I think you have a valid point; however, I think you may be discounting one form that the semantic web might take.

Right now, engineers and programmers are trying to design large complex systems to take all the information on the web and try to process it in the same way that a human might do. They're trying to make the "Holy Grail" of semantic web apps - the general, all purpose "sorting" and "feeding" device that will slice, dice and process content for a wide range of uses, a kind of "Google search appliance" that will be the "killer app" for Web 3.0.

In effect, they're trying to use statistical models to create something capable of processing and decision making in much the same way the human brain can do and seem to think that there are mathematical patterns that will become apparent in these relationships if they just dig hard enough. The Netflix Prize, where most of the teams are using this approach, is a good example of this - it can get you so far in recommending movies to users, but comes up against all the intangible things about the human brain and quirks of the content itself and hits a "brick wall".

My own thinking is that the semantic web will actually be several different applications that each take a look at a specalized "slice" of the pie, where the particular qualities of the content one is looking at or the desired information one needs is taken into account.

The model I'm developing, based on the academic field of Psychohistory, would look at psychological aspects of content and how it could be used by advertisers, marketers or in finance. It works for very specific applications, but isn't useful for others - human psychology and the language associated with it has a particular set of parameters associated with it. It becomes meaningless if you try to apply the model to general searching for web sites or pushing business content to users, for example.

posted February 1, 2008

Francisco: What comes first, the logic or the conclusions (or what's in a game)? You might soon have a system with a complex enough set of rules that approximate the effect of a dedicated group of people. The choice of rule sets is a form of conclusion. Finding feasible sets is of interest. It might not be called semantic web, but an effort toward trusted sources might still exist. Thanks.

Links:

• http://cadsmith.blogspot.com/2008/02/game-testing-all-in-one-schultz-others...

posted February 2, 2008

Experience is the key to time. Without experience our own mental actions suffer and begin to deteriorate. When one isolates prisoners into solitary confinement, or uses any form of sleep deprivation, or sensory deprevation, personal cognizance suffers, emotions blur and reason fails. It is torture without bruises.

In considering computational reasoning webs, semantic or otherwise, one is considering the stimuli (inputs), and responses (actions) as indicators of effective reason. Computer programming is often done with the conclusion and working backwards to a solution matrix, and it is that form of functional programming that can never duplicate the mind. Neural nets are one example where the solution evolves, generating the solution spontaneously. However our endeavors in this area so far have been primarily focused on specific solution sets, and not on the general matrix, thus the "training set". However, this is a baby step toward self organizing systems, and a web or network of potential actions based upon results optimization.

Another reasonable sense of what is possible is the recent work of voice recognition. When neural nets are applied to tonal and envelop patterns (certainly a time function of probabilistic nature), usable results can be obtained.

If you choose to discuss the philosophy of the human brain, or the relevance of man, a computer will probably never relate, any more than we can relate to the direct communications or tools used by animals. Nor would it understand our proclivity to choose others as friends while excluding others as enemies.

Postulating a reasoning machine, one with understanding, learning ability and capacity for emotion is not without possibilities, and at the same time, the gamut of existence, the issues of resolution of conflicts in the eventual relationships that evolve will no doubt give rise to machines with inferiority complexes, multiple personalities and other traits, although their manifestation will be foreign to us as we have no relevant experience to match that of the reasoning machine.

Whether the machine would ever pass a Turing test or the Chinese room is very dependent upon our recognizing the device as human, yet even if we choose someone that is not of our own language and culture, we would be hard pressed to understand the options of choice that various viewers would make in those tests. That is the frame of reference of the participant (the occupant of the room) affects the outcome.

Just as the Christians of the 15 and 1600's misinterpreted many of the friendly actions of various island cultures in the south pacific, culture, experience and perception limit understanding, cognition and relevance.

And while you look at the theoretical boundaries of such concepts as the semantic web, or even the upcoming societies of virtual reality, the reality is that the machines, the environment and even the people continue to evolve, and it appears that the evolution is and will continue to accelerate for the foreseeable future.

