Al Fasoldt is right in part. We do need to critically analyze everything we read. He is, however, clearly fear-mongering when he singles out the Web, and Wikipedia in particular as especially untrustworthy.
As an american (I can't speak for anyone else), I am continually exposed to false information in TV and radio programs, books, magazines, casual conversation, and, yes, Al, on the Internet, and in the Wikipedia.
The ratio of fact to fiction in the Wikipedia will be higher than some sources of information, and lower than others. Likely, Al's readers are already familiar with the process of selecting one news source over another, as they chose his newspaper over his competitor's this morning, they must choose some TV and radio programs over others, etc.
The real discussion to be had here is on how, specifically, the Wikipedia process compares to the processes used to create and disseminate information by other sources. I'm not sure I can add much to that discussion, but I will say this -- When I see an error on the Wikipedia, I can amend it with a few mouse clicks and some typing. When I see an error in a textbook or in a newspaper or on TV, however, I can only send a letter to the editor, and hope that he sees fit to fix/acknowledge the error.
Without knowing your taste (beyond the selection you presented), and knowing that you have a small jazz collection at this point, I can reccommend to you the jazz canon (as I see it), as well as my personal favorites (of course).
The Canon -- The artists/albums a jazz collector should be familiar with regardless of specialty (dixieland through free jazz):
The Blues (Robert Johnson/Bessie Smith/etc.) -- Jazz starts with the Blues -- I've learned a lot about Jazz by listening to pre-Jazz Blues artists. Louis "Pops" Armstrong - material from the early days where he plays trumpet at least as much as he sings. Django Rheinhardt Fats Waller Billie Holiday Ella Fitzgerald Count Basie (w/ Lester Young) Duke Ellington Benny Goodman Septet (w/ Charlie Christian) Lester Young Coleman Hawkins Oscar Peterson Charlie Parker, anything John Burkes "Dizzy" Gilespie, "Sunny Side Up", anything he did with Charlie Parker Thelonious Monk (w/ Charlie Rouse) King Pleasure, "Moody's Mood For Love" Sonny Rollins, "Freedom Suite" Miles Davis, "Birth of the Cool", "Kind of Blue" Dave Brubeck (w/ Paul Desmond), "Time Out" Stan Getz (w/ Gilberto) Charlie Mingus Cannonball Adderley Wes Montgomery John Coltrane, "Giant Steps" Wayne Shorter, "Speak No Evil" Ornette Coleman, "Shape of Jazz to Come" Charlie Haden, "Liberation Music Orchestra"
My personal picks:
Tal Farlow Bobby Hutcherson Chic Corea, Origins band (w/ Avishai Coehn and Steve Wilson) Keith Jarrett Kenny Garrett, "Triology" Wynton, Ellis, and Brandord Marsalis James Moody Charlie Mingus, "The Clown" Oliver Nelson, "Blues and the Abstract Truth", "More Blues and the abstract Truth" Benny Carter Phil Woods Yusef Lateef, "Eastern Sounds" Charlie Parker, "Mango Mangue" Tony Bennett John McLoughlin (w/ Remember Shakti) John Scofield (w/ Joe Levano if you can find it)
These lists, are, needless to say, very incomplete.
Remember when buying albums that until a certain point (in the 40's?), most records were more like today's singles, containing only a couple of tracks. As a result, compilations are probably your best bet for jazz until after Charlie Parker. There are, of course, many exceptions to this...
As you collect, you might also notice a gap between Swing and Bebop, where there is this radical shift in style. In reality, there was an extensive recording ban (strike), that coincided with the inception of bebop. Recordings from this period are highly valued by collectors.
The GPL is an attempt to modify or extend traditional copyright. It is a set of limitations that make publications more free. It is copyleft. The DMCA has several provisions that make publications less free. I won't bother enumerating them here.
GPL != GNU. GPL is a tool that the GNU community uses to enable the production and protection of free publications. GPL relies on copyright, but the notion of freedom does not. This notion predates Gutenberg significantly;)
Guys, please don't ignore the "GNU" in GNU-Darwin.
The DMCA and laws like it threaten the existence of GNU. GNU-Darwin is part of GNU. By encouraging GNU-Darwin, Apple helps GNU. By choosing to enforce the DMCA, Apple threatens GNU. This contradictory stance must be evaluated by determining which is stronger, the support, or the threat.
Apple is using its lawyers to support the DMCA. Apple is not using its lawyers to support GNU (see the GNU commentary on the APSL).
