Re: XQuery question

Mikito Harakiri wrote:

> "Sebastian Schaffert" <schaffer_at_informatik.uni-muenchen.de> wrote in
> message news:b9vvpd$hjc$1_at_minotaurus.cip.informatik.uni-muenchen.de...
>

>> I repeat it, I am NOT claiming that XQuery is good. I am claiming that
>> an inherently tree or graph-shaped data representation language is often
>> superior to flat relational tuples. I do NOT say that XML is a
>> particularly good representation for tree- or graph-shaped data (Lisp
>> S-expressions would achieve the same). However, XML is well
>> standardized, it has a broad community and despite its deficiencies
>> (like DTD, processing instructions, unfortunate link mechanisms, etc) is
>> a reasonably good representation for such data.

Obvious: A graph in general cannot have a linear representation. In any linear representation you'll need a reference mechanism (binary relation). However, in many cases it is still much better to only have a few such references and keep the rest tree structured.

> Or XMP propellerheads do think that they can
> handle graphs too? Could you easily solve graph problems in XML? Find the
> shortest path in a graph, for example?

This is probably not a query that I'd ask to a graph structured database. When querying a database, I am usually interested in the *data items*, not in the properties of the structure. Did you ever ask a question to your relational database on "what is the minimum number of joins I need to prove that there is a connection between two data items"?

>
> If I want to learn about graphs, I'll grab a Graph Theory book. It
> described Matrices, Cuts, Flows, Graph Spectra etc. First Order Logic has
> little to do with this area.

Complete nonesense. Many issues of graph theory (e.g. coloring a graph) can easily be represented and solved in first order logic. Instead of fetching a book on graph theory, I recommend you to read a good book on mathematical logic (e.g. E. Mendelson, "Introduction to Mathematical Logic").

> Neither XML. A suggestion that some sloppy
> tag manipulation is a superior technique that naturally embrace
> hierarchical structures is just hillarious.

I was not talking about "sloppy tag manipulation". This is merely your personal impression of XML which obviously has nothing to do with reality.

Sebastian Received on Fri May 16 2003 - 13:15:22 CEST