Reading this discourse over the past few days, we seem to be taking one step backwards for every two forwards. Below I'm gonna present what my understanding of the current state of affairs is, and what some proposed solutions to it are. Perhaps this can migrate to the wiki with alterations refining our collective understanding.
But first, some general comments -- I'm not sure why people are comparing ontologies to traditional software paradigms - of course in a silo a non-ontology approach will suffice and will likely be more cost effective, yet this is missing the point entirely! We aren't comparing the cost of developing one application to another, the whole point is -- what if we want to update the software? connect it to another? These are the costs that need to be internalized and should be accounted for in the comparison; indeed it is here where the purported utility of ontologies should come to shine.
Of course, one can develop an app in a silo for cheaper, but what if two companies using different databases decide to merge? What if in the life sciences two or more different databases are required to interoperate. What if along your supply chain different people have different accounting / shipping / etc. software apps? Without an ontology, each of these scenarios requires far more time, effort and $$ than one with an adequately developed ontology. Hence, i submit that the cost analyses thus far have been missing the mark, and if proposals have been made to DoD or other potential clients without advancing these arguments, then they aren't really touting the benefits of ontologies.
On to point two - in this thread, people have offered Cyc as an example FO. I thought we'd moved away from FO meaning a unique foundation ontology and instead had decided that FO means Foundation OntologIES. In this respect, Cyc is clearly inadequate, since it only represents one philosophical paradigm, which as we've seen again and again does not meet everyone's needs. (Digression: access to Cyc is quite shoddy too, i (and my lab) have applied multiple times for a license to OpenCyc, never heard back .... will resist temptation to speculate why.) Maybe we should retire the term FO and replace it with FO's, for clarity?
Now on to my summary:
FO's should serve two purposes.
1- to maximize reuse and minimize "wheel reinventing" by providing a resource for would be developers to identify axioms they would like to reuse.
Note: This is stronger / more versatile than simple lexical matching / i.e. reusing a pre-developed notion of "wheel". Rather, one is reusing whatever it is that defines the wheel (vis-a-vis Aldo Gangemi's work on ontology patterns, my thesis on ontology design tools)
2 - facilitate interoperability among different philosophical paradigms
This means storing and perhaps dynamically generating mapping axioms between different ontologies. In the 1990's and early 2000's an Interlingua was offered as a potential solution. I think with time, it has become clearer that a Family of Interlingua ontologies is what is needed. One approach is captured in CoLoRe which focuses on what can be expressed in a particular language, and puts philosophical commitments secondary - to me this is one very promising methodology, haven't heard any other proposals.
Situation at Hand:
The ontology field is nascent and fractured.
There are few guides for would be ontology designers to aid in the development of an ontology.
There are few tools for "" """.
As a result of the fractured nature, reuse is highly inefficient, and once an ontology has been articulated it is unclear how they relate to others.
Mapping into the thus far created UO's is inadequate, since the UO's themselves aren't linked to one another.
Moreover, many ontologies today are still stuck in taxonomy mode, which again, might be useful in a silo, but hides much of the semantic intent external to the language of representation. This inadequacy is likely due to the lack of formal logic familiarity, coupled with a paucity of tools/guides.
1 - do nothing, hope the fractured nature will eventually meld into a reasonable whole
2 - solicit government funding to address the collective action problem arising from the tragedy of anticommons and the free rider problem.
3 - several isolated approaches have popped up - WonderWeb (for OWL), Ontology Design Patterns (for OWL), CoLoRe (for Common Logic), plus some of the other approaches presented at the OOR talk on ontolog last week
4 - combination of 2&3 (since the collective action problem is quite real)
Expanding on 4 - while each of these approaches provides a partial solution, there is no overriding force / initiative unifying them (unless the ontolog effort culminates in a true crowd sourced fashion). Moreover, each is funded by lab budgets, and AFAIK, no one has $ for a dedicated caretaker for these efforts (webmasters / someone who will maintain the wiki / update fields etc.). This is grunt work that most will bristle at, yet it is fundamental to the success of each project.
Detraction for do nothing -- ontologies might go way of AI in the 1980's. No certainty anything will come of it. Might fracture into irreconcilable camps.
Detraction for FO's - consensus is elusive -- that's fine, we don't need consensus! We're not trying to develop one size fits all. We're developing translation from "size 10 shoe" to "size 44 shoe"... I can't stress this enough: No consensus is a red herring at best, a straw man at worst!
In tandem, we need ontology design and semantic mapping tools to support the upcoming infrastructure.
Do people agree that we have a collective action problem here? We have a cost to be burdened, with distributed benefits. Traditional economic analysis would indicate this is often where government intervention is desired.
However, we also see a bottom-up / crowd sourcing approach, championed by ontolog, and seemingly gaining traction. However, I don't see why this would preclude the former, so long as a proposal is carefully delimited, and the moon is not promised :D.
On Sat, Feb 21, 2009 at 12:23 AM, Ron Wheeler <rwheeler@xxxxxxxxxxxxxxxxxxxxx>
Conklin, Don wrote:That is what I hoped that the wiki would do but it seems that most of
> I'll start from the end of your last post. There is little point in repeatedly referring to lost productivity on a grand scale when there is no one concerned about it who has the resources to address it. The closest match would require a call to 202-456-1414, ask for Obama. Good luck with that.
> My experiences with DOD affirm your comment in that they want an order of magnitude advance from semantic technologies. But the hard truth is that the more tightly a domain is focused, the more a relational DB solution will equal, or better a semantic solution. At much less cost and difficulty.
> The DOD entities I've dealt with typically seek to integrate some number of relational DB's, sometimes with unstructured data. But within a fairly tight domain (yes, there is a range to the domain focus).
> Interoperability, while given much lip service, lags the commercial world. The DOD XML Registry was going to be the solution...still waiting on that one. By the way, where is the ontology repository and how capable is it?
> Fielding large scale, operational semantic systems is not yet feasible because the supporting infrastructure is not in place. For example, look in the Sunday paper job ads for oracle DB administrators. Then look for ontologist ads...
> Unfortunately, as it takes money to make money, success breeds success. Fielded semantic systems that work will garner more funds for the application of those technologies.
> It would be great if the federal government CIO announced a major drive to develop and field semantic technologies. I'm not holding my breath and he is a lot more worried about the next windows virus that decimates federal PC's.
> It will take a near gorilla campaign of small successful semantic projects time to prove themselves while the supporting infrastructure matures.
> The same way that Apple makes sound loops available in garage bands for neophytes to compose into music may be a model for non-ontologists to grab pertinent ontologies to compose into domain ontologies. Then there will be a customer pull for the techologies often discussed here.
the vocal people here are more interested in angel choreography than
actually advancing the use of ontology.
Some lurkers are actually quietly going about making things happen or at
least that is what I get from the occasional post.
This was a hellofa long email to type on a blackberry. Good thing my
flight was delayed.