Aldo Bucchi wrote:
>I'm a newcomer to semantic wikis, and semantic related techs in general.
>I am flirting with the idea of implementing an intra-organization
>semantic wiki in my company... using a "semantic" wiki is appealing
>since it would allow me to collect unstructured knowledge over time in
>a cooperative manner. This knowledge base can evolve organically and
>feed other systems ( CRM, project management, ERP, etc )
>Ok, so I have been playing with ikeWiki for a while, and the user
>interface ( with the ajax goodies ) rocks. Plus, it's java, the
>architecture is remarkable and it is built on jena so I should
>(probably) be able to slurp the triples from other apps ( haven't
>tried to sparql it yet ). But, I'm a little troubled about the
>maturity/robustness of the project, and wether it might suffer
>considerable changes and leave me stuck in the mud.
IkeWiki is a system in constant development, so it is of course subject
to frequent changes. However, since we are using it for our own
knowledge management since last week, you can expect it to stabilize and
mature very quickly. Also, if there are major changes to the database
schema or other internal things, there will very likely be an "upgrade
path", since I'd otherwise upset my colleagues.:-) Upgrading will
probably be based on a IkeWiki-specific export format that you can use
to dump your repository and reload it somewhere else or in a newer
For the nearer future, I expect the following internal changes to happen
that might affect upgradability:
1. arbitrary file resources / attachments; this is a feature that we
pretty much need for our own knowledge management
2. different reasoning engine; while Jena is a nice framework, its major
disadvantage is that it is very slow and not scalable. Try loading a
medium-sized ontology with reasoning turned on and you'll see what I
mean. I have already implemented an (incomplete) OWL-Lite reasoner in
SWI Prolog that performs extremely well in cases where Jena already
Example: DOLCE ontology, if I load it in Jena with only OWL_RDFS
reasoning enabled, I get an OutOfMemoryError after a few minutes, even
when giving 1GB to the Java process; in SWI Prolog, I have almost
complete OWL-Lite inferencing (no consistency checking, only drawing
conclusions, as that's what I need), and it enumerates all possible
inferences within 1 second.
In the next weeks, I'll try to wrap a Jena model using JPL/JNI around it
and see how it performs. Since this requires a native library, it will
probably be optional rather than mandatory, with a fallback to the
existing Jena models.
>next on my to-be-reviewed list, Semantic Media Wiki and a dozen more.
>any opinions? suggestions?
>Am I too ahead of time?
Maybe, most Semantic Wikis are currently in the prototype phase.:-)
Anyway, it appears to me that Semantic MediaWiki has the largest
developer base currently. I don't agree with some of the principles in
it (supporting reasoning is mandatory for me), but it is a rather stable
system. IkeWiki is our testbed for trying out new ideas, so it is
changing from time to time.
| Dr. Sebastian Schaffert sebastian.schaffert@...
| Salzburg Research Forschungsgesellschaft http://www.salzburgresearch.at
| Knowledge Based Information Systems +43 662 2288 423
| Jakob-Haringer Strasse 5/II
| A-5020 Salzburg
PGP Key fingerprint =
13 1D 2E 4F 20 3E C9 1F 4C 57 52 87 8A 80 48 4D F5 E9 97 EC