Subject | Re: [Firebird-general] IBM moves the database goalposts - xml related |
---|---|
Author | Paulo Gaspar |
Post date | 2004-12-09T17:26:23Z |
Martijn,
You are missing the point. The issue is not storing but querying in an efficient way.
Don't think about relational data. XML is used to store structured and semi-structured data of a different nature.
Think about XPath or XQuery and then imagine querying your database for documents or document fragments holding a given XPath pattern.
If you are still clueless, you can then try to follow the xml-dev list for a while to get a feeling for it... but it is so hard to get some usefull information from all the purely academic-but-useless crap that goes around there...
Regards,
Paulo Gaspar
Martijn Tonies wrote:
You are missing the point. The issue is not storing but querying in an efficient way.
Don't think about relational data. XML is used to store structured and semi-structured data of a different nature.
Think about XPath or XQuery and then imagine querying your database for documents or document fragments holding a given XPath pattern.
If you are still clueless, you can then try to follow the xml-dev list for a while to get a feeling for it... but it is so hard to get some usefull information from all the purely academic-but-useless crap that goes around there...
Regards,
Paulo Gaspar
Martijn Tonies wrote:
>Hello Nando,
>
>
>
>>MT> XML is a data-exchange and representation format.
>>
>>Yep.
>>
>>MT> What else does XML offer?
>>
>>I don't mind. I just took the opportunity to highlight a possible
>>reason why (IMHO) the industry is looking at XML as a storage means.
>>
>>If you are able to do everything you need with a pure relational
>>database then good for you. I outlined why I cannot (conveniently) do
>>all I need and I have talked with many people that have the same needs
>>and look at XML. Give me a way to store signed documents and query
>>their contents and I won't mind about the format - XML is just as good
>>as anything else (although I find it excessively verbose but that's a
>>different matter).
>>
>>MT> If people use it for complete documents and store them as such,
>>MT> you're using it the wrong way. Why should IBM or anyone else
>>MT> accept that and make it query-able etc?
>>
>>Because there's market there. And huge too. I feel that the relational
>>model could easily provide tools to do the above, thus eliminating the
>>need to look at XML as storage. But such tools are currently not
>>available.
>>
>>
>
>I argue that xml as STORAGE is plain silly.
>
>First: a DBMS can do everything it like to store the data. It can carve
>it in tree bark if it likes. Firebird compresses the data, no problem there.
>I argued with the MySQL guys about this too because of their
>non-standard CHAR/VARCHAR stuff - they argued that storing in
>char is different and then some bla bla... I said: no, it isn't. As long as
>you pad it with spaces when retrieving :-)
>
>Second: why do you store XML if it's only representation? What
>does XML offer you above "plain" relational storage?
>
>Third: relation storage has a logical/mathematical basis. What is the basis
>for IBMs XML storage stuff? Or even: for ANY XML-based "DBMS".
>
>
>With regards,
>
>Martijn Tonies
>Database Workbench - developer tool for InterBase, Firebird, MySQL & MS SQL
>Server
>Upscene Productions
>http://www.upscene.com
>
>
>