 
going to be tough. I once wrote a servlet that processed CDR data from a 
Cisco switch, and I spent a lot of time getting the logic just right. If 
I understand you correctly, Swapneel, you need the database for storage 
only, correct ? Are you planning to use the Oracle text indexer, or are 
you implementing that yourself?

One last thing : JDBC may take you a shorter time to learn than the 2 
weeks I put down in an earlier email. On reflection, that is probably 
padding too much, but I recommend that you do not underestimate the time 
taken to get the JDBC connection going, especially for Oracle. I've had 
problems with 8.1.5 before and had to resort to DataDirect drivers. 
8.1.7 seems to be ok, though.

Regards,
chong


Peter Lin wrote:

> 
>overall, using JDBC with Tomcat is the easy part. Deciding how to implement your tables and application logic will be the hard part. If your data is not normalized and doesn't need to be, then the first thing you should look into is statistical analysis of text. there are several well tested algo's for doing this type of processing. Unfortunately I don't know the name of the algo's. I worked on integrating personalization applications a couple years back relating to filtering news.
> 
>If your needs aren't too complex, it shouldn't take too long to implement some form of statistical analysis. Using file system to store the entire text doesn't necessarily mean you can't store the text summaries in Oracle. Google for related topics and you should be to find some examples. If you're needs are more complex, you'll need to look into grammar based parsing, which is a slow process. Most of the comparison between grammar based and statistical parsing has shown that statistical is more effective. hope that helps.
> 
>peter
> 
> Swapneel Dange <s_dange@hotmail.com> wrote:hey peter, mike & chong !
>
>so if i stay with a small thing like SQLPLUS, the JDBC connectivity wont be 
>a tough thing to do as compared to the ORACLE implementation. right ? 
>because in last few days after consulting with some people in-house here, i 
>am thinking over the OPTION of SQLPLUS.
>
>do commment on this !
>
>Swapneel Dange
>505-642-4126
>http://www.cs.nmsu.edu/~sdange
>  
>


---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org