Re: [topicmaps-comment] mapping topic maps on a relation database

[Kal Ahmed <kal@techquila.com>]

At 16:35 17/04/2002 +0000, Murray Altheim wrote:
>Lars Marius Garshol wrote:
>
>>* franck legoux
>>| | Does someone has a model to map "topic maps" on a relational
>>| database ?
>>* Andreas Rittershofer
>>| | This is possible, but why not to store it in an XML-database like
>>| Xindice? See http://xml.apache.org
>>You could do that, but the benefit is much smaller than it seems.
>>Basically you would have to create a new DTD (call it XTM') and then
>>transform from XTM to XTM' before storing it. Otherwise performing
>>simple queries like "show me all composer born in Italy" is just plain
>>hell to compute, since you would have to do things like merging,
>>following three different kinds of references etc during the query
>>evaluation.
>
>
>Unless you always stored already-merged topic maps. But you're
>correct, this isn't a simple matter of storing an XML document.
>But simple queries of an already-merged topic map (such as "get
>all topics whose base name is 'harrier' in the scope of 'birds'"
>can certainly be done. You just need to be sure that the TM engine
>is hooked into the DB correctly when TMs are brought in and merges
>are requested. [I'm not looking forward to this part of my future...]
>
>
>>Using an XTM' DTD may be a viable option, but I think you'd find XML
>>Query about as easy to use for querying it as SQL for querying the
>>RDBMS topic map.
>>What you really want is a frontend that hides all the messy storage
>>details for you. All the topic map engines do this for you, in one way
>>or another. TM4J even has an XML storage option, but I don't know what
>>DTD that uses. The commercial ones also have RDBMS storage solutions,
>>but none of them publish their schemas, as far as I know.
>
>
>Kal used the Ozone XML database, so it's only a matter of work to
>reengineer it to use Xindice.


Sorry I've been away from this forum for a little while - so I'm in catchup 
mode ;-)

TM4J has two fully-featured backends - one based on storing the processed 
XTM file in memory and one which uses Ozone - however the Ozone backend 
does not make use of the Ozone XML database, it simply uses Ozone to 
persist Java objects (and that works quite nicely! :-). There is an 
experimental backend which is not complete yet which uses XPath expressions 
and a DOM storage model - if/when completed, it should be possible to hook 
that up to an XML::DB-compliant storage system.

In practical terms[1], I think that I agree with Murray that if you 
restrict yourself to management of "consistent" topic maps - ones with all 
merging resolved, use of XPath/XQuery over the topic map should be doable - 
perhaps not as efficient as if you also created other indexing structures - 
but not impossible.

Cheers,

Kal

[1] With the caveat that I haven't actually done it yet ;-)