"The Cyc debacle illustrates how much work you have to put into tagging to get very little out. After twenty years of that money sink, it's still useless."
And after five years (and yes a lot of cash), the Gene Ontology is an incredibly useful tool for biologists.
It's not the answer to everything, but it makes some things easier. This is enough.
a) telomere shortening may be involved, or related to some forms of cancer. I haven't read this article well enough to determine how much evidence there is for a causative link
b) telomere lengthening in nematodes is correlated with increased life span. Again, not causative. More over nematodes have a sterotyped development, and adult cells do not divide (other than germ cells, obviously).
Telomeres are an attractive hypothesis, but its not been shown yet. Correlation does not prove causation. And what is true of nematodes need not be true of other organisms.
In short, it is certainly true that telomeric shortening does not explain aging, and it has not been shown that they are definately involved in it.
"So how does OWL/RDF make 'sense' out of a tag that I just made up?"
OWL can be used as a schema language. It can be used to specify what a concept means and how it relates to (some) other concepts. From this you can apply automated techniques to work out how it relates to other concepts.
Now, clearly, we have to agree on somethings. We both have to agree to use the same metalanguage, like OWL, and we have to agree on some of the vocabulary in the language that we are making up. But, with a relational database, or indeed XML, you have to agree on all of it.
Its not a magic bullet, but its a small step forward.
Good grief, and you accuse the semantic web people of having their head in the clouds. Performance is a key issue in the success or failure of any system.
"Namely, the SQL DBMS (if that's what they're using) that the GO DB is using are inadequately implementing hierarchies "
SQL is based around a relational data model. The relational data model is not designed to store hierarchives, or graphs or any sort. You can not make graph queries against the data, and it wont work efficiently.
"Not sure what you're asking. Does XML/RDF/etc. provide for referential integrity in this fashion?"
No, RDF does not maintain referential integrity. Its designed to operate without it. This is considered to be important because the web is de-centralised. Unlike an RDBMS. If you throw away referential integrity bang goes one of the key advantages of RDBMS.
"Isn't that the point of RDF and OWL etc.? That everyone has to use the same XML tags (or XSD) in order for this whole thing to work?"
No. This is exactly not the point. If OWL, and RDF, required this then you could just use XML. OWL, in particular, is a mechanism for expressing a schema with a precise semantics. So you can define new "tags", or classes, and still calculate the relationship between these classes.
More over the semantics of OWL make an "open world assumption". They do not assume that everything that is true is known to the system, while a RDBMS is closed world. Again this does not sit well with a decentralised world.
"If the message is not agreed upon beforehand then you are correct in saying that there can be no automatic communication. Of course, XML does not make this problem go away."
No, XML has nothing to say about this. By OWL offers at least some facilties for making this sort of "semantic reconciliation" work.
"The semantic web is a way to interrelate data from disparate systems, no?"
Yes.
"As I said before -- if people didn't create 'semi-structured' data in the first place we would already have the semantic web."
If everyone agreed on a single relational schema to store all of their data, then, you are correct, we would not have this problem. But people don't. Get used to it.
You can certainly do this. Look at the Gene Ontology database. The problem is that relational model performs very badly with this sort of data. The GO DB stores transitive closure information, precached, in the RDBMS to make it work reasonably quickly. Which obvious breaks normalisation.
Besides which you often want to describe things in richer manner than just hierarchies. The GO database uses on additional relationship, namely "part of". It turn out that if you increase the expressivity of your relationships, asking questions of the data quickly becomes computationally intractable (if not impossible). Again RDBMS offers no support for this.
"The real solution is a system of distributed RDBMS. "
How will you build a distributed system which can still maintain referential integrity like a RDBMS.
Besides which you are completely wrong to suggest that you need no additional metadata. This would work but only if people were all using the same schema. If not, then you are left with the problem of relating the data in one relational schema with that in another. Schema reconciliation is either done by hand, or with metadata, if you want to do it automatically (which is hard!).