So while the discussion of what is possible, or where the walls stand or what they mean is fruitful, their exposition is only one possible view, of the millions or should I say billions possible on this planet of ours.

Regards,
Les H

posted February 6, 2008

There is some confusion or obfuscation in the definition of terms and the statement of the problem.

First, semantic technologies do not have to achieve the ability to form conclusions like the human mind to be useful.

Second, to state that semantics lack the dimension of time is, a fundamental misunderstanding of the concept, in my opinion. The term "semantic" refers to the relationship between terms and the things they refer to. In plain speak, it's about meaning. Terms have meaning because we ascribe it to them. We pass the meaning down through generations. Some meanings change, others remain unchanged, and collectively define human cultures and existence. "Sky" is "sky" the same today as it was for the first human to look up. Our understanding and technical knowledge of "sky" has changed, but we still understand sky at some fundamental level as the place where the clouds are. Ask any three year old.

In that sense, semantics actually introduce time to the equation. Semantics provide an anchor point in time. Democracy is today what it was to the ancient Greeks. Not in exact form or implementation, but in concept and core meaning--rule by the people. While technologies may have changed (such as voting machines, polling techniques and the like), the core concept of the people choosing their government remains.

As for forming conclusions, I don't know if that's necessary. The human mind that established the relationship between terms (metadata) and content (data) already formed the conclusions. The semantic technologies just provide a mechanism for automating the handling of the data according to those conclusions.

posted February 7, 2008

"The nice part about being a pessimist is that you are constantly being either proven right or pleasantly surprised."

George F. Will
US editor, commentator, & columnist (1941 - )

posted February 8, 2008

Well after reading the replies so far I'm coming to the conclusion that in reality there will not be a single "semantic web" that everyone will find useful (it will be tried but it will fail due to differences in "meaning" or interpretation)........but there will be one or more semantic webs for each user or like minded user group where meanings are broadly similar. Sounds like "personalisation" with attitude.

But I'm probably wrong.
Nigel

posted February 10, 2008

First I have to say that I think (therefore partially agree) that there's no 'machine' (A formula, a calculation or a system, even if it's cloned) that can mimick a (or any) 'human' entirely.

And then I have to STRONGLY disagree with your statement that I have quoted from your question, which gives me the feeling that you barely have any idea neither on the basis which natural language processing is originally established nor the AI efforts toward it.

Quoted from your original question : ``So all efforts faced toward semantic web are doomed to failure a priori if the aim is to extend our human way of thinking into machines, they lack the metaphorical speech, because only a mathematical construction, which will always be tautological and metonymic, and lacks the use of the time that is what leads to the conclusion or "stop". ``

I'd recommend you to read AI and NLP theory (not Neuro-Linguistic Programming but Natural Language Processing), and I'm also providing a link for you down below about Artificial Neural Networks, so that you can have an idea about how the 'human way of thinking' is imposed to the machines, for starters.

Some more keywords for you will be "Interlingua" and "Computational Linguistics" if you have any time and if you're really interested...

Links:

• http://en.wikipedia.org/wiki/Artificial_neural_network
• http://en.wikipedia.org/wiki/Natural_language_processing

posted February 14, 2008

Hello,

The limitations of the Semantic web relate primarily to the lack of several key constructs that would support the type of automation that is needed to make it work. The standards referenced, OWL & RDF do not provide for a reference architecture nor the actual best practices involved in providing semantic reconcilation (or integration). The focus on Shared Upper Level Ontologies does not recognize the need for the merger of semantic elements into sets and the possibility of dynamic or context driven ontology. The key thing to keep in mind here is that ontologies are not all written in stone - in other words they represent community and individual perspectives and as such must have the flexbility to evolve.

It is also important to recognize that Semantic Integration can and should be applied to the enterprise as well as the Internet.

Regards,

Stephen Lahanas

Links:

• http://www.slideshare.net/slahanas/semantic-integration-v1
• http://www.semantech-inc.com/index.html

posted February 14, 2008

Why do so many think the human way of thinking is the do all and end all of higher computational "thought processes"? No "metaphysical requirement" is needed when it comes to how our physical brain functions.

