From: Patrick D. <pa...@du...> - 2009-07-18 18:10:24
|
Stuart, Stuart A. Yeates<sy...@gm...> wrote: > On Sat, Jul 18, 2009 at 4:58 AM, Daniel Paul > O'Donnell<dan...@gm...> wrote: >> Obviously, in comparison to this group I have yet to put aside my >> childish things, but if anybody knows of anybody or any project that >> might find it useful to know about this, please let me know: > > Our experiences with ontologies have led us to look at migrating (in > the long term) from our currently topic-maps approach to one based on > RDF and related technologies. While I hesitate to call our use of > topic maps a mistake (we had no way to determine how the technologies > would play out, CIDOC-CRM went with topic maps, etc), I would strongly > advise you against going down the topic maps route today. > The use or non-use of topic maps really depends on the requirements of your project. By no means will every project benefit from the use of topic maps but like all technologies, it depends upon your goals. Selecting an inappropriate technology for your goals is never going to lead to happy results. For my part I would like to learn more about your project and its goals before venturing an opinion about the use of topic maps. As far as Stuart's misleading when not simply false statements about reasons to not use topic maps: > There are three main reasons for this: > > Firstly the amount of development work in the RDF community is at > least two orders of magnitude larger than the topic maps community. > There are many more tools for many more purposes and the tools are > being developed/matured significantly faster. Yes, the RDF development community is larger than that for topic maps. But then the SQL community is bigger than both. So why not use SQL? Or in terms of longeivity, you could always use COBOL, which still underlies all credit card transaction processing. Bottom line is what technology you *should* use depends upon *your* requirements. Downloading Protege (http://protege.stanford.edu/) isn't going to make you an ontologist/RDF expert. By the same token, downloading the recently released as open source Ontopia knowledge suite (http://code.google.com/p/ontopia/) isn't going to make you a good topic map author either. (The Ontopia software is a full release of previously commercial software.) No software of any description can compensate for lack of expertise or poor design. > > Secondly the RDF community takes an explicitly standards-based > approach, show a reluctance to redevelop things when standards can be > re-purposed (mime-types, language codes, dublin core, vcard, etc) The PSI (Published Subject Identifier) sets based on *standard* language and country codes don't seem to have any reluctance to rely on other standards. There is no evidence that the topic map community shuns re-using standards. And with many creating ontologies to the right and left I find the re-use argument more than a little amusing. As far as I can tell, every ontology author thinks their ontology is so compelling that all others will abandon their own efforts and flock to their banner. Given the 10,000 ontologies at Swoogle, http://swoogle.umbc.edu/, I really don't see any evidence for that proposition. BTW, in addition to being the convener of the WG in SC 34 on topic maps I am also the editor of OpenDocument Format 1.2 and the chair of the subcommittee that adopted RDF/RDFa for the extensible metadata features of ODF 1.2. So, as I said, it really depends on your requirements, whatever technology is being evaluated. > > Thirdly vicarious interoperability seems to work in RDF and not in the > topic maps community (by that I mean the ability to interoperate > between independent data-sets). Which independent data-sets? Do they conform to the defined syntaxes and semantics of ISO 13250? Did you use a topic map engine that conforms to ISO 13250? I freely grant that poor design can lead to topic maps that don't have the interoperability desired but no software can compensate for bad design or poor execution. Apologies for not having a contribution to make on your original question. I do intend to visit your site to learn more about your project and to see if I can remember any likely contacts who might know of related projects. Hope you are having a great weekend! Patrick -- Patrick Durusau pa...@du... Chair, V1 - US TAG to JTC 1/SC 34 Convener, JTC 1/SC 34/WG 3 (Topic Maps) Editor, OpenDocument Format TC (OASIS), Project Editor ISO/IEC 26300 Co-Editor, ISO/IEC 13250-1, 13250-5 (Topic Maps) |