Monday, August 22, 2005

On Web 2.0 and Ontologies

I have just searched "python" in Technorati. Suddenly, pictures from flickr bearing the tags "python" were presented to the right of my search results screen, along with lines from other blogs. Though I knew this feature of technorati/feedster, it was the 1st time I really saw it in action.


Web 2, with del.icio.us, last.fm, flickr and more is wrapping us. The change is so fundamental; Internet is actually becoming a river of digital consciousness, just as I wrote couple of years ago in an article about Microsoft Collaboration practices. Actually, what Microsoft tried to do in the collaboration area was something similar to web2.0, but it lacked something fundamental that I feel inside my veins: it lacked real people.

Today we see how web2 forms in front of our eyes. I am amazed and to that I add a bit of puzzlement. Ontologies are under serious attack; serious - because the attacks are managed by smart people with real arguments. Clay Shirky is one of them but I'll give an unexpected example: Adam Bosworth, Google VP of engineering (can be heard at IT Conversations). And there are more.
I'll sum up my intuitive understanding of their arguments: Ontology is reasonable when reality is objective. When Objective - bring a "reality domain expert" and he/she will describe reality (in triplets or quadruples - who cares…). And that's the paradigm of web 1.0 - we're putting stuff on the net and people are viewing it: passive to the objective digital world that was previously uploaded into the net by some "experts". But web 2.0 proved the ontological approach to be ridiculo.us! How many different tags (i.e. classifications) a single real entity (URI) has? All tags bearing different meanings to different people in different contexts. Web 2.0 is the post-modern of the web 1.0 modernism: there's no absolute truth, my interpretation of a URI is no better no worse than yours. In that case, objective description of reality in the form of ontology is pointless. Bosworth discusses very briefly RSS 1.0 which was based on RDF. He claimed it failed because it's complex and people don't think syllogism (that's from Clay…). RSS 1.0, he said, was the result of a standard committee, while RSS 2.0 was the outcome of a single man initiative (like Joshua Shachter's from del.icio.us). The differences are clear-cut: standards (IEEE etc) are formalization of objective world - objective but highly compromised! Web 2.0 has no standards! Standards and web 2.0 are almost antipodes. No standards and no compromises. I'm not taking position here!!! I'm not saying it's good or bad. But, post-modernism brought a lot of good things along with a lot of bad things, such as the potential "eradication" (that's probably extreme…) of solid moral ground. Web 2 has a potential to bring over huge moral problems - I am not aware of anyone who tackles this issue. When every point of view is legit, no point of view is valuable!

Also, a good question is what will be the future of standards. We, architects, are in a strange period, where standards are treated in contempt - and I agree they are highly problematic, complicated and useless... Bosworth, for instance, gives his pitch against the WS-* - they indeed are the most glorious yet ridiculous bunch of standards. How many of them exist? No one counts. And yes, they are as complex as RDF.


5 Comments:

Blogger Udi h Bauman said...

Very nice & informative post, thanks.

I don't see the morality problems, because I see the bottom-up subjective approach as an embodiment of democracy, in its ultimate form, like Popper's open society. Eventually, when the voice of everyone is heard & aggregated, the voice of the collective intelligence is heard, which I believe is likely to be more moral than a dictated morality.

Standards are being replaced by ad-hoc standards. See for example microformats.org, which do a great job in identifying & defining them.

The ultimate Web2.0 will be the semantic web. Taggs provide some semantic, but OWL, SWRL & reasoners will truly turn the Web into a digital concsiousness, & bring amazing collaboration between its nodes (humans, agents & services).

6:48 PM  
Blogger Udi h Bauman said...

More explicitely, I don't see the ontology & Web2.0 approaches as opposite but rather as complementing. I can see personal OWL ontologies, from which you get diverse opinions, but aggregated knowledge, information & services.

6:51 PM  
Blogger Muli Koppel said...

Hi Udi,

Thanks for the microformats. I'll look into this.

Regarding the ultimate web 2.0: you should confront yourself with the difficulties Clay Shirky's putting vehemently in his article " The Semantic Web, Syllogism, and Worldview " (http://www.shirky.com/writings/semantic_syllogism.html).

But let's tackle something we both doing a lot: tagging in del.icio.us. What kind of AI do you need to grasp their meaning? To understand they all refer to the same URI? This is an extremely difficult problem. Most of the Web 2.0 collaborative infrastructure suppliers explicitly claim they will not deal with meaning – they are leaving it to the users.
Semantic misunderstanding is something fundamental to humans. Austin's book on Pragmatics points brilliantly to the problem. I'm citing here from a short overview of the "problem" as described in "What is Pragmatics?" (http://www.gxnu.edu.cn/Personal/szliu/definition.html):
"The ability to comprehend and produce a communicative act is referred to as pragmatic competence (Kasper, 1997) which often includes one's knowledge about the social distance, social status between the speakers involved, the cultural knowledge such as politeness, and the linguistic knowledge explicit and implicit".

The challenges facing the Machines in their attempt to understand the above contextual considerations based on facts listed in ontologies are serious enough. Add to that a Babylonian tagging, where almost every individual has its own ontology and you get a mission impossible.

I know you can solve it, though... :)

7:46 PM  
Blogger Nikhil said...

The question is going to be who needs to understand the semantic web and how. With strikeiron and other services leading to the exposure of core capabilities to other machines (as WS) composible solutions are going to become the norm and the semantic web is going to become instrumental in describing that from a shared service perspective. From a usable, customer interaction perspective, its going to be more Web 2.0 - where the end results are shared. I already know companies that are driving to SWRL and shared ontologies - that will be enablers at the shared service level. The composition of this and the using of this is going to be the customer facing end of the web.

Tim O'Reilly provides an interesting article http://www.oreillynet.com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html?page=1

The implications of this from a business model perspective are going to be phenomenal. I also think that reuse is going to be the CNC machine of the IT world..

Cheers,
Nikhil

12:04 PM  
Blogger Muli Koppel said...

Hi Nikhil

thanks for your comment. Yes, the semweb is inevitable. And I invite you to have a look at my post on Faustian deals and magical clipboards - Ray Ozzie's live clipboard wouldn't work without a shared ontology. And yet, if I understand correctly, ontologies, like standards, are about to be micro-formated, meaning the definition of what's there will stem from one or two subjects that would agree first and the others would come next. That's different than the current situation, where committees define a "Customer" for years.

cheers
muli

4:52 PM  

Post a Comment

<< Home