Except that what I posted was a link to papers with a Microsoft researcher amongst the authors, and hence provides evidence that Microsoft does in fact do real research that gets peer reviewed and found to advance the "state of the art".
The vast majority of what is happening on the "bleeding edge" is happening it other places than *any* one place. Bleeding edge stuff is done by lots of people in lots of places, there is no single majority contributor.
Microsoft does real research, their staff get published in real journals, and present at real conferences.
They did Bayesian Spam Filtering a couple of years before Paul Graham reinvented a less useful version, as one example.
Just because the work of the research arm doesn't get reflected in their commercial products doesn't mean they don't do good research. It just means their marketers know that the public buys shiny things in preference to useful things.
If those people don't switch ISPs then they clearly don't care if some of the legitimate email destined for them is blocked. Or at least, they think the reduction in spam is worth that cost.
If they cared about the legitimate mail getting blocked they would change to a mail provider that didn't do that.
Just as they are free to but a phone that doesn't ring when the no caller-id information is provided for an incoming call.
You can use an independant (of ISP) email provider for your email. Again the "invisible hand" of the free market should ensure a supply of such services at a reasonable price, if people find the ISP filtering a problem.
And of course ISPs should see a market from which they can make more profit - though charging extra for not filtering sounds almost like extortion.
If people care enough about legitimate mail getting blocked by their ISP's filters they will change ISPs.
The invisible hand of market forces will mean that ISPs that offer the functionality you mention will get more customers and hence make more money and not disappear.
Of course, possibly, people care more about not wading through hundreds of spam emails then they do about missing out on email from places who can't configure a mail server correctly from the start or check the bounces in a timely fashion, or pay someone else to deliver the mail.
I quoted from a section a bit further down in that manpage - which indicates I just might have the read the damn thing.
That spamassassin has a limit that my sample data didn't reach isn't of real concern to me. I can't just magically create some emails, I only have the emails that I have recieved.
A bayesian filter should work reasonably well with unbalanced training data. A Paul Graham style "let's ignore the huge amount of research in the field and make stuff up" filter will have problems because it ignores the term in a Naive Bayesian Classifier that deals with the ratios of each type of item.
The testing was done a month after the actual emails were recieved. Using such resources would allow the filter the benefit of hindsight. As in foo sends lots of spam and ends up on a blacklist, but I recieved a bunch of spam from foo before it got on the list.
So it wasn't artificial. I mentioned in the article why I made that constraint.
I also didn't retrain bayesian filters on false-negatives before giving them later emails, which isn't normal use of them either.
I even did some stats stuff and found that there was a significant performance difference between some of the filters - but I don't trust my stats knowledge enough to publish such things without getting them checked. SInce I didn't get them checked, I didn't include them.
If the article was meant for a machine learning journal then obviously it's a joke. But it wasn't it was meant for freshmeat, the requirements are much lower.
That is replicating it. Biological processes are common in industry (well so they claimed when advertising the biochemistry courses before I gave up on chem. eng. and transferred to CS).
Of course you wouldn't harvest them for their 2 to 7 inch long bits of fibre optic.
You try and replicate the process the sponges use. It at least shows it is possible to make the stuff at cold temperatures, which as the article states (which you obviously didn't bother comprehending, and probably reading) makes doping the glass easier.
"equal productivity" means there are no productivity gains...
Large enterprises get that equal productivity at significantly lower cost since, being free software, they can install Linux on as many machines as they want without paying extra for the priviledge.
For smaller enterprises the cost savings are lower, since they require fewer Windows licenses in order to use Windows.
And just how do you know the file and MD5 sum wasn't compromised sometime during the day before the backup was made?
Or are you planning on doing a back up every second and hoping the cracker can't win the race?
*Every* change to a file on that server since March is suspect. Yes you can recover all the files prior to the crack, but you also want to good data from after the crack.
Hence they need MD5 sums from people who got the files from somewhere else (which didn't in turn get them from the compromised server...).
On the bright side, it means they will finally start using signed MD5 sums. Something which everyone should have been doing years ago.
I never claimed it was an MTA. I claimed that the email client I use (mh/nmh) was very much a collection of small programs. I claimed my MTA was too, but I was referring to qmail (it's no where near as split up as mh, though).
I suspect everyone who uses mh writes lots of scripts for it, it makes it so easy that it's hard not to get carried away.
You'd expect the a message is a file thing to slow everything down, and it does in some cases, but I currently have 1.7GB of email in over 250,000 messages under mh - and it runs just fine...
