Why is there such a rush of governments, the U.N., and so many others already pledging support for this initiative? It sounds like it could be a good idea, but shouldn't people wait to see whether the thing actually works? They have a whole boatload of new, innovative features like automatic peer-to-peer wireless, their own specialized Linux software, etc. -- surely it'll take time to work out the kinks...
Just a thought: automated webpage translation services like Babelfish and Google's language tools can translate arbitrary webpages yet preserve images and navigation. This only works because the web works via hyperlinked text documents. But if you want to read a website that uses AJAX for navigation, these systems will break -- they can't intercept your XmlHTTPRequests. And even if there was a browser extension that could do that, it would be unclear what text needed to be translated and when.
So here's one example of a useful tool that depends on HTML transparency, broken by more programmatic clientside techniques.
The big thing in python web programming right now is the introduction of Django, a mature RAD framework that shares lots of features with Rails. It's got a lot going for it; it'll be interesting to see how things turn out.
I find CherryPy's URL traversal scheme a bit clunky -- since you connect up objects to each other via attributes, you can't see the hierarchy of your site. At least with PHP you can use "ls" to discover what your URL space looks like. Django uses a really neat scheme that binds a table of named regular expressions to callable handlers, e.g.
The entire phenomenon of packaging, or even the existence of distros, is odd. The developers of a piece of software do source releases, then have to cross their fingers that developers from other distros will package their work. It would be nice if the original software's developers took the next step and made binary packages, but it's impossible for them to package for every linux distro in existence. Therefore, it's expected that other people are supposed to come along and package it for their own distribution.
This system is unsustainable for a mass-market Linux world. If I want a piece of software, I should be able to go to a website run by the people who make it, and download it from there. Currently, I go to the website, learn about the software, THEN use seperate non-web tools to search my distro's own rpm/deb database to try to find if it already has a distro-sanctioned package. If that fails, I then go back to the software's website and hunt around for the best solution, falling back on a source release if necessary.
This "packagers-are-not-developers" system is especially bad in two situations.
(1) there's so much useful software out there, a distro maintainer can't keep up. This might already be the case. Distro packages are always several versions behind what you can get on upstream's website. Upstream's website does a much better job at telling users what the software is. To keep your distro up to speed, you then have to decentralize its production (e.g. thousands of Debian volunteers) and potentially introduce horrendous organizational/managerial difficulties (e.g. three years between releases.)
And, (2) commercial developers are screwed, because they have to do all the packaging themselves. This applies particularly to closed-source, but also to any software developed by a small or specialized group of people. This problem is solved now because enough Linux users are developers such that popular software gets packaged; but a higher non-dev:developer ratio implies more and more people can't get software they want.
Both (1) and (2) are terrible barriers to mass Linux adoption, because they imply limits on the 3rd party development community when most users aren't developers anymore. Linux world needs to have a unified technical system so it can have a unified social system where a software's developer can make end-user-consumable packages themselves.
UnitedLinux failed... LSB seems weak... maybe we'll only achieve this once some WalmartLinux storms the world and everyone codes for it as a de facto standard. Until then, let's watch those commons stayin' on tragic...
The problem of deciding good and bad would still remain. Perhaps a Wikipedia-style approach with volunteer evaluators?
They're running a project like this, to teach computers from volunteers. (A.k.a., collect volunteer-supplied data for various machine learning tasks.) It's Open Mind. Looks like they're not doing natural language grammar per se, but something similar could be used for that.
Grammar's hard, though. Linguists have been working for some 50 years to explicitly write down the rules differentiating what we percieve as good versus bad; and they still haven't succeeded. We still can't enumerate or describe the intuitions we all have in our heads about this.
This sounds like a great idea, provided that the Knoppix can be user-friendly enough to figure out how to boot up.
There's really no surefire way to ensure that a user's harddrive-installed OS is secure for banking. Considering the staggering variety of adware/spyware/viruses on machines today, it must be quite easy for a malicious malware creator to make a program that hijacks name resolution (change DNS servers, or the HOSTS file) for perfect phishing, or they could install a keystroke logger, or whatever else. If they got their bank-website-hijacking malware on machines in whatever way all today's adware stuff gets on, they could easily phish thousands of bank transactions every day.