RDBMS are fine, but they are not a universal answer. If you think that the semantic web is trying to do something that an RDBMS is doing them I suspect you misunderstand the semantic web, or the problem. Or possibly RDBMS.
The semantic web is not about producing new web pages that people can read. Its about producing knowledge in a form that computers can understand. If you want to stick this information in a browser, current technology works okay.
Writing illegal in capitals does not make it special. Its up to society to decide what is legal and what is not. The fact that something is illegal does not make it in and off itself wrong. It just makes it illegal. And the fact that something is legal does not make it right. Cause and effect are backwards.
There is a problem with purely IP based income. If draconian laws are the only way that IP can work, then we need to rethink IP. It is already happening in some areas, particulary in scientific publishing for instance.
BTW I don't see what is wrong with the notion that software is a commodity. Most of it is, and is likely to become more so.
Perhaps he is irritated at the large amount of time and effort going to enforcing increasingly draconian efforts to support the failing business model of the movie industry.
Like the oil industry which requires us to make continual military interventions into many parts of the world to survive, the movie industry is asking society to spend enourmous amounts of cash and suffer attacks on its freedom so it can make money. This is the industry that told us that video recording was going to kill it. Actually it said that TV was going to kill it before that.
They should get a life, learn to compete with the new realities, and stop belly aching. Why they can not realise that there is money to be made out of enabling people to watch a film in comfort, with good sound and visuals I don't know. A wobbly camcorder image with the sound of hardening arteries from the popcorn eating kid in the next seat is not competition.
The industrial revolution was a total disaster for large swathes of the population. It was only with the advent of movements like the luddites and the chartist and finally the union movements which actually enabled the majority of the population to enjoy the benefits that increased production bought.
The notion that increasing wealth for one automatically results in a better society for all was wrong then and its wrong now. The problem in this case is that the increasing mobility of capital has not been associated with an increasing mobility of people. Money can move for cheaper wages but people can not move for higher ones.
"The case for species preservation should be made on hard ground, not on computer-generated squish."
Computers can also be trained to produce sensible results. The weather forecasts are notably better now than they ever where before.
It is trivially obvious that we can not accurately predict how many species will survive global warming. It is trivially obvious that we can criticise any studies attempting to do so.
But this work is not the result of hysterical conclusions, it's an attempt to gather data in difficult circumstances.
Should we listen to this sort of data? Well it is also trivially obvious that once species have died out it is too late to recover them. Our current world leaders, most of whom will be dead in the next 20 or 30 years could have little to gain by listening to such reports. But perhaps the rest of us should.
Making a copy for playing from alternative media is not a "incidental" or "transient" copy. The point is, if you use FTP for example to transfer some copyrighted article to me, a short term copy of it, or parts of it, will exist on every machine in between you and me. Hence it would be illegal.
This exemption covers this. Likewise if you are reading an electronic book, having your OS caching it in virtual memory would not be infringing it.
Storeing stuff long term on a hard drive would not be incidental or transitory.
This is why the US needs a strong data protection act. In Europe there are strong laws to prevent release of personal information without the direct agreement of the person. And to make this law at all useful it would be illegal for a company to release that information, or transfer it to another country which does not have similar strong laws which are enforcible. So this situation would never have happened.
Indeed, this caused all sorts of hassles with transatlantic companies. They could not transfer data to the US because it didn't have an equivalent law. In the end the "Safe Harbour" agreement came up, which means that personal data about me, gather in Europe, but exported to the US has stronger data protection, than personal data gathered about US citizens and kept in the US.
You really are missing the point. Most currencies are different sizes and different colours. Most people around the world use this to tell the demoninations quickly. As a result, US money is a total pain in the ass.
Besides which numbers are fine, but quite a large section of the population is blind or partially sighted. US money is inpenetrable to them.
And ask yourself this? If numbers are so easy...then why the hell are your coins all different sized??
"So, tell me, if I'm a counterfitter, why wouldn't I just copy the older bills and 'age' them in the washing machine?"
As they get older, they will get less common. As they get less common, people will look at them more. As the people look at them more, they will get harder to pass. As they get harder to pass, people will not want to counterfeit them.