That's a quantum computer between your ears! The only metaphysical aspect of human brains is; we still do not understand how it really works. As soon as we do, we will create an AI that will definetly hurt a few egos.

posted March 2, 2008

Let me start by saying that you are probably right about your thoughts on the feasibility of the semantic web and weather the “machine world” will ever be able to mimic the un predictability of man or even be able to reproduce the lines of thought associations that are constantly running in a human brain and that are logic to only that individual. I don’t think we will ever be able to mach this, but we might be able to create something that is good enough to be better than the old fashion way.

My biggest fear is that really becomes possible to create the truly semantic web, that gives you exactly what you are looking for in the way you want and expect it. This would take something very important away, and that is the option to find something you are not looking for, to be presented with something that you did not expect in a way you have not thought about.

Some of my best experiences on the web and in the real world have been centred on me getting, finding or trying something that I didn’t know, thought I didn’t want or like or wasn’t looking for. All the best presents I have ever gotten have all been the ones that I didn’t wish for in the first place

posted March 3, 2008

I too am a "rugged epistemologist", but my hardware can't quite calculate why!

Links:

• http://en.wikipedia.org/wiki/Epistemology#The_regress_problem

posted March 3, 2008

Dear Francisco,
Oren Yosifon puts it quite right, the challenge is not to create a "replace human X" (fill in your favorite X) but to see where the real limits of a given technology are. That is already a very tough question. 40-some years of research in AI have been very productive, not to create the publicly widely misunderstood "artificial intelligence = human intelligence" but to create a host of sophisticated planning, decision, reasoning and modelling tools on which many technologies today rely.
The most challenging bridge to gap in the area of ontology in my field (computer-aided architectural design) is a consistent (or least-error prone) translation between the fluid use of words by architects in the design process and ontologies as used by information systems. So no replacement stuff but elimination of a whole of tedious activities that distract architects such as information search, validation, and verification of results.

Hope this is helpful,

kind regards,

Henri Achten

posted March 4, 2008

So what is your question? What I read was more of a thesis statement.

posted March 10, 2008

The Semantic Web is flawed... this topic is widely debated around the world and many others have structured a much clearer picture of this issue than you, me or any other individual in this QA, though I didn't really even see a question here, more like a pedantic regurgitation of materials read. Boiled down, "meanings are not fixed and tied to context". The Resource Descriptive Framework is unambiguous, while the ambiguity of natural language is fundamental. Part of the debate goes... there exists unambiguous languages, some subsets of mathamatics, certain programming languages... Beyond all that... Certain Semantic Web initiatives claim they will capture knowledge, which is much more significant than just establishing a Web ontology based on unambiguous languages. Most human knowledge exists in tacit form, without language all together. However, while knowledge and semantics are the widely used terms associated with the Semantic Web, in actuality the W3C does not use either of those terms to define it. The term they use is data. Language is used to define the data that establishes each individual's ontology. The Semantic Web is then a knowledge base of every user's data ontology, to establish an overarching Web Ontology, which is referenced by individual semantics.

Here is Tim Berners-Lee's statement on the name "Semantic Web"...

"I don't think it's a very good name but we're stuck with it now. The word semantics is used by different groups to mean different things. But now people understand that the Semantic Web is the Data Web. I think we could have called it the Data Web."

Clearly not everyone understands this.

Much of this argument in respect to AI can be found in the Neats vs. Scruffies labels given in AI research to how intelligent systems should be designed (wikipedia article link attached). Many involved have found successes in combining the two approaches. This issue addresses your concerns about computational intractability and the Computational complexity theory. This leads to what the Web really is, a Complex Adaptive System (link attached).

While many of the points you pulled together have some validity, it is unfortunate that you've fixated so much on the element of time, language and the flawed notion that the Semantic Web is supposed to "replace thinking" or "extend human thinking into machines".