Of course, only the GNU folks can determine what is best for them. it appears that they have done so, and that they made the determination that (perhaps) loosing one project is better than loosing the entire organization (or its mission / goals).
I can't see anything contradictory or stupid about this decision. The only contradictory thing I can see here is Apple's behavior.
As a musician and a software developer, I know intuitively what OSS and the internet community in general has to offer musicians. I recently downloaded Audacity, for example, and I was able to use it out-of-the-box with no problems -- something I can't quite say for some of its proprietary counterparts.
But as a software developer, I'm missing the community websites, mailing lists, and general online discussion that I enjoy in other domains. Where do music software developers "hang out" online? (I'm specifically referring to those people developing software tools for musicians, as opposed to those developing file-trading and music-player applications.)
There are plenty of websites about music out there, but 99% appear to be fansites or e-business sites for music industry companies.
From time to time I toss around the idea of putting together a mailing list / website for music technology discussions (maybe slashcode is good for this?). If I did, would anyone use it? Or does something comparable already exist, and I am just missing it?
BTW, the musical community needs OSS badly. Leaving technology standards up to the music industry has left us with crappy, slowly evolving protocols with little innovation. With all due respect, MIDI was great in its day, but it really doesn't meet the tests of time, IMHO. Also, I know that musicXML is on the way, but I'm still skeptical:)...
If you want "usability", connect the user to the developer in a meaningful way, by adding the following features to your project.
1. Feedback Button 2. Zero Defect Tolerance
1. The article mentions user reports, a.k.a., the Feedback button. It's really easy to implement, and every application should have one. Make it big and obvious and easy. You can't fix the defects if you don't know what they are. Your users will be *More than happy* to tell you all about the defects, if you make it clear how to do so, and that you are ready to listen.
2. This is not a joke. Any bug, and I mean any bug, once confirmed, should be a "stop the presses" event. If you can't fix the bug in 24 hours, create a build of the application with the failing feature removed. The best programmers, the ones that do 90% of the work in many OSS projects, know this intuitively. Making it an explicit policy would help any project greatly.
And don't forget -- confusing UI and incorrect documentation are defects too. I would define a defect, in fact, as any promise made by your application that it fails to fulfill.
Only government intervention can really ensure that such a system exists and until we demand it, it won't happen.
Well, I've got to disagree with you here, Steve. The legal system provides us with this great thing called a contract, and Clinton & friends gave us e-signatures so that we can create them electroniclly.
Your holding station doesn't have to hold the actual money, just the legally-binding promises to deliver that money upon content delivery. This means that the holding station doesn't have to be a full-fledged bank or lendor (credit card company). It just needs a lawyer.
To minimize legal fees, a user ranking system (trust metric) can be used, as can be seen on Ebay, Amazon, Advogato,/., etc. Content vendors can specify that their donations must come predominantly from "trusted" sources. You may be more concerned about the content providers taking the money and running -- well, the trust metric applies to them too.
The other key is that I see this work as being incremental. It's not like the provider will go develop their product for two years and then announce that they want $500,000 for it. More likely, products will be produced a tiny piece at a time ($100 - $10,000). The losses from abusers of the system will thus be rather small, especially if first-time producers are limited to smaller amounts.
A while ago, a friend of mine asked me how he could apply a sound economic model to the distribution of digital (a.k.a. easily reproducible) media. He wanted a system that fully accepted the near-uselessness of DRM technology. I told him about the "Street Performer Protocol".
This is the only model that makes sense to me in that it is clear, well-defined, and simple, yet complete. As the world "gets smaller", the information (knowledge) economy seems to be converging on a sort of minimum -- where the moment a piece of private information becomes public, it becomes public with a capital P (anyone who wants it will get it whether you like it or not). Digital technology allows the game of telephone to be played ad infinitum, and the message at the end of the line is the same as it was at the beginning. Sure, we can try to stretch the Copyright and Patent laws to fight this, but isn't the more intelligent solution to adapt to the new environment in a profitable way?
I have heard economists argue that "secrets" will become the most profitable asset in the information economy (as if they aren't already). This certainly applies to international politics and military affairs already.
In any case, it seems to me that SPP is in sync with all of this. And of course it applies to source code! I think that distributed development deserves a distributed payment system, based on SPP or something like it...
