Mailing List Archive: 49091 messages
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search
 

The Semantic Web - A solution?

 [1/10] from: depotcity::telus::net at: 12-Apr-2001 0:08


Personally, I don't see how XML and the Semantic web will gel? It would require endless streams of coding to get my Rebol powered agent (http://www.LFReD.com ) to get all these disconnected ontologies and tags etc etc. to work together. I don't know, maybe I'm missing something, but the complexity of it all is daunting. The more complex it becomes, the less likely the majority would/could be bothered with it. Carl said "It's better to conquer first, then explain why later." I propose a "Rebol Semantic Standard"... a universal Rebol "ontology" if you will, that at least Rebols can agree on. What do you think? Terry Brownell Orion Alliance Inc. Some additional thoughts... The basic premise of the Semantic web is sound... subject, verb and object. I tend to use subject and predicate for simplicity and scope. Lets look at Carls example... mix until well distributed Now I don't know about you, but the word that comes to mind is "blend". Blend is a predicate... It means "to mix together" it could be written thus... blend: ["to mix" "blend" "stir together" etc etc.] If "blend" becomes too general, then additional, more industry specific predicates could be added. To Blend is a universal predicate and spans across any language in the universe. To "mix until well distributed" means the same thing no matter how you say it. We could call it "Snortlock" as long as we all agree to the meaning. Snortlock milk and chocolate. .... Carl wrote... Hi Scott: Thanks for posting the reference. The "semantic web" is why I created REBOL. In 1996 that was the vision, and it still is today. I agree with you... REBOL actually offers a better solution to this problem than XML. In REBOL the generation and the interpretation can be written within one portable language. So the sending side and the receiving side, as well as the message itself are all REBOL. The semantic problem is deeper than the XML guys realized. Tagging solves only half the battle. The association back to meaning is difficult, and the more you can give that association functional power, the better off you are. For instance, I can easily deal with the type and unit semantics: [ milk 2 cups chocolate 2 tsp. ] but when I do: [ mix until well distributed drink and enjoy ] it's more "function" than "data". REBOL handles this well. Don't get me wrong. I am not blasting XML. It is a heavyweight solution to the problem for folks who have the time and money to deal with it. But, for those of us on a tight budget and timeline, REBOL provides a more economical approach. Unfortunately, you don't want to enter this argument. It's better to conquer first, then explain why later. Thanks again, -Carl

 [2/10] from: gchiu:compkarori at: 14-Apr-2001 12:18


On Fri, 13 Apr 2001 09:37:42 -0500 "Scott" <[t-man--onemain--com]> wrote:
> "Why" not "what" was my question. I've actually read the > article in
<<quoted lines omitted: 3>>
> the question > was dodged by the authors.
I'm sure Carl could answer this better since he created Rebol to build his vision of the "Semantic Web". I agree with the authors that the Semantic Web is the killer application. The problem looks overwhelming when trying to solve it at a global level, but at a local level it might be more doable. Take for instance weather reports, I'm normally not interested in the weather in California. I'm just interested in local weather conditions for tomorrow in Wellington - probabilities of rain, snow, sunshine, and wind speed. I know several weather forecasters, and so I just have to twist their arms to provide the weather in a machine easily parsable format such as XML. At the moment, the weather report is presented as it would be read over the radio. There is only one agency that provides train transportation, and they publish their timetables on the web. That's a bit harder, but it's doable to parse their timetables. So, if I'm planning a trip, to say, up North for a picnic, all my agents have to do is: Book 2 return tickets after 11 am to Carterton if weather okay for tomorrow. Return before TV X-Files starts. I'll have one agent that grabs the weather, another to look at timetables, another to book seats, and another to look at my favourite TV programmes. -- Graham Chiu

 [3/10] from: gchiu:compkarori at: 13-Apr-2001 16:13


On Thu, 12 Apr 2001 19:41:09 -0700 "Terry Brownell" <[depotcity--telus--net]> wrote:
> Selective breeding of roses has never been able to > produce a blue-colored > rose. >
No biological advantage. Blue would be unable to attract the appropriate insect life required for pollination. Perhaps we should be looking to genetic programming to solve the semantic web problems and sort out the blue from the red roses. -- Graham Chiu

