Perhaps the fact that it is difficult, that there is no simple solution, is what stops them from being able to fulfill that desire. Perhaps there is a lot of marketing to spread the idea that they don't. I wonder in whose interest that would be?
Now another thought. The problem with current social networks is that they are too small. On any centralised network the network operator is always listening. So the number of possible groups that could be made on a site with N users is the size of the Powerset of N. Which is a huge number. Just as a matter of interest here are some figures:
Now take a site with N users, remove the operator you have P(N)/2 number of groups that connot be made.
Next think of the possible groups that could be made if the whole of humanity could be linked together... The current social networks are just peanuts compared to what is possible....
The incestuous relationship is even weirder than one can imagine. Last year the Chief of the office of the Minister of Culture - the one ran by Christine Albanel no less, the Minister who is passing this hadopi law - was found dead in the appartment of the chief of TF1 international after having consumed a coctail of cocaine and ghb - known as a rape drug according to reports.
The Internet is a packet-switching network. As far as I can tell the iPhone has just as much connectivity as any home computer, it's not sandboxed into some crappy WAP corner nobody cares about.
No there are serious restrictions on how you can use your iPhone. No P2P is one such restriction. The other is you cannot use your iPhone as a modem to link to your computer. That is what decided me against it.
Well ok, HDTV pictures would fit on the 17" PowerBook's laptop screen, as they could fit on any laptop screen if resized correctly. The question I was raising, not very clearly though, is whether if the pictures of HTDV are 1920×1080 pixels, viewing them on a 1680x1050 pixel screen would entail some loss of quality? It would seems so doing simple maths: 220 pixels in the length and 30 pixels in the height.
But according to wikipedia both 1920x1080 and 1680x1050 are considered WUXGA. So are we missing out a lot on these new 17" PowerBook Pro's?
Sorry for the ignorant question. If WUXGA is 1920x1200 and the max supported resolution on the 17" is 1680x1050 then how come it is still WUXGA. Now I did do a little research on Wikipedia WUXGA right before, but that does not tell all the story. For example it says that North American HDTV uses 1920×1080 pictures, so it would seem that those would not quite fit on the 17" laptop, no?
Look I completely agree that one has to start with realistic things. RSS and atom are very simple, and good at what they do: really they just offer very simple file system metadata: see What Atom is all about.
Semantic Web services are a lot more complex. But at least they are RESTful. Now if a large percentage of the population finds its difficult to close xml tags, then they won't be using either atom or anything else. But that does not mean that there are not some very cool things to do in the mashup area.
Oracle is building Semantic Web technology into its database and open source mappers are appearing a little all over the place. See D2RQ as a good example, or Open Link Virtuoso. It's easy to create a mapping, I wrote mine and set up a server in one week.
Concerning ontologies versus tagging, there is no either/or here. It is simple to create a relationship for tagging. Here goes
Back to the main thread of this discussion: the HTML work.
To get back to Tim Berner's Lee's remarks: my thought here is the following:
I think I am getting what this is about: standardise the interpretation of tagsoup.
If every browser interpreted tagsoup identically then one could think of tagsoup as a form of xhmtl. Tagsoup pages would be displayed identically across browsers, and one could work with the resulting xhtml DOM tree.
One major advantage of producing your site in clean xhtml would then simply be that the rendering of a clean xhtml page would be a lot faster, as it would not have to go through the extra translation to xhtml. It would of course be easier to maintain too, as the structure of xhtml would be clearer than whatever weird tagsoup rules end up being decided as the standard ones.
SOAP is not! the semantic web! If you don't know that, then I suggest you turn the lights on before you eat stuff lying around you. You may have been in the toilets when you tried to eat that cake;-) . Don't want to think what you put in your mouth.
RDF is more RESTful than plain xml in many ways, since the terms in the language are URLS and so you can get their meaning. It is also very compatible with Atom. I am on the Atom Protocol list and my name is on the Atom Spec. So I know the debate well. I have a small AtomOwl ontology with XQuery transform, so you can transform andy atom document into XML. (ouch! xml. Damn people here find that too difficult! Are the people here your friends?)
Think of RDF as databases + URIs. Take a plain old vanilla mysql database, add a RDF mapping, and you can query it with sparql. I know, I have done this for the Roller Blog database. (you know, the database that publishes atom feeds)
Clearly the Semantic Web is not targeted at people who have difficulty writing xhtml.
