- From: Soumitra Sengupta <soumitra@b-bop.com>
- Date: Thu, 01 Feb 2001 10:01:30 -0800
- To: Dream Catcher <isc70409@leonis.nus.edu.sg>
- CC: xml-dev@lists.xml.org, xmlschema-dev@w3.org
- Message-ID: <3A79A47A.40FD72D0@b-bop.com>
Unfortunately the answer is "It Depends". There are both technical and business reasons. The technical reasons for using one technology over the other has been explained very well in several emails and you can look up Ronald Bourret's articles to get more information on this. The business reasons typically have to do with the cost of ownership and safety and security of the data. Organizations have made very large investments in RDMS technology both from the people standpoint and technology standpoint. That is why, IMHO the ODMS vendors had a very hard time breaking into the database market. In the end you have to cut through all the marketing language and technical religion and understand your requirements well and then pick a solution that works for you. Unfortunately there is no one magic bullet. Here are some rules of thumb that has worked for me: If you have a few documents and do not need to search across them, a persistent DOM implementation will work If your XML is "data centric" (no mixed content etc.), semi-structured and you need a robust, scalable and high performance storage, RDMS or other DBMS types systems will work well for you If you have all kinds of XML content and in large volume, then there are products from several vendors including our company that will meet your needs. If the data is for human consumption, a file system and a good text search engine will do the job just fine If the data is for the consumption of automated application, you will need a farly expressive and structured query language like QUILT or XQuery to get the job done. In which case you will require a XML DB. Hope this was helpful. Soumitra Dream Catcher wrote: > Although I am developing an application for securing XML data. This > morning I woke up and came out a general question: Is it really a better > idear to store data in XML structure instead of relational table? As we > know to bulid a DOM tree is quite expensive for main memory...Any advice > and opinion are cherished...:-)
Received on Thursday, 1 February 2001 13:06:22 UTC