The problem is that its very hard to know if you are "using" stuff or not. The licenses are not simple, and as he says, you more or less have to wipe the hard drive when systems get passed around. Thats a big cost...you are having to re-install a machine, which already works fine, just in case its got some software on which the recipient is not going to use anyway.
Having to comply with complex licenses reduces the usefulness of the product. Imagine if a car manufacturer licensed all the bits of a vehicle separately....
"Why load this big program at every application launch?"
Much of the loading time is loading classes not the JVM per se. These are loaded on demand, rather than for every application.
"Safety? Given Java's lack of pointer types, it shouldn't need a JVM monitoring it to be safe. Again, there are lots of safe languages (Haskell, Clean, and the aforementioned dynamic languages) that don't require a runtime monitor."
Java's type system is not as rich as the other type systems that you mentioned. It needs some checking at runtime (ClassCast, and ArrayBounds).
The second point is that you are assuming that you trust the compiler. The original idea of Java was mobile code, and all that jazz. If you do not trust the compiler then you have to do more checks at runtime, even though they could and have been checked at compile time.
Java is a reasonable compromise. It was also there at a good place, and good time. Some of the original motivations for Java (mobile code) seem irrelevant now. This is way things work.
Slashdot Mirror
"The Cyc debacle illustrates how much work you have to put into tagging to get very little out. After twenty years of that money sink, it's still useless."
And after five years (and yes a lot of cash), the Gene Ontology is an incredibly useful tool for biologists.
It's not the answer to everything, but it makes some things easier. This is enough.
Phil
These reports suggest that
a) telomere shortening may be involved, or related to some forms of cancer. I haven't read this article well enough to determine how much evidence there is for a causative link
b) telomere lengthening in nematodes is correlated with increased life span. Again, not causative. More over nematodes have a sterotyped development, and adult cells do not divide (other than germ cells, obviously).
Telomeres are an attractive hypothesis, but its not been shown yet. Correlation does not prove causation. And what is true of nematodes need not be true of other organisms.
In short, it is certainly true that telomeric shortening does not explain aging, and it has not been shown that they are definately involved in it.
Phil
It's far from shown that telomere shortening is the direct cause of even cellular sensenence, yet alone old age.
Mice have telomeres about 10 times the length of humans, but live (and are only capable of life) which is 10 fold shorter than humans.
Phil
"So how does OWL/RDF make 'sense' out of a tag that I just made up?"
OWL can be used as a schema language. It can be used to specify what a concept means and how it relates to (some) other concepts. From this you can apply automated techniques to work out how it relates to other concepts.
Now, clearly, we have to agree on somethings. We both have to agree to use the same metalanguage, like OWL, and we have to agree on some of the vocabulary in the language that we are making up. But, with a relational database, or indeed XML, you have to agree on all of it.
Its not a magic bullet, but its a small step forward.
Phil
"Performance is an implementation detail."
Good grief, and you accuse the semantic web people of having their head in the clouds. Performance is a key issue in the success or failure of any system.
"Namely, the SQL DBMS (if that's what they're using) that the GO DB is using are inadequately implementing hierarchies "
SQL is based around a relational data model. The relational data model is not designed to store hierarchives, or graphs or any sort. You can not make graph queries against the data, and it wont work efficiently.
RDBMS are good for some things, but not this.
Phil
"Not sure what you're asking. Does XML/RDF/etc. provide for referential integrity in this fashion?"
No, RDF does not maintain referential integrity. Its designed to operate without it. This is considered to be important because the web is de-centralised. Unlike an RDBMS. If you throw away referential integrity bang goes one of the key advantages of RDBMS.
"Isn't that the point of RDF and OWL etc.? That everyone has to use the same XML tags (or XSD) in order for this whole thing to work?"
No. This is exactly not the point. If OWL, and RDF, required this then you could just use XML. OWL, in particular, is a mechanism for expressing a schema with a precise semantics. So you can define new "tags", or classes, and still calculate the relationship between these classes.