I suggest you read up on some of the works published by Marvin Minsky (link attached). Especially with respect to that completely flawed theoretical "demonstration" that the Semantic Web 1) should be viewed as "a language with capabilities similar to human language" and 2) "proven false based on the Turing machine's halting problem". The ridiculousness of that claim is hard for me to even repeat let alone swallow. I'm failing to see your technically expertise on the Semantic Web (or computer science) and what makes you believe you will "prove the Semantic Web does not exist" as if you're establishing some kind of thesis.

The purpose of the Semantic Web has never been to correctly compute the answer for all possible inputs with respect to the human language. As I said in the beginning, there are many issues with the Semantic Web, its theory and practices that are being questioned, but not one point you raised truly touches upon this or demonstrates that you have a clear understanding of your topic. I wish there was some semblance of an actual question in this posting to answer, because instead I'm left picking apart your biased and wildly drawn conclusions.

Links:

• http://en.wikipedia.org/wiki/Neats_vs._scruffies
• http://en.wikipedia.org/wiki/Complex_adaptive_system
• http://web.media.mit.edu/~minsky/

posted March 11, 2008

It appears that you are trying to make a point about human intelligence vs. artificial intelligence rather than about the semantic web.

You (at least the text you posted) state(s) that human reasoning cant’ be captured by any formal logic because humans always arrive at a conclusion in a finite amount of time, even when the reasoning is inconclusive. In my professional field, Law, this is actually very common.

Though your assumption about humans only holds for the pedantic, you would have a point if the conclusions thus reached would always be correct. This is, however, not the case. In my professional field, Law, this is actually very common.

I think it is entirely feasible to write a computer program that presents a possibly incorrect conclusion after a finite amount of time.

posted March 11, 2008

I have 2 questions:
First, I don't quite understand tne necessity of the assertion that because the Semantic web is limitted to purely formal logic which does not 'contemplate' (consider) the concept of time a reasoning machine will be led to inconsitencies. And somehow that a lack of consistency will prohibit a machine from ever achieving the capacity of human reason?
And second with regard to time. Time, by my lay understanding, is a vector in a space. In a historical analysis it becomes a vector, usually poorly fitted, to the transformations occurring in the observed data series by way of statistical methods. In the case of reasoning systems like the human mind that observe initial states, inputs and outcomes and use the resulting networks to postulate potential outcomes for partially matching inputs and initial states time is but a vector by which the state transofrmation is sequenced.
As I read through the question and the many learned replies I was tickled to think that time is generally an arbitrary mathematical vector applied not necessarily for fitness of purpose, but rather becasue it is a familiar presumed constant. Where as the relations that exist in the semantic web are generally unquantifiable and subjective to the nodes in that web. Those arcs are the observed properties of the space as defined by the members of the space, the position of the observer, and the motivation or motion of the observer. I find irony in that that seems so very much more human, more subjective than any relation presumed to be defined by a univeral constant with direct numerical properties.
The web, semantic or otherwise is indeed stateless and as such very much lacking in a time vector. (Excepting the cases where that information is supplemented by meta data.) It is not however lacking in vectors by which a reasoning machine might assemble relations. Is the act of reasoning unlike applying the transform learned by observing the change in a machine state ( e.g a mind, a model, a taxonomy, a map, etc ) over a vector (arbitrary or subjective) to anticipate the resulting state when a set of values are applied to the inputs of that machine state?

posted March 11, 2008

Maybe for the interested people among us, you can check the rest out on the URL at the bottom.

THE SIXTH EUROPEAN SUMMER SCHOOL ON ONTOLOGICAL ENGINEERING AND THE SEMANTIC WEB (SSSW-2008): REGISTRATION NOW OPEN

VENUE
Cercedilla, near Madrid, Spain.