It is targeted at groups who have valuable information to exchange in a very flexible manner. For those people, the Semantic Web is taking off. Think mashups. Think databases.
This may seem weird but should not be so surprising. Autism is
classified as a neurodevelopmental disorder that manifests itself in markedly abnormal social interaction, communication ability, patterns of interests, and patterns of behavior... ...autism manifests itself in delays in "social interaction, language as used in social communication, or symbolic or imaginative play".
Since autism is clearly related to language learning, we studied it in Philosophy, when I was at Birckbeck College. Children that are autistic have difficulty comprehending that others can see the world differently from the way they do. They will not understand for example that if a character in a muppet show hides something, the other characters in the show won't know that it is hidden.
I have just been reading Wittgenstein's Philosophical Investigations where through a series of questions he gets to the complexity of language learning, how much of a social process it is, how much it involves games - should in fact be seen as a set of overlapping games. When playing with a human being, there is always immediate feedback between a child and the people and objects around it, which involves smiles, cuddles and frowns, movements, hopping up and down, hiding, etc. The people on kids programs try the best to do that, but they can never directly respond to the child's immediate emotions, and they are in the end only ever a two dimensional picture on a box. So that the objects they move don't have a physical presence for the child. If those objects fall they can't hurt the child, if the people speak about an object, the child can't participate, if the person lies the child can't be deceived.
Children placed all day in front of a TV may not cry, but there is something fundamental that they will be missing.
Yes the real test of how useful this is would be to see which mice - the normal group or the happy group - fared better in a natural environment over a time span where natural selection could play a role. The happy mice may be happy for a while but not last very long. Sometimes it's not the right time to be happy: such as when (being a mouse) a cat is staring you right in the face.
In my November 2005 I posted BlogEd, ZFS and OSX, where I described an odd software defect that occured to me on my 17" laptop, and that dissapeared after I reinstalled a new hard drive, after my old one died on me. I am not sure what the defect was due to, but it is clear that with ZFS hard drive corruptions would be detected much earlier. This would be quite a serious advantage.
Ontologies have allready been hugely successful: it's called Object Oriented Programming. The two
are very similar, and you can read up about the similarities in detail on my blog [1] [2] [3]
The Semantic Web does not expect everyone to agree on an ontological framework, just as OO programming
does not require everyone to use everyone else's classes. When you write an little java ontology (also known as a class library), you put
your 'ontology' in a special name space which allows mixing and matching. To get your library widely adopted you need to
do a lot more than write it. You need:
to have a good ontology that others want to use
put it up in a namespace that people trust (say the W3C, the IETF, the OMG, or something like that)
work with others to make it known
solve a pressing issue
In Java one well known and respected channel for this to happen is the Java Community Process.
People can use other channels of course, and often do. But from time to time everyone agrees that
there is no need to keep re-inventing a framework and they then decide to go to some standards body
to agree on some common mapping, to help interoperability. The interoperability in Java was there all along of course.
But what was needed is a Convention to use a certain vocabulary.
Exactly the same will be true in the Semantic Web. The best ontologies will survive. Processes will be put in place to
help foster good ones.
Also I think you should not think that the only use of ontologies is to annotate web pages. The semantic web is here to
help you speak about resources in general, not just web resources. You may have more luck thinking about the Semantic
Web in terms of a way of doing what SQL currently does [4] (see SPARQL) but in a much much more scalable way.
In any case there is a lot of work going on in creating OWL models for Atom. The two are not at all incompatible - and with a little
more work they could have been nearly indistinguishable.
Another really good book that covers all the bases is Service Oriented Computing which gives a very good view as to how the Semantic Web, Agents, Web Services and RESTful apis fit together. This is a really serious book, but it helps get an understanding of the problems that are attempting to be solved.
This is very good news for open standards, as it is going to force adminstrators who want to reduce their work load to emphasize standards in the functionality they provide.
The INSEAD business school in France for example has outsourced their computer administration to some large company that ONLY supports Windows for example. This position is going to become less and less tenable as OSX users start growing and asking for their OS to be supported too. In order to reduce the work load of the support company and avoid having to deal with the subtleties of every OS they will, in the end be forced to work with standards: ie ldap, WEB-DAV, imap, correct html, etc... Once that is in the door for linux and Open Solaris users will be comfortably open.