 [4/10] from: t-man:onemain at: 13-Apr-2001 1:49


howdy-- I have been following Berners-Lee's "semantic web" idea for at least a year now. Tim is a great read, and I wish he'd write more, but I think he's a little too dreamy about his semantic web. The fact is that people are much better at making sense of data than computers or networks or standards or ____ will be for a long while, and when that long while is up, it will likely be unnoticable to most. I can't begin to imagine the stupid amount of endless tagging and validation and resource descripting that will have to occur for the current scheme for all of that to yield a simple "Oh, yeah?" button. I'd still be deciding for myself anyway. The Semantic Web is a universal space for anything which can be expressed in classical logic. --TBL http://www.w3.org/DesignIssues/Rules.html I think I'd rather have a "What for?" button, myself. And most importantly a way to ask somebody I trust, RIGHT NOW, regardless of where they are or what they are doing. Email is sorta slick that way, newsgroups and lists are, too. But instant messaging from my current device to that person's device is really the ticket. Text is almost always more efficient and more considerate of peoples' time. My semantic web is already up and running, and I suspect that most peoples' are. It's simply the infrastructure that ties the knowledge and reasoning power of those I trust together. That is a force that the semantic web will not surpass anytime soon. REBOL is a fantastic agency for a real semantic web. Not one that tries to prove things to itself, but one that brings people together in a way that maximizes collaborative efforts and minimizes the ill effects of old business practices like the telephone and meetings and airline flights and commutes. The web has proven that what is truly valuable is attention, and attention takes time and anything that saves that goes straight to productivity. TBL, is in the clouds at 50,000 feet or something: Instead of asking machines to understand people's language, it involves asking people to make the extra effort. http://www.w3.org/DesignIssues/RDFnot.html I don't want to make the extra effort, personally. I don't think there's a payoff for the extra effort. Sure, it makes the web look prettier from the clouds, but I'd rather see a million bots with a basic concept of how to parse a sentence out looking for answers to my questions than a quadrillion pages with extra markup. Sure, if there were a standard that all documents adhered to that would aid these bots, then that would be cool. The old-fashioned standard of sentences and paragraphs and titles and sections and outlines, etc. I don't think any more writing between the lines is going to be solving anything. I keep hitting my "What for?" button, and nothing's coming back. REBOL is conquering first, because it won't have to explain why. Carl put it very politely, and now that I re-read his post at the bottom it seems I'm just reiterating what he has already said (terser, too, of course :-) T ----- Original Message ----- From: "Terry Brownell" <[depotcity--telus--net]> To: "Rebol List" <[rebol-list--rebol--com]> Sent: Thursday, April 12, 2001 2:08 AM Subject: [REBOL] Re: The Semantic Web - A solution?
> Personally, I don't see how XML and the Semantic web will gel? It
would require endless streams of coding to get my Rebol powered agent (http://www.LFReD.com ) to get all these disconnected ontologies and tags etc etc. to work together. I don't know, maybe I'm missing something, but the complexity of it all is daunting. The more complex it becomes, the less likely the majority would/could be bothered with it.
> Carl said "It's better to conquer first, then explain why later." > > I propose a "Rebol Semantic Standard"... a universal Rebol "ontology"
if you will, that at least Rebols can agree on.
> What do you think? > > Terry Brownell > Orion Alliance Inc. > > Some additional thoughts... > > The basic premise of the Semantic web is sound... subject, verb and
object.
> I tend to use subject and predicate for simplicity and scope. > > Lets look at Carls example... > > "mix until well distributed" > Now I don't know about you, but the word that comes to mind is
blend .
> Blend is a predicate... It means "to mix together" it could be written
thus...
> blend: ["to mix" "blend" "stir together" etc etc.] > > If "blend" becomes too general, then additional, more industry
specific predicates could be added.
> "To Blend" is a universal predicate and spans across any language in
the universe. To "mix until well distributed" means the same thing no matter how you say it. We could call it "Snortlock" as long as we all agree to the meaning.

 [5/10] from: t-man:onemain at: 13-Apr-2001 1:55


----- Original Message ----- From: "Graham Chiu" <[gchiu--compkarori--co--nz]> To: <[rebol-list--rebol--com]> Sent: Thursday, April 12, 2001 11:13 PM Subject: [REBOL] Re: The Semantic Web - A solution?
> On Thu, 12 Apr 2001 19:41:09 -0700 > "Terry Brownell" <[depotcity--telus--net]> wrote:
<<quoted lines omitted: 13>>
> [rebol-request--rebol--com] with "unsubscribe" in the > subject, without the quotes.
Pardon my naivete, but why is the semantic web a problem that ought be solved?