DATE
6-12 July, 2008

DEADLINE FOR PROVISIONAL REGISTRATION
20 April, 2008

INTRODUCTION
We are pleased to announce that **REGISTRATION IS NOW OPEN** for the
sixth European Summer School on Ontological Engineering and the
Semantic Web. This summer school, presented by leading researchers in
the field, is the successor to the well-received schools held in 2003,
as part of the activities of the OntoWeb Network Consortium, and in
2004-2007, as part of the activities of the KnowledgeWeb Network of
Excellence. The school represents an opportunity for postgraduate
students to equip themselves with the range of theoretical, practical,
and collaboration skills necessary for full engagement with the
challenges involved in developing Ontologies and Semantic Web
applications. To ensure a high ratio between tutors and students the
school will be limited to 50 participants. This year the school will
be sponsored by a number of European projects including LUISA, NEON,
SUPER and X-MEDIA and also by STI International.

http://kmi.open.ac.uk/events/sssw08/

posted March 12, 2008

Francisco - people will think I paid you for this question!

Here are some thoughts that support Nigel Deighton's answer to your question. He believes "there will not be a single 'semantic web' that everyone will find useful (it will be tried, but it will fail due to differences in "meaning" or interpretation)........but there will be one or more semantic webs":

The Semantic Web being discussed by the W3C (World Wide Web Consortium) is a difficult proposition for sure! The volume of work required to create ontologies that represent human knowledge is mind boggling. So, I am in agreement that a "Bottom-Up" development of a useful Semantic Web is suspect.

However, a "Top-Down" approach is not only possible, but alive and kicking today! What I mean by "Top-Down" is the recognition that certain raw materials are available to create "Semantic Web Light". The best explanation I've found about what we're really expecting from Web 3.0 was in a blog by Alex Iskold - 'Top-Down: A New Approach to the Semantic Web' (September 2007). Alex says, "Here is what we are really looking forward to with the Semantic Web:

* Spend less time searching
* Spend less time looking at things that do not matter
* Spend less time explaining what we want to computers

A consumer focus and clear benefit for businesses needs to be there in order for the Semantic Web vision to be embraced by the marketplace." This can happen without a standardized, ontological, Web-wide implementation. It can happen one website at a time. Where each website owner recognizes specific market segments, learns to identify the presence of a visitor representing a segment and displays more of what that visitor is likely to care about and less of what they don't.

I believe my company is delivering on that promise today.

We were in stealth mode with our Semantic Web technology until January 1, 2008. We secured patent pending in November 2007 for a business method that enables licensees of our product, Semanticator™, to create semantic personas or profiles based on pre-session, detectable attributes. Then, once a persona is detected, websites using our technology are able to deliver a more meaningful experience. We call this Semantic Marketing.

Our oldest implementation launched in August 2007 for the automotive industry. With two personas, we were able to impact 62% of visitor interactions (so you don't have to go nuts building out personas to be effective). Of those 62%, we have seen a 41% increase in time on site, a 30% reduction in bounce rate (visitors who enter and exit the same page) and a 26% increase in conversion!

So, my answer to whether the Semantic Web is viable has to be ABSOLUTELY. But, like Randy Riddle said in his answer, there will be brick walls. What we have to do is use these massive human brains that some are trying to mimic, look at it differently and knock those walls down. If you want to join others discussing Semantic Marketing, please check out the Linked-In Group or the Google Group. We'd love to hear your opinions!

Links:

• http://www.readwriteweb.com/archives/the_top-down_semantic_web.php
• http://www.semanticator.com
• http://groups.google.com/group/semantic-marketing

posted March 12, 2008

When you say "semantic" are you implying meaningful?

How can a diverse collection of interrelated and wholly unrelated information be meaningful? Unless, of course, it is a reflection of humankind.

posted March 19, 2008

Don't know who said this first, but asking if a computer can think is like asking if a submarine can swim.

The semantic web is not trying to be AI. That will be Web 6.0.

It's not about getting the web to think, it's about organizing an institutional memory in a more meaningful way. It won't be perfect, and it can't duplicate the way human minds work. Relationships are based on properties, and many of the properties of any object or concept are really perceptions - beauty is in the eye of the beholder.

But it has the potential to be better than what we have now, which is a fairly haphazard relationship between memes based on caprice, a billion personal opinions with no logic directing them, and attempts to rank number one on Google.