More over the semantics of OWL make an "open world assumption". They do not assume that everything that is true is known to the system, while a RDBMS is closed world. Again this does not sit well with a decentralised world.
"If the message is not agreed upon beforehand then you are correct in saying that there can be no automatic communication. Of course, XML does not make this problem go away."
No, XML has nothing to say about this. By OWL offers at least some facilties for making this sort of "semantic reconciliation" work.
"The semantic web is a way to interrelate data from disparate systems, no?"
Yes.
"As I said before -- if people didn't create 'semi-structured' data in the first place we would already have the semantic web."
If everyone agreed on a single relational schema to store all of their data, then, you are correct, we would not have this problem. But people don't. Get used to it.
Phil
You can certainly do this. Look at the Gene Ontology database. The problem is that relational model performs very badly with this sort of data. The GO DB stores transitive closure information, precached, in the RDBMS to make it work reasonably quickly. Which obvious breaks normalisation.
Besides which you often want to describe things in richer manner than just hierarchies. The GO database uses on additional relationship, namely "part of". It turn out that if you increase the expressivity of your relationships, asking questions of the data quickly becomes computationally intractable (if not impossible). Again RDBMS offers no support for this.
Phil
"The real solution is a system of distributed RDBMS. "
How will you build a distributed system which can still maintain referential integrity like a RDBMS.
Besides which you are completely wrong to suggest that you need no additional metadata. This would work but only if people were all using the same schema. If not, then you are left with the problem of relating the data in one relational schema with that in another. Schema reconciliation is either done by hand, or with metadata, if you want to do it automatically (which is hard!).
RDBMS are fine, but they are not a universal answer. If you think that the semantic web is trying to do something that an RDBMS is doing them I suspect you misunderstand the semantic web, or the problem. Or possibly RDBMS.
Phil
The semantic web is not about producing new web pages that people can read. Its about producing knowledge in a form that computers can understand. If you want to stick this information in a browser, current technology works okay.
Phil
Writing illegal in capitals does not make it special. Its up to society to decide what is legal and what is not. The fact that something is illegal does not make it in and off itself wrong. It just makes it illegal. And the fact that something is legal does not make it right. Cause and effect are backwards.
There is a problem with purely IP based income. If draconian laws are the only way that IP can work, then we need to rethink IP. It is already happening in some areas, particulary in scientific publishing for instance.
BTW I don't see what is wrong with the notion that software is a commodity. Most of it is, and is likely to become more so.
Phil
Perhaps he is irritated at the large amount of time and effort going to enforcing increasingly draconian efforts to support the failing business model of the movie industry.
Like the oil industry which requires us to make continual military interventions into many parts of the world to survive, the movie industry is asking society to spend enourmous amounts of cash and suffer attacks on its freedom so it can make money. This is the industry that told us that video recording was going to kill it. Actually it said that TV was going to kill it before that.
They should get a life, learn to compete with the new realities, and stop belly aching. Why they can not realise that there is money to be made out of enabling people to watch a film in comfort, with good sound and visuals I don't know. A wobbly camcorder image with the sound of hardening arteries from the popcorn eating kid in the next seat is not competition.
Phil
As sourceforge is not, to my knowledge, open source
I am unclear how it is representative of open source.
It's not free software either.
This is one of the reasons that Tim Perdue created gforge incidentally.
Phil
The industrial revolution was a total disaster for large swathes of the population. It was only with the advent of movements like the luddites and the chartist and finally the union movements which actually enabled the majority of the population to enjoy the benefits that increased production bought.
The notion that increasing wealth for one automatically results in a better society for all was wrong then and its wrong now. The problem in this case is that the increasing mobility of capital has not been associated with an increasing mobility of people. Money can move for cheaper wages but people can not move for higher ones.
Phil
"The case for species preservation should be made on hard ground, not on computer-generated squish."
Computers can also be trained to produce sensible results. The weather forecasts are notably better now than they ever where before.
