| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
about the whole enterprise, although there is no question about the technological enormity of the target pursued. To our feelings, the administrators running the World Wide Web Consortium rushed the matter by recommending the
OWL as an ontology language standard fit for structuring the Web data, documents, and applications. Since, beside the well-knowing merits, the language has bad conceptual faults which make it falling short of a wide commercial use.First, any ontology formal language designed to formulate classes and their relationships has to include an ontology of relations describing their real features accompanied with the formal attributes.
Danny:
I'm afraid I'm unclear about what you mean by "real features" in this context, do you have any examples of these in other languages?
As I see it the OWL language attempts to find the sweet-spot between expressive capability and decidability in the context of the web. From my own little experiments, I'd suggest it does quite a good job.
Asha:
This is not the case with the OWL created as adding better vocabulary than developed by XML, RDF, RDF Schema, DAML+OIL in formal description of classes, individuals and properties, like cardinality, equality, and other formal characteristics. The language is lacking many significant features of relations, both formal and real.Among real, first of all, the relation of cause and effect is fatally missing.
Danny:
Cause and effect can be modelled using OWL - this is something found in OWL-S:
http://www.daml.org/services/owl-s/1.0/owl-s.html
Note too that the languages (RDF and OWL) are designed to be extensible, so additional constraints could be added as required.
Asha:
Such defects come from the approach used, purely set-theoretical and formal logical, while any content-based (world) ontology distinguishes internal and external relations, avoiding their reduction to relational properties.'Being a parent' is merely a relational property, while 'parentage' (parenthood, but not the act or process of parenting) is a relation of parent to child. Or, more generally, the relational property of being a cause is just a monadic reduction of causality, the relation of cause to effect.
Danny:
I can't actually see any part of that which couldn't be represented fairly directly using RDF/OWL As to the processing of the material thus represented, maybe what you describe falls outside of the capability of OWL's Description Logic - but that doesn't mean such logic couldn't be layered on top.
Asha:
So, instead of classes and relational properties, we have rather to speak of classes and their relationships as two sorts of distinct entities.
Danny:
Within OWL DL, yes. Within OWL Full (which encompasses 'generic' RDF) properties are classes too. Instantiation of properties isn't possible directly, that is lacking - but given that tractability is a key aim, there's a potential Pandora's box of undecidability if this were possible.
Asha:
Besides, it will make a poor language to use the class/property distinction, where all properties are reduced to relational properties as a substitute of the relations of classes. In fact, there are several sorts of property, substantial (material, structural), stative (qualitative, quantitative), active (functional), along with relational (causal, spatial, temporal) properties, all such properties of a substantial class (genus) are to be the properties of its species and individuals. With such distinction, it makes a difference whether we just state that 'a class (a thing) have attributes (properties)' or that 'a thing has the properties to be a substance, to be in a state, to be acted or acting upon, and to be related to something else'. The later is a universal ontological statement, while the former is a general logical proposition. That is, from one side, to build a formal ontology language, one may develop a set-theoretic (logical) ontology constructed as a formal logical system composed of its objects-primitives, classes, individuals, and properties, logical syntax (notation techniques, formation and transformation rules), and formal semantics (model theory), as the OWL is doing.
From the other side, one can develop a true ontological system involving as its primitives Entity or Thing and the kinds of entities or things along with fundamental definitions, axioms and real-world semantics seeking the interpretation of the theory and its truth conditions in the world of things, entities, or beings.
Danny:
The only difference I see there is in the 'real-world semantics' part. Like I said, I'm not entirely sure what you mean by this.
Asha:
Next demerit is tightly connected with the first. Any general theoretical model or conceptual scheme or description framework is built with some kinds of concepts, individuals, predicates, attributes, or relations. If the model represents the traits and features of the real world, its predicates represent real properties. In so far as the scheme describes formal objects, its predicates give description of formal properties, attributes, or predicates. As a consequence, there are two sorts of predicates: formal, logical attributes or real-valued, ontological predicates. Failing to see such difference can result in another conceptual mess and confusion. Proposing a general conceptual scheme by using the class (object)/property (relation) distinction, the data type languages like OWL are missing the meat of things. For such a formal data structure might not adequately represent the whole gamut of the world entities with their various interrelationships. We must clearly differentiate between classes, properties, definitions as logical forms and entity types as ontological classes; for the definition and class and property will always be in one of these thing types represented by the classes of ontological predicates and signified by the generic terms and expressions, such as thing, being or entity, substance or object, state, property or quality and quantity, change or action or process, and relation, like causality, time, space. Any meaningful statements or propositions is always about one or another of these types of entity, in two principal ways. If the same type of predicate is asserted of its kinds (a quantity predicated of magnitudes), it signifies its genus, a thing's essence. On the other side, if one type of predicate is asserted of another type, as equality and inequality relations are predicated of quantity, sameness or difference relations, of substances, it signifies rather a property than an essence.
Danny:
That appears to be the same point again, that somehow the Semantic Web languages are lacking 'reality'. Most things we do in this field are very approximate models, and grounding in the real world is limited. Having said that, I don't see any problem in creating the kinds of predicates you describe here.
Asha:
Thus, to construct a high quality Web ontology language as a computational ontology incorporating the basic rules of reasoning about the world and its generic entities, we first need to formulate the universal entity framework supporting all the major classes and relations in the single hierarchy of things, as it is implicitly intended to achieve by the SUO project, how successfully or unsuccessfully, this is another topic. And designing such a (machine processible) ontology language system requires a formal inquiry into the world's objects, states, changes, and relations as the fundamental classes of ontological predicates specifying diverse kinds of entity and realms of the world.
Danny:
Ok, and while that magnum opus continues, the crude approximations of RDF and OWL can be used to carry out useful work on the web. OWL most certainly is not perfect, but there is enough for it to useful, and significantly, be comprehensible to most web developers. Whether there is a single universal ontology is a moot point. There's certainly a lot more that can be done in the field, but I would suggest that we can learn a lot from applying what we have already, approaching the problem from the direction of what we do know, rather than that of what we don't.
But whatever, I don't see any reason to reject the Semantic Web languages on the grounds that they can't fully describe the universe, any more than you might reject the relational model for storage of census data. The Object Oriented programming paradigm often involves a crude modelling of real-world objects, yet it is still useful. Your argument begins by suggesting that OWL is unfit for commercial application because of its conceptual shortfalls. If anything its shortfalls, which are on the whole simplifications, put it more in reach of commercial developers away from the Ivory Towers.
The bottom line is that the W3C's initiatives and the loftier aims you describe aren't mutually exclusive. I'd go further and suggest that the former may indeed help bootstrap development of the latter.
Cheers,
Danny.
----
http://dannyayers.com