The semantic web will provide the platform for the next generation of the internet. It's a step on the road, not a destination.

posted March 21, 2008

I would venture to say that the Limit for what we can do with technology is way beyond what our current brain capacity even is. I'm not worried about a "limit." The idea is to keep pushing it and profit from it.

James
from
http://FaceySpacey.com - "Your One Stop Social Media Shop"

posted March 21, 2008

First I want to apologize I only browsed the article, I am going to read it more calmly soon)
But anyway I wanted to congratulate you for the idea of formally demostrate that semantic web is doomed to failure using only computers. I bet you are right, If I can help with the demostration let me know. I a pretty good math background, among other skills.

Following for what about I understood about goedels theoreme is that it formally demostrates that every system/paradigm such as logic and math need to be based on arbitrary assumptions, like 1=1. And the other point I guess you are refering to is that in self referential systems (such as language) it is impossible to avoid inconsistencies, the only way is to have a meta rules. Guess you read Hoffstater books

(excuse me misspellings, I am foreign).

On other thing I always think is that the computer is not going to be able to make it ever, it is just a sofisticated calculator, no awarnes, no autonomy, etc. And we are still lacking lots of neuroscience knowledge to even imaging how we manage to do the things we do.

But still I think this techniques are pretty usefull
Of course is the same old same old, never as good as a human being but, by having to have certain domaing knowledge, in this case language, and try different techniques (HMM,N-grams, Taggin, ,fuzzy NN, etc..) and search for the best value of the parameters, such as the influence of inertia while searching for a minimum or maximum.
decide proper balances between accuracy and response time,avoid unnecesary overlearning, apply different euristics and many other tunings that are needed to tune the programs in the best possible way. you wont have a perfect result, but it will be pretty usefull for humans since a large part of the work is done by machines.
I came from classis AI, so do not think I specially like this way of doing things, but I will delighted to hear some new ideas.

I did not get your last sentence about the turing machine and the halting problem, is it about demostrating that this theory is impossible using a technique like the imposibility to determine if a program will ever stop using another program to supervise it, or am I on the wrong path?

Sorry If I was not pretty clear, I am overwhelmed by work. But I am very intrested in your theories and ideas.
I have some other new things I am examining, that I can chat with you about some other time.

Best Regards
MS

posted March 22, 2008

I used to think like you.

Then I read about Description Logic, and OWL (Web Ontology Language) and changed my mind. OWL Full takes the paradigm of Open World Assumption: under this Open World Assumption, if a statement can not be proved to be true using current knowledge, we can not draw the conclusion that the statement is false. Failure to derive a fact does not imply the opposite.

The language of logic programs with strong negation allows us to postulate the closed world assumption for some predicates and leave the other predicates in the realm of the open world assumption.

open world assumption and its opposite, the Closed World Assumption, apply within the field of Knowledge representation to express the extent to which knowledge within a system is viewed to be complete. Completeness allows inferences to be drawn from what is not recorded in the system, while incompleteness prevents this. The open world assumption or OWA asserts that a system's knowledge is incomplete, so that if a statement cannot be inferred from what is expressed in the system, then it still cannot be inferred to be false. Heuristically, the open world assumption applies when we represent knowledge with a system as we discover it and the reality described by the system can never be known to have been fully described. In the OWA, statements about knowledge that are not included in or inferred from the knowledge explicitly recorded in the system may be considered unknown, rather than wrong or false.

Semantic Web languages such as RDF(S) and OWL make the open world assumption.

So, while It is unlikely that any reasoning software will be able to support complete reasoning for any language, the same holds by the uncertainty principle predicted by Goedel, and in the end both systems can be seen as equipollent under this point of view.

As for Turing's halting problem, QWL Full has been developed to avoid it using defined semantics to help it. This, and the OWA, should offer guarantees of a coherent, non-contradictory system.

posted March 22, 2008

To expand on what Henri Aachen said, "40-some years of research in AI have been very productive, not to create the publicly widely misunderstood "artificial intelligence = human intelligence" but to create a host of sophisticated planning, decision, reasoning and modelling tools on which many technologies today rely."

