Re: Have you seen this story?
Kingsley, On Thu, Apr 9, 2009 at 3:09 PM, Kingsley Idehen kide...@openlinksw.com wrote: Aldo Bucchi wrote: Hi guys, I didn't find that post even challenging ( and as some of you might know I really like to argue ), because it makes a fundamental mistake and all drips from there: Do the manufacturers of, say, a new form of carbon nanotubes, use it as material for their own tools? Well, the answer is: not necessarily (and most probably, not at all). At least not in its raw form. It needs processing, it might be more expensive and the tools probably won't make the job better than the old ones. But the material is still better than alluminium, but tools are complex and require other skills that these developers need not necessarily have. It needs to take its place on the low level of a complex industry and value will eventually flourish. This is not different than Linked Data in this context. So, why can someone come to such blunt observation by relating creator dogfooding to the ultimate value of the technology? One could argue that this is closely related to the semantic curse. The answer appears when you try to answer this simple question: * How is this material better? Which inevitably leads you, at least, to: * What do these materials have in common? * What specific qualities of value, present in both, are being improved? We only recently did that for Linked Data! So, the fundamental and shared flaw here has been to attribute a magical, one-of-a-kind nature to something instead of characterizing it in terms of the previously existing alternatives, which results in confusion and... well, what do you expect if we start from there ;) He might be right that there were mistakes, but the real flaws were related to non-specific communication from the SW community ( there was not clear definition of the what is this, what does it compare to and why its better ) and then a lack of deep analysis on part of the writer, who got stuck in his myopia and is calling carbon nanotube developers snake oil salesmen because they don't use the material in their labs. However, I do believe in dogfooding and I do it mostly for personal purposes. But one thing is to support it, another to demand it. OTOH. I like to think that these weren't mistakes. I mean, that the time this project took to lift off due to poor communicational strategies was not in vain. It would have been awfully hard and controversial to explain Linked Data in terms of distributed database technology back in the days. While it would have been certainly understood by a much larger audience, in terms of its development it probably would have entered a state of enthropy and evolved into several JSR kind of process, not to mention strategic oppositions from industry leaders and the inevitable competition ( which, when it comes to standardization processes, is not usually welcome ). Aldo, No argument re. the above, as you know anyhow :-) In more concrete terms. We didn't give M$ a chance to create RDF-MS Edition by staying off the radar. ( I hope so ) Hmm. ADO.NET's Entity Frameworks is RDF-MS salvo #1 Project M is salvo #2. They get it right at #3 and by then it will be them playing well with the Linked Data Web. IE is no doable on the Linked Data Web :-) Ah good examples. Now, regarding IE. Of course it is not the same game so their trick would have been different... But how about Windows Live + Semantic Office + a centralized registry to provide IDs for things. I agree that they will now have to play along with Linked Data web. But they didn't see it coming, if they had... ( well we'll never know I'm just fantasizing here ). Semantic was a great codename, but for the wrong reasons! Great codename for a great Thing. Stinker of a name for discerning meaning of the Thing :-) Semantically correct, but comunicationally impaired ;) Anyway, Semantic Web issues are now water under the bridge in my world view, the big MO is behind Linked Data and we should simply carry this into related and vital realms such as OWL (the TBox side is very important), but do so with pragmatism and coherence. We know this works, based on the Linked Data journey experience. This community has succeeded were alternative approaches have failed. We must never forget that when moving forward. Yes I agree, this is old stuff. I was just looking back for a second. There are so many things I never got to say that I sometimes try to slip them in here. I have to admit that regarding the *Macro* TBox aspect I am only now starting to pay attention. My imagination needs some more training to understand where this is going to take us in the mid term. ( I have groked the ABox aspect for 5 years now, but the other one quickly leads me to AI and out of my comfort zone ). This community is amazing ;) I would put all my chips on the table for you anytime now. ( wait... I already am! ) A. Kingsley
Re: Have you seen this story?
