[Sigia-l] information architecture in practice

[Ted Han]

On 6/29/05, Alexander Johannesen <alexander.johannesen at gmail.com> wrote:
> A *lot* of my work is about taxonomies, thesaurii and Toic Maps, and I
> try to avoid the pitfalls of RDBMS (relational databases) and SQL
> (database query language) as much as possible. There are other methods
> I find a lot simpler. Have a look at this;
> 
> <term id="346">
>    <label>Forgery</label>
>    <use-instead refid="123" />
> </term>
> 
> <term id="123">
>    <label>Fraud</label>
> </term>
> 
> This is a thesaurus XML. From this point on the XML will probably feel
> more workable than RDBMS and SQL for sure. Here's a taxonomy excerp;
> 
>    <item name="Arts and Humanities">
>            <item name="Architecture" />
>            <item name="Art" />
>            <item name="Humanities" />
>            <item name="Language" />
>            <item name="Literature" />
>            <item name="Music" />
>            <item name="Philosophy" />
>            <item name="Religion" />
>    </item>
> 
> That's really all you need for datamodel and dataset. Next is to
> choose some good tools for this, and I like XSLT but anything will do,
> really. It's all about relations between items, so a good hint at
> working with this in XML is to read up on the canonical XML
> specification, and pay attention to id and refid attributes. Also, you
> can have a look at Topic Maps for a format and datamodel that supports
> all this and more (http://shelter.nu/tm.html)
 
I'm sorry if i'm just missing something or whatever, but how exactly
does this avoid problems created by SQL and RDBMS (and which
problems)?  Cause i'm fairly confident that i could code the example
used above without terribly much trouble (of course you'd require some
scripting language to do some of the redirection, but thats the same
functional role as the XSLT).

-T