The prevalence of malware seems to indicate that people can't control or trust the programs on their own hard drives. If that's the case, they can't trust any of their online interactions. Since Knoppix kills your harddrive and all its flexibility, it's much more secure.
What would be funny is if more and more institutions started demanding the use of bootable OS's. Our PC's would be reduced to a BIOS, monitor, and keyboard... reminds you of the Apple II days, where you had to boot half of the operating system off a floppy every time you turned on the computer.
THe Lookout program was mentioned a bit back on Joel on Software and interestingly enough, it looks like they used the Apache Foundation's open-source Lucene full-text indexer to create a good search plugin for Outlook.
I think an often overlooked reason that floppies have survived so long is because speed and size aren't all that's important to make a media useful. The system of floppy disks for data storage and sharing still beats out all these newfangled systems -- USB drives, CD's, Zip, networks -- in several ways.
1) Read/write is transparent. The burning step for CDRW is terrible; you should be able to directly open, save, and erase files just like any other drive. Then you don't need to copy files to your harddrive to work on them and then back again when you're done; that eventually invites confusion. The most prevalent network transfer protocols require separate download/upload steps.
2) The media is physically robust. Unlike CD's, a protective case isn't critical for floppies. Floppies do not start flaking out after being scratched a number of times. They're easier to transport and share -- I can put them in a backpack and run around all day without the flimsy plastic case breaking. And the fact you can write on them with a normal pen increases usefulness too: labelling is really helpful for yourself and essential for sharing.
And unlike USB drives, floppies have a standardized size, so you can stack them and store them in standardized cases.
3) The media is cheap, which facilitates sharing. USB drives cost lots of money; to give your data to someone you can't just hand them a spare drive. Floppies, even the older high quality ones, are cheap enough to give away.
With cheap media, you can afford to use a labelled disk as a unit of classification -- you don't need to fill up the disk to get your money's worth. USB drives can't do this (yet).
Expensive drives inside computers paired with cheap disks is much better than expensive combined disk+drives that can be swapped between computers. A good universal physical medium should be usable on all computers; it's not like the act of transferring files is something that only the rare person with a usb stick wants to do. You should only have to have a cheap disk to transfer files; you should not have to invest in a special drive.
To transfer files I once had to go around knocking on doors, looking for someone with a USB drive. This is ridiculous. (I am more likely to have a spare floppy, or only have to go knocking around for a floppy!)
4) Media reading/writing is (was) universal. CD drives are universal, but not always for writing. USB is pretty good now, but it can be a pain to find the plug in the back of the machine; I've also had weird OS hangups on certain systems (esp. older windows). Networks aren't always available in all environments -- especially figuring out which server or transfer protocol to use that will work for your particular situation.
Universality was definitely a bane of Zip drives and other floppy replacements -- a media type is useful only if everyone else has it.
5) They're dead easy to use. The CD burn step and usb issues were mentioned above. Further, network transfers are a pain. I've had the most annoying experiences just figuring out how to network transfer a file from one computer to another. Maybe you can upload/download via ftp -- if you have a server around, and you even know what ftp is? Maybe use email -- which requires extra space in someone's mailbox, and through web interfaces is often even clunkier than ftp? And the login steps are definitely extraneous. Store on a network drive -- if you have a server available nearby? Computers still can't universally detect each other's presence and sling around individual files without depending on some remote server. The easiest and most common way to transfer files I've observed on campus is to have an AOL IM signon on each computer, then use its file transfer mechanism. This is ridiculous. If files still fit on floppies this situation would be so much easier.
Obviously, it's possible to solve the peer-to-peer transfer problem via better and more universal pr
o snapshot of telephone penetration matters, because the issue is not amount but rate. If you care about the digital divide, and you believe that access to communications can help poor countries to grow, then pontificating about who has or hasn't made a phone call is worse than a waste of time, it actively distorts your view of the possible solutions because it emphasizes a statist attitude.
Wrong: it doesn't imply hopelessness, but rather encourages us to take action to change. Do you think that Kofi Annan wants us to throw up our hands and not care about the rate of improvement? No! By recognizing the magnitude of the problem, we can realize how important <i>more</i> improvement is. Just because things are improving doesn't mean we shouldn't be concerned about the huge inequalities that exist.