 [6/10] from: gchiu:compkarori at: 13-Apr-2001 21:32


On Fri, 13 Apr 2001 01:55:37 -0500 "Scott" <[t-man--onemain--com]> wrote:
> Pardon my naivete, but why is the semantic web a problem > that ought be > solved?
Implementation is the problem. -- Graham Chiu

 [7/10] from: t-man:onemain at: 13-Apr-2001 9:37


----- Original Message ----- From: "Graham Chiu" <[gchiu--compkarori--co--nz]> To: <[rebol-list--rebol--com]> Sent: Friday, April 13, 2001 4:32 AM Subject: [REBOL] Re: The Semantic Web - A solution?
> On Fri, 13 Apr 2001 01:55:37 -0500 > "Scott" <[t-man--onemain--com]> wrote:
<<quoted lines omitted: 6>>
> Graham Chiu > --
Why not "what" was my question. I've actually read the article in Scientific America now, and thought I would find the answer in the "What is the Killer App?" link, but after reading that, I think the question was dodged by the authors. T

 [8/10] from: alan_otterstad:mikronvinyl at: 12-Apr-2001 16:09


I have definitely GOT to vote for SNORTLOCK....hehehehe...good word Terry...excellent alan Personally, I don't see how XML and the Semantic web will gel? It would require endless streams of coding to get my Rebol powered agent ( http://www.LFReD.com ) to get all these disconnected ontologies and tags etc etc. to work together. I don't know, maybe I'm missing something, but the complexity of it all is daunting. The more complex it becomes, the less likely the majority would/could be bothered with it. Carl said "It's better to conquer first, then explain why later." I propose a "Rebol Semantic Standard"... a universal Rebol "ontology" if you will, that at least Rebols can agree on. What do you think? Terry Brownell Orion Alliance Inc. Some additional thoughts... The basic premise of the Semantic web is sound... subject, verb and object. I tend to use subject and predicate for simplicity and scope. Lets look at Carls example... mix until well distributed Now I don't know about you, but the word that comes to mind is "blend". Blend is a predicate... It means "to mix together" it could be written thus... blend: ["to mix" "blend" "stir together" etc etc.] If "blend" becomes too general, then additional, more industry specific predicates could be added. To Blend is a universal predicate and spans across any language in the universe. To "mix until well distributed" means the same thing no matter how you say it. We could call it "Snortlock" as long as we all agree to the meaning. Snortlock milk and chocolate. .... Carl wrote... Hi Scott: Thanks for posting the reference. The "semantic web" is why I created REBOL. In 1996 that was the vision, and it still is today. I agree with you... REBOL actually offers a better solution to this problem than XML. In REBOL the generation and the interpretation can be written within one portable language. So the sending side and the receiving side, as well as the message itself are all REBOL. The semantic problem is deeper than the XML guys realized. Tagging solves only half the battle. The association back to meaning is difficult, and the more you can give that association functional power, the better off you are. For instance, I can easily deal with the type and unit semantics: [ milk 2 cups chocolate 2 tsp. ] but when I do: [ mix until well distributed drink and enjoy ] it's more "function" than "data". REBOL handles this well. Don't get me wrong. I am not blasting XML. It is a heavyweight solution to the problem for folks who have the time and money to deal with it. But, for those of us on a tight budget and timeline, REBOL provides a more economical approach. Unfortunately, you don't want to enter this argument. It's better to conquer first, then explain why later. Thanks again, -Carl

 [9/10] from: gchiu:compkarori at: 13-Apr-2001 12:01


> Personally, I don't see how XML and the Semantic web will > gel? It would
<<quoted lines omitted: 7>>
> it becomes, the > less likely the majority would/could be bothered with it.
As complex life evolved from simple carbon chains, our current software agents will evolve with time to manage this complexity. -- Graham Chiu

 [10/10] from: depotcity:telus at: 12-Apr-2001 19:41


Selective breeding of roses has never been able to produce a blue-colored rose. Terry Brownell

Notes
  • Quoted lines have been omitted from some messages.
    View the message alone to see the lines that have been omitted