It is trivially obvious that we can not accurately predict how many species will survive global warming. It is trivially obvious that we can criticise any studies attempting to do so.
But this work is not the result of hysterical conclusions, it's an attempt to gather data in difficult circumstances.
Should we listen to this sort of data? Well it is
also trivially obvious that once species have died out it is too late to recover them. Our current world leaders, most of whom will be dead in the next 20 or 30 years could have little to gain by listening to such reports. But perhaps the rest of us should.
Phil
"However, I understand that that book was one of Atwood's most popular works probably because it was a favorite among feminists. "
Or because its a rich novel, with a compelling plot, and the wonderful use of language that you would expect from her.
Also it was made into a film in which Natasha Richardson gets her kit off.
Phil
Making a copy for playing from alternative media
is not a "incidental" or "transient" copy. The point is, if you use FTP for example to transfer some copyrighted article to me, a short term copy of it, or parts of it, will exist on every machine in between you and me. Hence it would be illegal.
This exemption covers this. Likewise if you are reading an electronic book, having your OS caching it in virtual memory would not be infringing it.
Storeing stuff long term on a hard drive would not
be incidental or transitory.
It really is a daft law.
Phil
John Peel is only on holiday. Besides I think
David Stafford has his moments.
This is why the US needs a strong data protection act. In Europe there are strong laws to prevent release of personal information without the direct agreement of the person. And to make this law at all useful it would be illegal for a company to release that information, or transfer it to another country which does not have similar strong laws which are enforcible. So this situation would never have happened.
Indeed, this caused all sorts of hassles with transatlantic companies. They could not transfer data to the US because it didn't have an equivalent law. In the end the "Safe Harbour" agreement came up, which means that personal data about me, gather in Europe, but exported to the US
has stronger data protection, than personal data gathered about US citizens and kept in the US.
It's a strange world.
Phil
You really are missing the point. Most currencies are different sizes and different colours. Most people around the world use this to tell the demoninations quickly. As a result, US money is a total pain in the ass.
Besides which numbers are fine, but quite a large section of the population is blind or partially sighted. US money is inpenetrable to them.
And ask yourself this? If numbers are so easy...then why the hell are your coins all different sized??
Phil
"So, tell me, if I'm a counterfitter, why wouldn't I just copy the older bills and 'age' them in the washing machine?"
As they get older, they will get less common. As they get less common, people will look at them more. As the people look at them more, they will get harder to pass. As they get harder to pass, people will not want to counterfeit them.
Phil
The problem is that its very hard to know if
you are "using" stuff or not. The licenses are not simple, and as he says, you more or less have to wipe the hard drive when systems get passed around. Thats a big cost...you are having to re-install a machine, which already works fine, just in case its got some software on which the recipient is not going to use anyway.
Having to comply with complex licenses reduces the usefulness of the product. Imagine if a car manufacturer licensed all the bits of a vehicle separately....
Phil
Has anyone got any experience of the accessibility options. Is the screenreader usuable?
Phil
"Personally I am a response individual, and a left wing anti capitalist...."
Although I sometimes have problems with typing....
"responsible"
"responsible organisations and individuals or a bunch of left wing anti capitalists."
Personally I am a response individual, and a left wing anti capitalist....
Phil
"Why load this big program at every application launch?"
Much of the loading time is loading classes not the JVM per se. These are loaded on demand, rather than for every application.
"Safety? Given Java's lack of pointer types, it shouldn't need a JVM monitoring it to be safe. Again, there are lots of safe languages (Haskell, Clean, and the aforementioned dynamic languages) that don't require a runtime monitor."
Java's type system is not as rich as the other type systems that you mentioned. It needs some checking at runtime (ClassCast, and ArrayBounds).
The second point is that you are assuming that you trust the compiler. The original idea of Java was mobile code, and all that jazz. If you do not trust the compiler then you have to do more checks at runtime, even though they could and have been checked at compile time.
Java is a reasonable compromise. It was also there at a good place, and good time. Some of the original motivations for Java (mobile code) seem irrelevant now. This is way things work.
Phil