Well, the success of an online community is more important than its size. The fact that it can work in either a small place like Blacksburg or a big one like Seattle says something about the entire idea. I'm wondering what elements are common to these different examples. Thanks for the comments.
Many of these sites actually have started with a newspaper site, or later get taken up by a newspaper for funding reasons. It's definitely a good idea, except that sometimes they could need to do things beyond the scope of a newspaper -- like host websites for community groups, provide a forum to criticize media coverage of a particular event, etc.
Good point. One of the questions I'm researching is, can it be successful among poorer communities as well? You then have to deal with digital divide issues at the same time -- very tricky.
No, it's the immature obsession with sex so prevalent among male, adolescent-minded geeky community that's offensive. Only an overwhelmingly male subculture will be so disgusting across the board. While making dirty sex jokes isn't always bad, the fact that EVERYONE here is making them, and laughing at them (all +5 funny), the environment becomes intimidating and hostile towards women. The all-male reinforces the sexism of the computer science world.
$ grep 'res\\' *|wc -l grep: Debug: Is a directory
75
Unfortunately, there is no 'res' directory for these 75 references to it. The project descriptor files refer to dozens of files THAT DO NOT EXIST.
There are no makefiles, there are no.ico's or.bmp's.... No explanation how or where to get them, no instructions how to make the thing go. If all the "Source Code" link on a website does is give you raw, instruction-less code without data and descriptor files critical to actually _building_ something, the GPL hasn't been fulfilled.
For the sake of everyone, Streamcast, get this together!
Take this as a warning. No matter how "cool" a piece of software is, if it is proprietary software, it is absolutely worthless. It can disappear at any moment, and it contributes nothing to the progress of computer science.
Give me a break. The world's a lot bigger than computer science -- real people in the real world actually use computer software to actually do things. Not everyone is living in luxury enough to sit around and tinker with the guts of their operating systems. Playing around with Linux is fun, but programming is NOT the end-all be-all of programs. It's a shame that the insightful parts of your post about corporate power have to be overshadowed by your techno-elitism.
..is AllegroHack, for Windows and Linux. It's really sweet, and IMHO more complete than Falcon's Eye (the last time I played it a monkey's picuture was that of a guard!)
Also, Allegrohack works with SLASH'EM a really fun version of nethack (if a wee bit unbalanced) where you get more abilities and fun stuff like TNT.
Take a look at the comments to the osnews.com article. Every frickin' one is incredibly positive and written in the same buzzword-rich, press-release style as the interview itself, extolling the virtues of REBOL. The only non-exuberantly-positive posts are ones asking about negative aspects to which there's no answer.. Is a response like that possible on a fairly moderated messageboard? It looks like the posters are selling the product instead of commenting on their experiences with it...
A sufficient battery backup system with lots of warning indicators when the battery is running low should solve the problem. Maybe it's a bad example, but I remember the solid state memory devices in those old NES cartridges lasting for years and years.
But seriously, it shouldn't be too hard to hack this together. Anyone up for writing the drivers?
I recently finished reading Ray Kurzweil's The Age of Spritual Machines in which he says that humans will start using their computers more and more, with better and better interfaces (speech recognition, then neural implants, etc.) and more and more convincing computer personalities. Slowly but surely, we would merge into the machines and become one as we naturally use the computer to enhance our memory, reasoning and finally thought processes.
Cyc could fit nicely into such a seemingly far-fetched scenario: we start using it to get data or suggestions, and as it grows more and more intelligent and useful, we set up a "direct connection" for even better response time... combined with a proliferation of other AI programs, the blur between biological humans and our computer-based intelligent offspring will blur as to become indistinguishable.
Just a thought: Perhaps, one day all the stuff going on right now will be in the history books. Cyc will be known as one of the great pioneers of the new age of humanity. Now THAT would be cool.
How is this any different from what humans do? We're fed an incredible amount of facts and assumptions about life and the world that we take for granted -- as demonstrated by how hard it is to program common sense into cyc. The reason our thoughts seem "creative" or "spontaneous" is because of the vast, interconnecting and rich set of ideas, facts and experiences we've accumulated and cross-relate, without even thinking about it.
What differentiates The Linux Developers Group from Eazel? Eazel's only product of value was given away for free, as is GnuCash, which from the LDG's website seems to be its only product. I absolutely applaud your efforts for Free Software, but what is your plan to be profitable?