This is exactly my point. There are many web browsers out there, and not one is built from program components connected linearly via pipes. Is this because everyone building a web browser is an idiot who just doesn't understand this magical Unix Philosophy? Or is it because this design strategy of small commands interconnected by linear pipes does not scale to the implementation of modern, large programs?
I don't know of a GUI environment that supports such development. Web browsers use plugins, those plugins could be seperate programs which communicate with the browser via pipes - but that makes drawing to the appropriate locations on the screen difficult.
I never claimed the unix philosophy was magical. It's just nice from a user point of view. Wouldn't it be nice to be able to tell your web browser to execute "favourite editor" inside the text area you type things like these comments in. That's harder than it is with a terminal program since it has to know where to draw itself. It could be done, but require a different GUI windowing practice. Doing it now is too difficult, the ground work doesn't exist and hence there's no benefit.
Ditto for compilers. A compiler is a very common, useful, complicated tool, and of the hundreds I've used (GNU or otherwise), not one has been implemented by a set of small programs interconnected by pipes. I've never seen a register allocator connected to an interference graph calculator connected to a register spiller, etc. etc. Every single compiler is a monolithic application, whether written by the GNU project or not.
Compilers are one of the few application which need to be fast - they don't wait on user input or on the network. Converting representations of the data structures generated by compilers into text, piping them, and then converting the text back into data structures would be slow.
The unix philosophy doesn't mean you can't optimise the things that actually need optimising.
Compilers use multiple processes, preprocessors, compiler proper, assembler, and linker in most cases - these can be bundled in one binary, but you still get access to the pieces. Which allows you to combine them in other ways (starting at the assembler for example).
The compiler does one thing and does it well. It compiles a programming language into assembler. Now GCC is internally modular allowing different languages and yes it would be nice if those modules were seperate processes so you could add a new language via a filter program (though the FSF wouldn't want that since it would be an end run around the GPL).
The philosophy isn't "break everything down into the smallest reusable modules and make them seperate programs". It's just "programs should do one thing only and deal with text if at all possible".
And it isn't the fastest way of doing things, but for lots of tasks it is fast enough. CGI scripts are examples of the unix philosophy - the web server runs other programs to do other tasks. Of course it's more efficient to avoid the fork and exec, so we have things like mod_perl and mod_php these days.
My web browser is a monolithic app, not connected by pipes.
Your web browser probably wasn't implemented by people following the unix philosophy. The Unix GUI environment doesn't support small tools as well as the command line environment.
GCC is a couple of monolithic applications, optionally connected by pipes, but never reconnected in any useful way (cpp notwithstanding).
GCC is GNU software. GNU software does not follow the unix philosophy. GNU software follows the add a command line option for everything and the kitchen sink philosophy.
My newsreader and mailreader again, monolithic applications.
My newsreader, runs my prefered editor when I need to write a post. For a little while I used my own hacked up newsreader which had seperate unix commands for most actions.
My mail client consists of 57 individual command line programs (one for listing messages, one for changing folders, one for showing messages, one for marking messages, etc.)
My MTA, again, a monolithic application.
Mine is a collection of programs that each do one thing, and do it well.
Not one large program I use is a shell script, or collection of small, interchangeable programs.
But they do exist, and are more useful in my opinion, but have a steeper learning curve.
Re:This makes me think of .....
on
More on Spintronics
·
· Score: 2, Informative
Your thought is incorrect. Einstein assumed gravity is limited by the speed of light. An experiment done late last year involving Jupiter passing in front of a quasar seemed to confirm that assumption. Though some believe the expirement was flawed.
The GPL gives a net benefit of rights, but to say "it only grants rights" is wrong, wrong, wrong.
Name one thing I can do with some source code copyrighted by someone else wihout any license or agreement with them, that I can not do with some source code licensed to meunder the GPL.
Just one will do.
If you can't then the GPL hasn't taken any rights away, has it?
All it has done is add some conditional rights, if you don't want the rights you don't have to worry about the conditions. Nothing has been taken away (except maybe in the warranty disclaimer).
Slashdot Mirror
Maybe you should go back to "Comprehension for Dummies 101".
What the hell does LiveJournal have to do with the fact that the posts on slashdot are not all made by a single person?
http://searchthis.mozdev.org/
Except that what I posted was a link to papers with a Microsoft researcher amongst the authors, and hence provides evidence that Microsoft does in fact do real research that gets peer reviewed and found to advance the "state of the art".