Why will OSX users be better able to change support organisations? Because these users will probably have a lot more political clout, and certainly won't be the kind of users to be happy with some clever perl hack around a stupid administration. Also it will be difficult for support organisations to argue that OSX is more difficult to support, as all my experience points to the contrary.
slow garbage collection? the demo at Java One on the real time jvm that is used to control the airplane made by Boeing shows that even garbage collection can be tightly controlled.
Bloated? the windows international 5.0 jvm is a 15MB down only. For something that powerful this is really incredibly small. And it is much smaller than the.net equivalent I gather.
Not free? There are many implementations of the jvm from many vendors. BEA for examples has its own JVM that runs on multiple OSes. Their latest one will even run on Intel directly on the metal. IBM has its jvm. Many other do. And all of these are compatible. The only problem to date has been having a gnu jvm. There are a few, and they are a little behind: not surprising given the speed at which things have been moving in java, and the amount of investment it has received. But recently sun has backed one open source jvm at apache. Also I think Sun is open sourcing their own jvm...
I completely agree. The banks and all these instituations should be penalised for using such numbers as proof of identity.
Private institutions such as banks or the government should instead be giving out kryptokeys (also knowns as token cards) that give unique one time time-limited passwords to proove the person's identity.
That is what we had when I worked at AltaVista. At Sun they have exactly the same system, and I believe most security conscious institutions work that way. I would be really surprised if the technology is not far enough that one cannot now get such a display embedded on a credit card.
My suggestion to you: don't put too much weight on Tim Bray's bet. If you look carefully at his rdf.net challenge you will notice that the wording leaves him ample space to maneuvre were things to turn out agains him:
This has to happen before January 1, 2006, and
I am the sole judge and jury, but
Ill publicize anything thats submitted formally, and my comments on it,
so Im doing this in the open, except for
Im busy, so I may exercise fairly brutal triage on incoming proposals
and take a while to get to the ones really worth looking at, and
If theres serious money in it, the recipient of RDF.net is morally
obligated to find a way to cut me in for a piece of the action.
I like the last one: if someone has a idea that is going to make a lot of money they can have rdf.net for free if they cut him in on the action. wow! here is a man who really does not believe anything is going to happen:-)
I work with Tim Bray, but I seriously disagree with this position of his.
If you had gone back to the days before xml was invented
you could have made exactly the same argument against xml:
"SGML was not a success, therefore XML can't be". I have blogged about this falacious argument at length.
You can work with the Semantic Web without having to take on the most difficult problems of AI. You can use it to work on some really simple problems very effectively.
Speaking of "frauds", "ignorants" and "snake oil" when speaking of this project is really simplistic and (dare I turn the arrogance of the above poster against him?) stupid.
All those XML documents out on the web are allready machine understandable information. So are you also agains XML?
RDF adds to XML simply the ability to combine namespaces from different ontologies (= xml libraries) if you will in a well defined manner, which is a great way for everyone to save time.
Yes. This is the perfect slashdot thread to speak about re-inventing the wheel. It looks like there are at least 30 different Weblog Server out there, if not more. And some guy got pissed off because James Gosling wrote his own agregator. Weird.
Isn't that what the "free" version of Google's Blogger does already?
No it is not. That requires a thin client: namely a web browser, and it is connected to a thick server: Google (think about the number of machines they have, then you will understant what I mean by thick). In order to write a blog you have to be connected to their server. You have to write your blog into an html form, which is not very featurefull. You have to be on the internet to post it to their servers.
BlogEd on the other hand is a thick client. You can write and manage your blog without even being connected. You are not going through a third party to manage your blogs.
Each has its advantages. There is a lot more than can be done to improve BlogEd, and I don't want to deny that Blogger does have many cool features. I was just pointing out that this is a way to post your blogs that can easily be overlooked.
And on the Java side there is also BlogEd James Gosling's Client Side Blog Editor, which just ftps the html to the server. A very useful tool for thin servers.
All the listed Weblogs are server side.