Unfortunately many concepts and critiques of the "Semantic Web" are warmed over from debates in the history of AI, and the lessons learned there are important.

There's a major current in AI recently that re-invigorates early work in decision theory and probability modeling, where the uncertainty in any empirical fact is primary, and the idea of a proof is replaced with inference in a Bayes framework. Comparisons with human ability are fraught, and recent successes tend to mimic pre-concious perceptive abilities. The hard problems that can be solved are e.g. autonomous navigation, image understanding, and -- relevant to the Semantic Web -- some aspects of NLP such as machine translation.

One view of AI is that is is a way to incorporate semantics (that is, people can attribute meaning) into the automated tools we build. Humans will continue to excel as tool-building creatures as we have for millenia, despite the dubious myth of the possibility to create "intelligent" entities.

Links:

• http://www.nips.cc
• http://en.wikipedia.org/wiki/Judea_Pearl
• http://www.cs.ubc.ca/~murphyk/Bayes/bnintro.html

posted March 26, 2008

Your attempt to prove that the semantic web does not exist cannot succeed because it in fact already does exist. The “Semantic Web” exists as a concept that implies a community. The community encompasses those currently in discussion about the concept as well as knowledge discovered and created under its aegis. The development and distribution of automated processing machines that reduce associated concepts to practice provide means for employment of those in the community, its structures and its processes.

The “Semantic Web” exists because a group of like-minded individuals imagined its existence. You can argue the concept's academic and economic virtues and will no doubt find diverse views.

With apologies to Schuppanzigh and Beethoven, "Do you think I worry about your wretched proofs when a concept intrigues me?"

Links:

• http://www.la-jolla-music-society.com/content/view/732/125/

posted April 5, 2008

Ye gods!

Which consultant wants the gig to set standards for that then!

posted April 19, 2008

I think that "ontology" as commonly used means an Aristotleian ontology, which suffers from confusion that metadata is substantial in the same way as actual data. My opinion is mostly informed by close reading of Heidegger's Being and Time. The problem is that meaning is interpretive, and it varies sometimes continuously and sometimes integrally, and that variance always has a significant subjective factor: what does the beholder choose to ignore/attend. Can the semantic web be defined as a recursive convergent system?

posted April 25, 2008

Semantic refers to a mapping between terms and experiences humans use to give meaning to those terms, that is representational (vertical), and a mapping (a la Saussure) between terms and other terms whose use define that term relationally (horizontal meaning). So we can map time just as well as anything else. Semantic web has NOTHING to do with formal logic or philosophies of any sort (though people who make a living doing philosophical critiques of technologies--Searl, Fodor, and lesser lights with insist otherwise). The Semantic web is a human defined step, with the size and exact nature of the step up for grabs till some standards are agreed on. The average step being now suggested is labeling all data on web pages (and behind them in databases) as to the functional domains it can support useful to user operations in--addresses will be mapped onto location finding, physical delivery, and similar functions, etc.

There is nothing particularly grand, grandiose, or philosophic about this. To say that "the Semantic Web is impossible" is exactly, logically, the same as saying "Street Addresses in Geneva Switzerland are impossible". Both statements reveal a COMPLETE lack of reading about the Semantic Web and a complete lack of attempts to understand it, even at the beginner level.

In other words, do your homework first, then post questions.

posted April 26, 2008

I quite agree with your thoughts. There are serious obstacles to Semantic Webs we may not be able to overcome in the foreseeable future. So if we speak of some omnipresent web of semantics, this would be really very hard to achieve. It all depends on how we define the Semantic Web. Still, there are specific, practical areas where semantic apps are possible and such applications are being built...

One more thought. If human brain and psychic could be represented by mathematical models, the psychology, linguistic and other human centric disciplines would be exact sciences. Which hasn't been the case during the last several thousand years.

posted April 27, 2008