The vast majority of what is happening on the "bleeding edge" is happening it other places than *any* one place. Bleeding edge stuff is done by lots of people in lots of places, there is no single majority contributor.
Microsoft does real research, their staff get published in real journals, and present at real conferences.
They did Bayesian Spam Filtering a couple of years before Paul Graham reinvented a less useful version, as one example.
Just because the work of the research arm doesn't get reflected in their commercial products doesn't mean they don't do good research. It just means their marketers know that the public buys shiny things in preference to useful things.
I take it you have personally advanced the state of the art more than this.
You can't. But that isnt the point.
If those people don't switch ISPs then they clearly don't care if some of the legitimate email destined for them is blocked. Or at least, they think the reduction in spam is worth that cost.
If they cared about the legitimate mail getting blocked they would change to a mail provider that didn't do that.
Just as they are free to but a phone that doesn't ring when the no caller-id information is provided for an incoming call.
You can use an independant (of ISP) email provider for your email. Again the "invisible hand" of the free market should ensure a supply of such services at a reasonable price, if people find the ISP filtering a problem.
And of course ISPs should see a market from which they can make more profit - though charging extra for not filtering sounds almost like extortion.
If people care enough about legitimate mail getting blocked by their ISP's filters they will change ISPs.
The invisible hand of market forces will mean that ISPs that offer the functionality you mention will get more customers and hence make more money and not disappear.
Of course, possibly, people care more about not wading through hundreds of spam emails then they do about missing out on email from places who can't configure a mail server correctly from the start or check the bounces in a timely fashion, or pay someone else to deliver the mail.
Do you really expect "newbie users of Linux" to be using Gentoo?
Redhat has a GUI interface to their up2date software for users who don't like typing commands.
I work in a university setting, and I can tell you that having a PHd will not save you from accidentally opening this virus.
Having a PhD is a sure sign of a lack of common sense, so that really isn't surprising.
I quoted from a section a bit further down in that manpage - which indicates I just might have the read the damn thing.
That spamassassin has a limit that my sample data didn't reach isn't of real concern to me. I can't just magically create some emails, I only have the emails that I have recieved.
A bayesian filter should work reasonably well with unbalanced training data. A Paul Graham style "let's ignore the huge amount of research in the field and make stuff up" filter will have problems because it ignores the term in a Naive Bayesian Classifier that deals with the ratios of each type of item.
The testing was done a month after the actual emails were recieved. Using such resources would allow the filter the benefit of hindsight. As in foo sends lots of spam and ends up on a blacklist, but I recieved a bunch of spam from foo before it got on the list.
So it wasn't artificial. I mentioned in the article why I made that constraint.
I also didn't retrain bayesian filters on false-negatives before giving them later emails, which isn't normal use of them either.
I did a ten fold cross validation.
I even did some stats stuff and found that there was a significant performance difference between some of the filters - but I don't trust my stats knowledge enough to publish such things without getting them checked. SInce I didn't get them checked, I didn't include them.
If the article was meant for a machine learning journal then obviously it's a joke. But it wasn't it was meant for freshmeat, the requirements are much lower.
They reduced the time span for the movie. After all the exciting rush to leave the shire just isn't as exciting when it takles place over two decades.
Turning LoTR into as movie without making such major changes would result in the most mind numbingly boring film ever made.
That is replicating it. Biological processes are common in industry (well so they claimed when advertising the biochemistry courses before I gave up on chem. eng. and transferred to CS).
Of course you wouldn't harvest them for their 2 to 7 inch long bits of fibre optic.
You try and replicate the process the sponges use. It at least shows it is possible to make the stuff at cold temperatures, which as the article states (which you obviously didn't bother comprehending, and probably reading) makes doping the glass easier.
"equal productivity" means there are no productivity gains...
Large enterprises get that equal productivity at significantly lower cost since, being free software, they can install Linux on as many machines as they want without paying extra for the priviledge.
For smaller enterprises the cost savings are lower, since they require fewer Windows licenses in order to use Windows.
Why would there be anything but text on the screen when you are coding?
Have you thought of using bigger fonts?
You know, like everyone else does who uses high resolutions.
And just how do you know the file and MD5 sum wasn't compromised sometime during the day before the backup was made?
Or are you planning on doing a back up every second and hoping the cracker can't win the race?
*Every* change to a file on that server since March is suspect. Yes you can recover all the files prior to the crack, but you also want to good data from after the crack.
Hence they need MD5 sums from people who got the files from somewhere else (which didn't in turn get them from the compromised server...).