Aldo Bucchi wrote: Kingsley, On Thu, Apr 9, 2009 at 3:09 PM, Kingsley Idehen kide...@openlinksw.com wrote: Aldo Bucchi wrote: Hi guys, I didn't find that post even challenging ( and as some of you might know I really like to argue ), because it makes a fundamental mistake and all drips from there: Do the manufacturers of, say, a new form of carbon nanotubes, use it as material for their own tools? Well, the answer is: not necessarily (and most probably, not at all). At least not in its raw form. It needs processing, it might be more expensive and the tools probably won't make the job better than the old ones. But the material is still better than alluminium, but tools are complex and require other skills that these developers need not necessarily have. It needs to take its place on the low level of a complex industry and value will eventually flourish. This is not different than Linked Data in this context. So, why can someone come to such blunt observation by relating creator dogfooding to the ultimate value of the technology? One could argue that this is closely related to the semantic curse. The answer appears when you try to answer this simple question: * How is this material better? Which inevitably leads you, at least, to: * What do these materials have in common? * What specific qualities of value, present in both, are being improved? We only recently did that for Linked Data! So, the fundamental and shared flaw here has been to attribute a magical, one-of-a-kind nature to something instead of characterizing it in terms of the previously existing alternatives, which results in confusion and... well, what do you expect if we start from there ;) He might be right that there were mistakes, but the real flaws were related to non-specific communication from the SW community ( there was not clear definition of the what is this, what does it compare to and why its better ) and then a lack of deep analysis on part of the writer, who got stuck in his myopia and is calling carbon nanotube developers snake oil salesmen because they don't use the material in their labs. However, I do believe in dogfooding and I do it mostly for personal purposes. But one thing is to support it, another to demand it. OTOH. I like to think that these weren't mistakes. I mean, that the time this project took to lift off due to poor communicational strategies was not in vain. It would have been awfully hard and controversial to explain Linked Data in terms of distributed database technology back in the days. While it would have been certainly understood by a much larger audience, in terms of its development it probably would have entered a state of enthropy and evolved into several JSR kind of process, not to mention strategic oppositions from industry leaders and the inevitable competition ( which, when it comes to standardization processes, is not usually welcome ). Aldo, No argument re. the above, as you know anyhow :-) In more concrete terms. We didn't give M$ a chance to create RDF-MS Edition by staying off the radar. ( I hope so ) Hmm. ADO.NET's Entity Frameworks is RDF-MS salvo #1 Project M is salvo #2. They get it right at #3 and by then it will be them playing well with the Linked Data Web. IE is no doable on the Linked Data Web :-) Ah good examples. Now, regarding IE. Of course it is not the same game so their trick would have been different... But how about Windows Live + Semantic Office + a centralized registry to provide IDs for things. I agree that they will now have to play along with Linked Data web. But they didn't see it coming, if they had... ( well we'll never know I'm just fantasizing here ). In a nutshell, Microsoft has always been interested in Save To Web, this is why they took to WebDAV so quickly etc.. Thus, they will certainly play well re. Save To Linked Data Web and Annotate Linked Data Web, which is already taking shape in the realm of Chemistry. They must be tired of having their own lunch served back to them (Ajax and Smart Tags are classic examples). As for Office in general, they have formal object models that are easily mapped to ontologies, and you will see a lot of this exposed in the RDF-ization realm (us and others). Personally, I want you to be able to Save To Data Space(s) that are Internet accessible and Web friendly (re. representational dispatch). I want total data irregularity and maximum hyper-orthogonality (quoting: Ted Nelson). I never liked paper! I never like the restrictions of the DBMS engines or Spreadsheets, and I hated being teased by Lotus Notes. This is why Virtuoso is a Universal Server that exposes Linked Data Spaces :-) Kingsley Semantic was a great codename, but for the wrong reasons! Great codename for a great Thing. Stinker of a name for discerning meaning of the Thing :-) Semantically correct, but comunicationally impaired ;) Anyway, Semantic Web
Contd: Have you seen this story?
Aldo Bucchi wrote: Kingsley, [SNIP] Anyway, Semantic Web issues are now water under the bridge in my world view, the big MO is behind Linked Data and we should simply carry this into related and vital realms such as OWL (the TBox side is very important), but do so with pragmatism and coherence. We know this works, based on the Linked Data journey experience. This community has succeeded were alternative approaches have failed. We must never forget that when moving forward. Yes I agree, this is old stuff. I was just looking back for a second. There are so many things I never got to say that I sometimes try to slip them in here. I have to admit that regarding the *Macro* TBox aspect I am only now starting to pay attention. My imagination needs some more training to understand where this is going to take us in the mid term. ( I have groked the ABox aspect for 5 years now, but the other one quickly leads me to AI and out of my comfort zone ). But it will be a comfortable zone (as its value will become much clearer) now that we have a Linked Data substrate :-) This community is amazing ;) I would put all my chips on the table for you anytime now. ( wait... I already am! ) LOL! Kingsley A. Kingsley Regards, A On Thu, Apr 9, 2009 at 10:00 AM, Tom Heath tom.he...@talis.com wrote: Hi Daniel, 2009/4/9 Daniel Schwabe dschw...@inf.puc-rio.br: Dear all, this may be old stuff, but I was surprised to read http://www.intelligententerprise.com/blog/archives/2009/02/semantic_web_sn.html... Me too! He does have some points... In 99% of cases with respect to me he doesn't ;) As I say in my response on his blog (copied into that post of mine that Juan refers to) I agree that we, the Semantic Web community, have not always done as much as we could in the dog food department, but that has been changing rapidly since 2006 and we should keep up/increase the pace. I won't comment on that blog post any further here; it's already sapped too many hours of my life :) Cheers, Tom. -- Dr Tom Heath Researcher Platform Division Talis Information Ltd T: 0870 400 5000 W: http://www.talis.com/ -- Regards, Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen President CEO OpenLink Software Web: http://www.openlinksw.com -- Regards, Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen President CEO OpenLink Software Web: http://www.openlinksw.com