As for practicality, please note that SPP is not new or untested. Public Radio & Television, for example, has been doing it for decades: "We'll give you a quality stream of news/entertainment if and only if you pay us $X by date Y". And guess what -- it works. The government backs out of more of its commitment to funding public media each year, and yet the industry is here.
Probably the name is the worst part of the whole idea. I thought SPP was bad, but "Ransom" -- that's near idiotic -- the kind of name that makes great soundbites for the RIAA. Yeah, "Ransom" sucks. The idea of SPP is great though -- I just wonder why more folks aren't on the bandwagon yet?
BTW, the whole Stephen King experiment is an awful example of this, since there are so many external contributing factors. A fair first experiment with this concept would use a medium that is commonly distributed in digital format. While people do read from computer screens frequently, they do not tend to read novels on the computer. A more fair test would be in the distribution of music, software applications, software documentation, digital images, etc.
OK -- rant done.
Re:WebObjects is Web Applications done right.
on
Thursday Release Party
·
· Score: 2, Insightful
It simply is the easiest, most effective, and best product ever put out by Apple or anyone in the web applications / database backed website space.
OK, I aggree with kwerle -- I think all that WO has gone to your head;)
I will say that the OSS world would do well to quit copying M$ for a while and take a look around at all of the great, relatively hidden, proprietary software out there. Every time I see JSP and Jakarta Struts, I think fondly about my short time as a WO programmer. If all you know is VB and ASP, maybe that stuff looks good, but anyone who's had time to get to know WO knows better;)
Now, WO isn't *that* great -- EOF is the real diamond there, and even it could use improvements.
The great part is that this technology is so transparent, so clean, that I imagine it should be fairly simple to re-implement as OSS. Perhaps this is what the Tapestry project is trying to do...
And yes, WO is cheap, but remember that free as in pretzels isn't the point. OSS means being able to develop a community that is committed and resillient. A lot of people have been burned by NextStep's demise and business decisions (to a certain extent justifiable) by Apple that have restricted software development for one reason or another.
Well, maybe with all of these hackers messing with OSX, the ideas in WO will finally make it into mainstream middleware...
Low level languages will always coexist with high level ones.
Well, always is a pretty strong word, but otherwise you've really hit the nail on the head here. I think that the issue is not abstract vs. low-level, but rather how to keep the programmer focused on the task at hand (telling the computer what to do), while eliminating the mundanities (semicolon here, attach function A to class B using GUI tool blah...).
I think the answer is layers (which is what we have currently) taken to the extreme. I'd love to see an environment where one can design the application in purely abstract terms, perhaps with a visual representation. When a program unit needs to be modified for enhanced performance, one could "drill down" into the next lower-level language to write custom code for the job. If that's not enough, drop down another level, etc.
In a sense, this relationship already exists between many VM-based languages and the platforms they run on (Java and Lisp both have C/C++ at their core, for example). I'd like to see these relationships further fleshed-out and more accessible to the programmer.
Now there are tools out there that are supposed to provide this sort of flexibility. If anyone has experince with any of them, I'd be interested in hearing about it...
I didn't have to install anything unusual to use the following web services demos.
Also, web services are a back-end b2b communication platform first and foremost. Like any other back-end protocol, the front-end can be whatever you want -- http, http + applet, Flash, thin client, thick client, C, C++, C#, C&%^$&*?, whatever.
I check my mail using both a thick client, and a web browser. Both work fine and have their uses.
Sounds like your real beef is with.Net. Don't forget,.Net != web services. There are many of us developing web services with no intent to use.Net in any way.
Not only is this a great idea, it goes way beyond spam. How about "delete-as-off-topic" or "delete-as-rtfm" buttons specific to a given mailing list? The same algorithm could be used for these cases.
Take it a step further to organize your entire mailbox. How about "categorize-as-tech-support" or "categorize-as-jboss-related". Many of us already push our email around into folders for the purpose of organization. I can't see why this algorithm can't be used to assist that process as well.
The power of this system is that it is feedback-based. The software uses known science (statistics) to mold itself to your own preferences, by paying attention to the input that you have to make to use the application in the first place.
Why do you think there are businesses whose sole function is to track and to report on the input people make to the various machines in their lives (computer/websites/tv/etc.)? This information is powerful and we need more examples of the ethical use of it. Note that his system is completely "individual" and doesn't require sharing user input with others through a central server.
I haven't read the entire article, but I really think this is a great idea.
Slashdot Mirror
Al Fasoldt is right in part. We do need to critically analyze everything we read. He is, however, clearly fear-mongering when he singles out the Web, and Wikipedia in particular as especially untrustworthy.
