Re: Semistructured data models

From: Jan.Hidders <hidders_at_hcoss.uia.ac.be>
Date: 24 Jul 2002 11:13:54 +0200
Message-ID: <3d3e6fd2$1_at_news.uia.ac.be>


In article <4a11bbd0.0207232246.74bd1fc4_at_posting.google.com>, Chaitanya Mishra <cmishra_at_cse.iitk.ac.in> wrote:
>
>I am doing a project on semistructured databases right now. One thing that
>doesn't seem to be defined precisely is whether semistructured data is a
>graph or a tree. All the earlier papers (the lore project ones etc) state
>that it should be represented as a graph. But quite a lot of papers assume
>that the data model is a tree. Could anyone help me out.

Yeah, well, if we knew exactly what the structure was it wouldn't be *semi*-structured, would it now? :-) But seriously, the reference "Data on the Web" that was given is indeed an excellent one, and you will see there that both definitions are valid and used.

>Also is XML currently the most popular semistructured data format?

Without a doubt. Note by the way that XML is somewhat "inbetween" on the question whether it represents a graph or a tree. The fundamental structure is a tree (it has to because it is basically a text format) but you can use identifiers to let any element point to any other element and thus represent a graph. This is also supported in XPath where you can write a path like person/_at_spouse=>person/_at_name to make jumps in the tree.

>I've also heard of something called the RDF (Resource Description
>Framework). does it have anything to do with semistructured data as well?

Yes. It is a standard for writing down in an XML notation semantical information about XML data on the web like "who created what" and "what is this page about" and "what type of information is found where". Because they have also added the notion of schemas for such information and given a formal semantics for it, it has more or less grown into a complete data model of its own. I would even say that as a data model it makes more sens than XML. If you want know more there is a nice introduction at

  http://www.w3.org/TR/1999/REC-rdf-syntax-19990222/

and more information in general at

  http://www.w3.org/RDF/

>Finally, has any work on parallel semistructured database systems been
>done?

Not that I know of.

  • Jan Hidders
Received on Wed Jul 24 2002 - 11:13:54 CEST

Original text of this message