That is missing out on a very useful category of Weblog editors: client side only editors. This is really useful for those of you who have a web server that does not have enough space to put up php or other server side magic: check out James Gosling's BlogEd. The nice thing about BlogEd is you can write and manage your blog whithout being connected to the web. It produces simple html which is the ftp-ed to the server at minimal cost.
There is still a lot of ways it can be improved. But the idea is certainly very original. And it is free: available under a BSD licence.
Slashdot Mirror
A study on Facebook done recently shows that people care a lot more than one thinks about privacy
http://preibusch.de/publications/Bonneau_Preibusch__Privacy_Jungle__2009-05-26.pdf
Perhaps the fact that it is difficult, that there is no simple solution, is what stops them from being
able to fulfill that desire. Perhaps there is a lot of marketing to spread the idea that they don't. I wonder in whose
interest that would be?
Now another thought. The problem with current social networks is that they are too small. On any centralised network the network operator is always listening. So the number of possible groups that could be made on a site with N users is the size of the Powerset of N. Which is a huge number. Just as a matter of interest here are some figures:
P(100) = 2^100 = 1267650600228229401496703205376
P(250) = 2^250 = 18092513943330655534932966407607485602073435104006338131165247501236\
42650624
P(1000) = 2^1000 = 10715086071862673209484250490600018105614048117055336074437503883703\
51051124936122493198378815695858127594672917553146825187145285692314\
04359845775746985748039345677748242309854210746050623711418779541821\
53046474983581941267398767559165543946077062914571196477686542167660\
429831652624386837205668069376
Now take a site with N users, remove the operator you have P(N)/2 number of groups that connot be made.
Next think of the possible groups that could be made if the whole of humanity could be linked together... The current social networks are just peanuts compared to what is possible....
The incestuous relationship is even weirder than one can imagine. Last year the Chief of the office of the Minister of Culture - the one ran by Christine Albanel no less, the Minister who is passing this hadopi law - was found dead in the appartment of the chief of TF1 international after having consumed a coctail of cocaine and ghb - known as a rape drug according to reports.
No P2P, that's resrtricted by the providers that Apple restricted itself to. You can't even use your iPhone as a modem to your laptop.
The Internet is a packet-switching network. As far as I can tell the iPhone has just as much connectivity as any home computer, it's not sandboxed into some crappy WAP corner nobody cares about.
No there are serious restrictions on how you can use your iPhone. No P2P is one such restriction. The other is you cannot use your iPhone as a modem to link to your computer. That is what decided me against it.
But according to wikipedia both 1920x1080 and 1680x1050 are considered WUXGA. So are we missing out a lot on these new 17" PowerBook Pro's?
What am I missing?
Look I completely agree that one has to start with realistic things. RSS and atom are very simple, and good at what they do: really they just offer very simple file system metadata: see What Atom is all about.
Semantic Web services are a lot more complex. But at least they are RESTful. Now if a large percentage of the population finds its difficult to close xml tags, then they won't be using either atom or anything else. But that does not mean that there are not some very cool things to do in the mashup area.
Oracle is building Semantic Web technology into its database and open source mappers are appearing a little all over the place. See D2RQ as a good example, or Open Link Virtuoso. It's easy to create a mapping, I wrote mine and set up a server in one week.
Concerning ontologies versus tagging, there is no either/or here. It is simple to create a relationship for tagging. Here goes
http://blogs.sun.com/bblfish :tag "semantic", "cool", "blog" .
I have tagged my blog with three tags.
Back to the main thread of this discussion: the HTML work.To get back to Tim Berner's Lee's remarks: my thought here is the following:
I think I am getting what this is about: standardise the interpretation of tagsoup.
If every browser interpreted tagsoup identically then one could think of tagsoup as a form of xhmtl. Tagsoup pages would be displayed identically across browsers, and one could work with the resulting xhtml DOM tree.
One major advantage of producing your site in clean xhtml would then simply be that the rendering of a clean xhtml page would be a lot faster, as it would not have to go through the extra translation to xhtml. It would of course be easier to maintain too, as the structure of xhtml would be clearer than whatever weird tagsoup rules end up being decided as the standard ones.
SOAP is not! the semantic web! If you don't know that, then I suggest you turn the lights on before you eat stuff lying around you. You may have been in the toilets when you tried to eat that cake ;-) . Don't want to think what you put in your mouth.