As an american (I can't speak for anyone else), I am continually exposed to false information in TV and radio programs, books, magazines, casual conversation, and, yes, Al, on the Internet, and in the Wikipedia.
The ratio of fact to fiction in the Wikipedia will be higher than some sources of information, and lower than others. Likely, Al's readers are already familiar with the process of selecting one news source over another, as they chose his newspaper over his competitor's this morning, they must choose some TV and radio programs over others, etc.
The real discussion to be had here is on how, specifically, the Wikipedia process compares to the processes used to create and disseminate information by other sources. I'm not sure I can add much to that discussion, but I will say this -- When I see an error on the Wikipedia, I can amend it with a few mouse clicks and some typing. When I see an error in a textbook or in a newspaper or on TV, however, I can only send a letter to the editor, and hope that he sees fit to fix/acknowledge the error.
Without knowing your taste (beyond the selection you presented), and knowing that you have a small jazz collection at this point, I can reccommend to you the jazz canon (as I see it), as well as my personal favorites (of course).
The Canon -- The artists/albums a jazz collector should be familiar with regardless of specialty (dixieland through free jazz):
The Blues (Robert Johnson/Bessie Smith/etc.) -- Jazz starts with the Blues -- I've learned a lot about Jazz by listening to pre-Jazz Blues artists.
Louis "Pops" Armstrong - material from the early days where he plays trumpet at least as much as he sings.
Django Rheinhardt
Fats Waller
Billie Holiday
Ella Fitzgerald
Count Basie (w/ Lester Young)
Duke Ellington
Benny Goodman Septet (w/ Charlie Christian)
Lester Young
Coleman Hawkins
Oscar Peterson
Charlie Parker, anything
John Burkes "Dizzy" Gilespie, "Sunny Side Up", anything he did with Charlie Parker
Thelonious Monk (w/ Charlie Rouse)
King Pleasure, "Moody's Mood For Love"
Sonny Rollins, "Freedom Suite"
Miles Davis, "Birth of the Cool", "Kind of Blue"
Dave Brubeck (w/ Paul Desmond), "Time Out"
Stan Getz (w/ Gilberto)
Charlie Mingus
Cannonball Adderley
Wes Montgomery
John Coltrane, "Giant Steps"
Wayne Shorter, "Speak No Evil"
Ornette Coleman, "Shape of Jazz to Come"
Charlie Haden, "Liberation Music Orchestra"
My personal picks:
Tal Farlow
Bobby Hutcherson
Chic Corea, Origins band (w/ Avishai Coehn and Steve Wilson)
Keith Jarrett
Kenny Garrett, "Triology"
Wynton, Ellis, and Brandord Marsalis
James Moody
Charlie Mingus, "The Clown"
Oliver Nelson, "Blues and the Abstract Truth", "More Blues and the abstract Truth"
Benny Carter
Phil Woods
Yusef Lateef, "Eastern Sounds"
Charlie Parker, "Mango Mangue"
Tony Bennett
John McLoughlin (w/ Remember Shakti)
John Scofield (w/ Joe Levano if you can find it)
These lists, are, needless to say, very incomplete.
Remember when buying albums that until a certain point (in the 40's?), most records were more like today's singles, containing only a couple of tracks. As a result, compilations are probably your best bet for jazz until after Charlie Parker. There are, of course, many exceptions to this...
As you collect, you might also notice a gap between Swing and Bebop, where there is this radical shift in style. In reality, there was an extensive recording ban (strike), that coincided with the inception of bebop. Recordings from this period are highly valued by collectors.
Good Luck!
A) How exactly is Apple "choosing to enforce the DMCA"?
;)
They complained.
B) The GPL only has meaning because of copyright.
The GPL is an attempt to modify or extend traditional copyright. It is a set of limitations that make publications more free. It is copyleft. The DMCA has several provisions that make publications less free. I won't bother enumerating them here.
GPL != GNU. GPL is a tool that the GNU community uses to enable the production and protection of free publications. GPL relies on copyright, but the notion of freedom does not. This notion predates Gutenberg significantly
Guys, please don't ignore the "GNU" in GNU-Darwin.
The DMCA and laws like it threaten the existence of GNU. GNU-Darwin is part of GNU. By encouraging GNU-Darwin, Apple helps GNU. By choosing to enforce the DMCA, Apple threatens GNU. This contradictory stance must be evaluated by determining which is stronger, the support, or the threat.