On the bright side, it means they will finally start using signed MD5 sums. Something which everyone should have been doing years ago.
I never claimed it was an MTA. I claimed that the email client I use (mh/nmh) was very much a collection of small programs. I claimed my MTA was too, but I was referring to qmail (it's no where near as split up as mh, though).
I suspect everyone who uses mh writes lots of scripts for it, it makes it so easy that it's hard not to get carried away.
You'd expect the a message is a file thing to slow everything down, and it does in some cases, but I currently have 1.7GB of email in over 250,000 messages under mh - and it runs just fine...
This is exactly my point. There are many web browsers out there, and not one is built from program components connected linearly via pipes. Is this because everyone building a web browser is an idiot who just doesn't understand this magical Unix Philosophy? Or is it because this design strategy of small commands interconnected by linear pipes does not scale to the implementation of modern, large programs?
I don't know of a GUI environment that supports such development. Web browsers use plugins, those plugins could be seperate programs which communicate with the browser via pipes - but that makes drawing to the appropriate locations on the screen difficult.
I never claimed the unix philosophy was magical. It's just nice from a user point of view. Wouldn't it be nice to be able to tell your web browser to execute "favourite editor" inside the text area you type things like these comments in. That's harder than it is with a terminal program since it has to know where to draw itself. It could be done, but require a different GUI windowing practice. Doing it now is too difficult, the ground work doesn't exist and hence there's no benefit.
Ditto for compilers. A compiler is a very common, useful, complicated tool, and of the hundreds I've used (GNU or otherwise), not one has been implemented by a set of small programs interconnected by pipes. I've never seen a register allocator connected to an interference graph calculator connected to a register spiller, etc. etc. Every single compiler is a monolithic application, whether written by the GNU project or not.
Compilers are one of the few application which need to be fast - they don't wait on user input or on the network. Converting representations of the data structures generated by compilers into text, piping them, and then converting the text back into data structures would be slow.
The unix philosophy doesn't mean you can't optimise the things that actually need optimising.
Compilers use multiple processes, preprocessors, compiler proper, assembler, and linker in most cases - these can be bundled in one binary, but you still get access to the pieces. Which allows you to combine them in other ways (starting at the assembler for example).
The compiler does one thing and does it well. It compiles a programming language into assembler. Now GCC is internally modular allowing different languages and yes it would be nice if those modules were seperate processes so you could add a new language via a filter program (though the FSF wouldn't want that since it would be an end run around the GPL).
The philosophy isn't "break everything down into the smallest reusable modules and make them seperate programs". It's just "programs should do one thing only and deal with text if at all possible".
And it isn't the fastest way of doing things, but for lots of tasks it is fast enough. CGI scripts are examples of the unix philosophy - the web server runs other programs to do other tasks. Of course it's more efficient to avoid the fork and exec, so we have things like mod_perl and mod_php these days.
mh or nmh, depending on which machine I happen to be logged into.
My web browser is a monolithic app, not connected by pipes.
Your web browser probably wasn't implemented by people following the unix philosophy. The Unix GUI environment doesn't support small tools as well as the command line environment.
GCC is a couple of monolithic applications, optionally connected by pipes, but never reconnected in any useful way (cpp
notwithstanding).
GCC is GNU software. GNU software does not follow the unix philosophy. GNU software follows the add a command line option for everything and the kitchen sink philosophy.
My newsreader and mailreader again, monolithic applications.
My newsreader, runs my prefered editor when I need to write a post. For a little while I used my own hacked up newsreader which had seperate unix commands for most actions.
My mail client consists of 57 individual command line programs (one for listing messages, one for changing folders, one for showing messages, one for marking messages, etc.)
My MTA, again, a monolithic application.
Mine is a collection of programs that each do one thing, and do it well.
Not one large program I use is a shell script, or collection of small, interchangeable programs.
But they do exist, and are more useful in my opinion, but have a steeper learning curve.
Your thought is incorrect. Einstein assumed gravity is limited by the speed of light. An experiment done late last year involving Jupiter passing in front of a quasar seemed to confirm that assumption. Though some believe the expirement was flawed.
Name one thing I can do with some source code copyrighted by someone else wihout any license or agreement with them, that I can not do with some source code licensed to meunder the GPL.
Just one will do.
If you can't then the GPL hasn't taken any rights away, has it?
All it has done is add some conditional rights, if you don't want the rights you don't have to worry about the conditions. Nothing has been taken away (except maybe in the warranty disclaimer).