RDF is more RESTful than plain xml in many ways, since the terms in the language are URLS and so you can get their meaning. It is also very compatible with Atom. I am on the Atom Protocol list and my name is on the Atom Spec. So I know the debate well. I have a small AtomOwl ontology with XQuery transform, so you can transform andy atom document into XML. (ouch! xml. Damn people here find that too difficult! Are the people here your friends?)
Think of RDF as databases + URIs. Take a plain old vanilla mysql database, add a RDF mapping, and you can query it with sparql. I know, I have done this for the Roller Blog database. (you know, the database that publishes atom feeds)
Clearly the Semantic Web is not targeted at people who have difficulty writing xhtml.
It is targeted at groups who have valuable information to exchange in a very flexible manner. For those people, the Semantic Web is taking off. Think mashups. Think databases.
This may seem weird but should not be so surprising. Autism is
Since autism is clearly related to language learning, we studied it in Philosophy, when I was at Birckbeck College. Children that are autistic have difficulty comprehending that others can see the world differently from the way they do. They will not understand for example that if a character in a muppet show hides something, the other characters in the show won't know that it is hidden.I have just been reading Wittgenstein's Philosophical Investigations where through a series of questions he gets to the complexity of language learning, how much of a social process it is, how much it involves games - should in fact be seen as a set of overlapping games. When playing with a human being, there is always immediate feedback between a child and the people and objects around it, which involves smiles, cuddles and frowns, movements, hopping up and down, hiding, etc. The people on kids programs try the best to do that, but they can never directly respond to the child's immediate emotions, and they are in the end only ever a two dimensional picture on a box. So that the objects they move don't have a physical presence for the child. If those objects fall they can't hurt the child, if the people speak about an object, the child can't participate, if the person lies the child can't be deceived.
Children placed all day in front of a TV may not cry, but there is something fundamental that they will be missing.
Yes the real test of how useful this is would be to see which mice - the normal group or the happy group - fared better in a natural environment over a time span where natural selection could play a role. The happy mice may be happy for a while but not last very long. Sometimes it's not the right time to be happy: such as when (being a mouse) a cat is staring you right in the face.
In my November 2005 I posted BlogEd, ZFS and OSX, where I described an odd software defect that occured to me on my 17" laptop, and that dissapeared after I reinstalled a new hard drive, after my old one died on me. I am not sure what the defect was due to, but it is clear that with ZFS hard drive corruptions would be detected much earlier. This would be quite a serious advantage.
The Semantic Web does not expect everyone to agree on an ontological framework, just as OO programming does not require everyone to use everyone else's classes. When you write an little java ontology (also known as a class library), you put your 'ontology' in a special name space which allows mixing and matching. To get your library widely adopted you need to do a lot more than write it. You need:
In Java one well known and respected channel for this to happen is the Java Community Process. People can use other channels of course, and often do. But from time to time everyone agrees that there is no need to keep re-inventing a framework and they then decide to go to some standards body to agree on some common mapping, to help interoperability. The interoperability in Java was there all along of course. But what was needed is a Convention to use a certain vocabulary.
Exactly the same will be true in the Semantic Web. The best ontologies will survive. Processes will be put in place to help foster good ones.
Also I think you should not think that the only use of ontologies is to annotate web pages. The semantic web is here to help you speak about resources in general, not just web resources. You may have more luck thinking about the Semantic Web in terms of a way of doing what SQL currently does [4] (see SPARQL) but in a much much more scalable way.
Henry Story
[1] Java Annotations and the Semantic Web
[2] UML, MOF, MDA, OWL: how they all fit together
[3] Would a little DOAP help?
[4] SPARQL to ignite web 2.0
In any case there is a lot of work going on in creating OWL models for Atom. The two are not at all incompatible - and with a little more work they could have been nearly indistinguishable.
Another really good book that covers all the bases is Service Oriented Computing which gives a very good view as to how the Semantic Web, Agents, Web Services and RESTful apis fit together. This is a really serious book, but it helps get an understanding of the problems that are attempting to be solved.
The INSEAD business school in France for example has outsourced their computer administration to some large company that ONLY supports Windows for example. This position is going to become less and less tenable as OSX users start growing and asking for their OS to be supported too. In order to reduce the work load of the support company and avoid having to deal with the subtleties of every OS they will, in the end be forced to work with standards: ie ldap, WEB-DAV, imap, correct html, etc... Once that is in the door for linux and Open Solaris users will be comfortably open.