Apple is using its lawyers to support the DMCA. Apple is not using its lawyers to support GNU (see the GNU commentary on the APSL).
Of course, only the GNU folks can determine what is best for them. it appears that they have done so, and that they made the determination that (perhaps) loosing one project is better than loosing the entire organization (or its mission / goals).
I can't see anything contradictory or stupid about this decision. The only contradictory thing I can see here is Apple's behavior.
Thanks. They don't exactly have a Developer category, but this certainly is a step in the right direction.
As a musician and a software developer, I know intuitively what OSS and the internet community in general has to offer musicians. I recently downloaded Audacity, for example, and I was able to use it out-of-the-box with no problems -- something I can't quite say for some of its proprietary counterparts.
:) ...
But as a software developer, I'm missing the community websites, mailing lists, and general online discussion that I enjoy in other domains. Where do music software developers "hang out" online? (I'm specifically referring to those people developing software tools for musicians, as opposed to those developing file-trading and music-player applications.)
There are plenty of websites about music out there, but 99% appear to be fansites or e-business sites for music industry companies.
From time to time I toss around the idea of putting together a mailing list / website for music technology discussions (maybe slashcode is good for this?). If I did, would anyone use it? Or does something comparable already exist, and I am just missing it?
BTW, the musical community needs OSS badly. Leaving technology standards up to the music industry has left us with crappy, slowly evolving protocols with little innovation. With all due respect, MIDI was great in its day, but it really doesn't meet the tests of time, IMHO. Also, I know that musicXML is on the way, but I'm still skeptical
If you want "usability", connect the user to the developer in a meaningful way, by adding the following features to your project.
1. Feedback Button
2. Zero Defect Tolerance
1. The article mentions user reports, a.k.a., the Feedback button. It's really easy to implement, and every application should have one. Make it big and obvious and easy. You can't fix the defects if you don't know what they are. Your users will be *More than happy* to tell you all about the defects, if you make it clear how to do so, and that you are ready to listen.
2. This is not a joke. Any bug, and I mean any bug, once confirmed, should be a "stop the presses" event. If you can't fix the bug in 24 hours, create a build of the application with the failing feature removed. The best programmers, the ones that do 90% of the work in many OSS projects, know this intuitively. Making it an explicit policy would help any project greatly.
And don't forget -- confusing UI and incorrect documentation are defects too. I would define a defect, in fact, as any promise made by your application that it fails to fulfill.
Only government intervention can really ensure that such a system exists and until we demand it, it won't happen.
/., etc. Content vendors can specify that their donations must come predominantly from "trusted" sources. You may be more concerned about the content providers taking the money and running -- well, the trust metric applies to them too.
Well, I've got to disagree with you here, Steve. The legal system provides us with this great thing called a contract, and Clinton & friends gave us e-signatures so that we can create them electroniclly.
Your holding station doesn't have to hold the actual money, just the legally-binding promises to deliver that money upon content delivery. This means that the holding station doesn't have to be a full-fledged bank or lendor (credit card company). It just needs a lawyer.
To minimize legal fees, a user ranking system (trust metric) can be used, as can be seen on Ebay, Amazon, Advogato,
The other key is that I see this work as being incremental. It's not like the provider will go develop their product for two years and then announce that they want $500,000 for it. More likely, products will be produced a tiny piece at a time ($100 - $10,000). The losses from abusers of the system will thus be rather small, especially if first-time producers are limited to smaller amounts.
A while ago, a friend of mine asked me how he could apply a sound economic model to the distribution of digital (a.k.a. easily reproducible) media. He wanted a system that fully accepted the near-uselessness of DRM technology. I told him about the "Street Performer Protocol".
This is the only model that makes sense to me in that it is clear, well-defined, and simple, yet complete. As the world "gets smaller", the information (knowledge) economy seems to be converging on a sort of minimum -- where the moment a piece of private information becomes public, it becomes public with a capital P (anyone who wants it will get it whether you like it or not). Digital technology allows the game of telephone to be played ad infinitum, and the message at the end of the line is the same as it was at the beginning. Sure, we can try to stretch the Copyright and Patent laws to fight this, but isn't the more intelligent solution to adapt to the new environment in a profitable way?
I have heard economists argue that "secrets" will become the most profitable asset in the information economy (as if they aren't already). This certainly applies to international politics and military affairs already.
In any case, it seems to me that SPP is in sync with all of this. And of course it applies to source code! I think that distributed development deserves a distributed payment system, based on SPP or something like it...
As for practicality, please note that SPP is not new or untested. Public Radio & Television, for example, has been doing it for decades: "We'll give you a quality stream of news/entertainment if and only if you pay us $X by date Y". And guess what -- it works. The government backs out of more of its commitment to funding public media each year, and yet the industry is here.
Probably the name is the worst part of the whole idea. I thought SPP was bad, but "Ransom" -- that's near idiotic -- the kind of name that makes great soundbites for the RIAA. Yeah, "Ransom" sucks. The idea of SPP is great though -- I just wonder why more folks aren't on the bandwagon yet?
BTW, the whole Stephen King experiment is an awful example of this, since there are so many external contributing factors. A fair first experiment with this concept would use a medium that is commonly distributed in digital format. While people do read from computer screens frequently, they do not tend to read novels on the computer. A more fair test would be in the distribution of music, software applications, software documentation, digital images, etc.
OK -- rant done.
It simply is the easiest, most effective, and best product ever put out by Apple or anyone in the web applications / database backed website space.
;)
;)
OK, I aggree with kwerle -- I think all that WO has gone to your head
I will say that the OSS world would do well to quit copying M$ for a while and take a look around at all of the great, relatively hidden, proprietary software out there. Every time I see JSP and Jakarta Struts, I think fondly about my short time as a WO programmer. If all you know is VB and ASP, maybe that stuff looks good, but anyone who's had time to get to know WO knows better
Now, WO isn't *that* great -- EOF is the real diamond there, and even it could use improvements.
The great part is that this technology is so transparent, so clean, that I imagine it should be fairly simple to re-implement as OSS. Perhaps this is what the Tapestry project is trying to do...
And yes, WO is cheap, but remember that free as in pretzels isn't the point. OSS means being able to develop a community that is committed and resillient. A lot of people have been burned by NextStep's demise and business decisions (to a certain extent justifiable) by Apple that have restricted software development for one reason or another.
Well, maybe with all of these hackers messing with OSX, the ideas in WO will finally make it into mainstream middleware...
Low level languages will always coexist with high level ones.
Well, always is a pretty strong word, but otherwise you've really hit the nail on the head here. I think that the issue is not abstract vs. low-level, but rather how to keep the programmer focused on the task at hand (telling the computer what to do), while eliminating the mundanities (semicolon here, attach function A to class B using GUI tool blah...).
I think the answer is layers (which is what we have currently) taken to the extreme. I'd love to see an environment where one can design the application in purely abstract terms, perhaps with a visual representation. When a program unit needs to be modified for enhanced performance, one could "drill down" into the next lower-level language to write custom code for the job. If that's not enough, drop down another level, etc.
In a sense, this relationship already exists between many VM-based languages and the platforms they run on (Java and Lisp both have C/C++ at their core, for example). I'd like to see these relationships further fleshed-out and more accessible to the programmer.
Now there are tools out there that are supposed to provide this sort of flexibility. If anyone has experince with any of them, I'd be interested in hearing about it...
just install this 300MB library+runtime first
.Net. Don't forget, .Net != web services. There are many of us developing web services with no intent to use .Net in any way.
I didn't have to install anything unusual to use the following web services demos.
Also, web services are a back-end b2b communication platform first and foremost. Like any other back-end protocol, the front-end can be whatever you want -- http, http + applet, Flash, thin client, thick client, C, C++, C#, C&%^$&*?, whatever.
I check my mail using both a thick client, and a web browser. Both work fine and have their uses.
Sounds like your real beef is with
Not only is this a great idea, it goes way beyond spam. How about "delete-as-off-topic" or "delete-as-rtfm" buttons specific to a given mailing list? The same algorithm could be used for these cases.
Take it a step further to organize your entire mailbox. How about "categorize-as-tech-support" or "categorize-as-jboss-related". Many of us already push our email around into folders for the purpose of organization. I can't see why this algorithm can't be used to assist that process as well.
The power of this system is that it is feedback-based. The software uses known science (statistics) to mold itself to your own preferences, by paying attention to the input that you have to make to use the application in the first place.
Why do you think there are businesses whose sole function is to track and to report on the input people make to the various machines in their lives (computer/websites/tv/etc.)? This information is powerful and we need more examples of the ethical use of it. Note that his system is completely "individual" and doesn't require sharing user input with others through a central server.
I haven't read the entire article, but I really think this is a great idea.