Why will OSX users be better able to change support organisations? Because these users will probably have a lot more political clout, and certainly won't be the kind of users to be happy with some clever perl hack around a stupid administration. Also it will be difficult for support organisations to argue that OSX is more difficult to support, as all my experience points to the contrary.
slow garbage collection? the demo at Java One on the real time jvm that is used to control the airplane made by Boeing shows that even garbage collection can be tightly controlled.
.net equivalent I gather.
Bloated? the windows international 5.0 jvm is a 15MB down only. For something that powerful this is really incredibly small. And it is much smaller than the
Not free? There are many implementations of the jvm from many vendors. BEA for examples has its own JVM that runs on multiple OSes. Their latest one will even run on Intel directly on the metal. IBM has its jvm. Many other do. And all of these are compatible. The only problem to date has been having a gnu jvm. There are a few, and they are a little behind: not surprising given the speed at which things have been moving in java, and the amount of investment it has received. But recently sun has backed one open source jvm at apache. Also I think Sun is open sourcing their own jvm...
These laptops look interesting. But perhaps they are just the first moves in what may be a Niagara laptop? Now that would certainly have my interest!
I completely agree. The banks and all these instituations should be penalised for using such numbers as proof of identity.
Private institutions such as banks or the government should instead be giving out kryptokeys (also knowns as token cards) that give unique one time time-limited
passwords to proove the person's identity.
That is what we had when I worked at AltaVista. At Sun they have exactly the same system, and I believe most security conscious institutions work that way. I would be really surprised if the technology is not far enough that one cannot now get such a display embedded on a credit card.
My suggestion to you: don't put too much weight on Tim Bray's bet. If you look carefully at his rdf.net challenge you will notice that the wording leaves him ample space to maneuvre were things to turn out agains him:
- This has to happen before January 1, 2006, and
- I am the sole judge and jury, but
- Ill publicize anything thats submitted formally, and my comments on it,
so Im doing this in the open, except for
- Im busy, so I may exercise fairly brutal triage on incoming proposals
and take a while to get to the ones really worth looking at, and
- If theres serious money in it, the recipient of RDF.net is morally
obligated to find a way to cut me in for a piece of the action.
I like the last one: if someone has a idea that is going to make a lot of money they can have rdf.net for free if they cut him in on the action. wow! here is a man who really does not believe anything is going to happenI work with Tim Bray, but I seriously disagree with this position of his. If you had gone back to the days before xml was invented you could have made exactly the same argument against xml: "SGML was not a success, therefore XML can't be". I have blogged about this falacious argument at length. You can work with the Semantic Web without having to take on the most difficult problems of AI. You can use it to work on some really simple problems very effectively. Speaking of "frauds", "ignorants" and "snake oil" when speaking of this project is really simplistic and (dare I turn the arrogance of the above poster against him?) stupid.
I am not a numerics person but I found this book:
A Numerical Library in Java for Scientists and Engineers
by Hang T. Lau
Champal & Hall/CRC
ISBN-1-58488-430-4
It looks pretty interesting.
All those XML documents out on the web are allready machine understandable information. So are you also agains XML? RDF adds to XML simply the ability to combine namespaces from different ontologies (= xml libraries) if you will in a well defined manner, which is a great way for everyone to save time.
BlogEd on the other hand is a thick client. You can write and manage your blog without even being connected. You are not going through a third party to manage your blogs. Each has its advantages. There is a lot more than can be done to improve BlogEd, and I don't want to deny that Blogger does have many cool features. I was just pointing out that this is a way to post your blogs that can easily be overlooked.
And on the Java side there is also BlogEd James Gosling's Client Side Blog Editor, which just ftps the html to the server. A very useful tool for thin servers.
All the listed Weblogs are server side. That is missing out on a very useful category of Weblog editors: client side only editors. This is really useful for those of you who have a web server that does not have enough space to put up php or other server side magic: check out James Gosling's BlogEd. The nice thing about BlogEd is you can write and manage your blog whithout being connected to the web. It produces simple html which is the ftp-ed to the server at minimal cost. There is still a lot of ways it can be improved. But the idea is certainly very original. And it is free: available under a BSD licence.