Slashdot Mirror
Why is there such a rush of governments, the U.N., and so many others already pledging support for this initiative? It sounds like it could be a good idea, but shouldn't people wait to see whether the thing actually works? They have a whole boatload of new, innovative features like automatic peer-to-peer wireless, their own specialized Linux software, etc. -- surely it'll take time to work out the kinks...
Just a thought: automated webpage translation services like Babelfish and Google's language tools can translate arbitrary webpages yet preserve images and navigation. This only works because the web works via hyperlinked text documents. But if you want to read a website that uses AJAX for navigation, these systems will break -- they can't intercept your XmlHTTPRequests. And even if there was a browser extension that could do that, it would be unclear what text needed to be translated and when. So here's one example of a useful tool that depends on HTML transparency, broken by more programmatic clientside techniques.
I find CherryPy's URL traversal scheme a bit clunky -- since you connect up objects to each other via attributes, you can't see the hierarchy of your site. At least with PHP you can use "ls" to discover what your URL space looks like. Django uses a really neat scheme that binds a table of named regular expressions to callable handlers, e.g.
(r'^polls/(?P<poll_id>\d+)/$', 'myproject.apps.polls.views.polls.detail')
and the handler is declared as
def detail(request, poll_id)
...
The entire phenomenon of packaging, or even the existence of distros, is odd. The developers of a piece of software do source releases, then have to cross their fingers that developers from other distros will package their work. It would be nice if the original software's developers took the next step and made binary packages, but it's impossible for them to package for every linux distro in existence. Therefore, it's expected that other people are supposed to come along and package it for their own distribution.
This system is unsustainable for a mass-market Linux world. If I want a piece of software, I should be able to go to a website run by the people who make it, and download it from there. Currently, I go to the website, learn about the software, THEN use seperate non-web tools to search my distro's own rpm/deb database to try to find if it already has a distro-sanctioned package. If that fails, I then go back to the software's website and hunt around for the best solution, falling back on a source release if necessary.
This "packagers-are-not-developers" system is especially bad in two situations.
(1) there's so much useful software out there, a distro maintainer can't keep up. This might already be the case. Distro packages are always several versions behind what you can get on upstream's website. Upstream's website does a much better job at telling users what the software is. To keep your distro up to speed, you then have to decentralize its production (e.g. thousands of Debian volunteers) and potentially introduce horrendous organizational/managerial difficulties (e.g. three years between releases.)
And, (2) commercial developers are screwed, because they have to do all the packaging themselves. This applies particularly to closed-source, but also to any software developed by a small or specialized group of people. This problem is solved now because enough Linux users are developers such that popular software gets packaged; but a higher non-dev:developer ratio implies more and more people can't get software they want.
Both (1) and (2) are terrible barriers to mass Linux adoption, because they imply limits on the 3rd party development community when most users aren't developers anymore. Linux world needs to have a unified technical system so it can have a unified social system where a software's developer can make end-user-consumable packages themselves.
UnitedLinux failed... LSB seems weak... maybe we'll only achieve this once some WalmartLinux storms the world and everyone codes for it as a de facto standard. Until then, let's watch those commons stayin' on tragic...
They're running a project like this, to teach computers from volunteers. (A.k.a., collect volunteer-supplied data for various machine learning tasks.) It's Open Mind. Looks like they're not doing natural language grammar per se, but something similar could be used for that.
Grammar's hard, though. Linguists have been working for some 50 years to explicitly write down the rules differentiating what we percieve as good versus bad; and they still haven't succeeded. We still can't enumerate or describe the intuitions we all have in our heads about this.
This sounds like a great idea, provided that the Knoppix can be user-friendly enough to figure out how to boot up.
... reminds you of the Apple II days, where you had to boot half of the operating system off a floppy every time you turned on the computer.
There's really no surefire way to ensure that a user's harddrive-installed OS is secure for banking. Considering the staggering variety of adware/spyware/viruses on machines today, it must be quite easy for a malicious malware creator to make a program that hijacks name resolution (change DNS servers, or the HOSTS file) for perfect phishing, or they could install a keystroke logger, or whatever else. If they got their bank-website-hijacking malware on machines in whatever way all today's adware stuff gets on, they could easily phish thousands of bank transactions every day.
The prevalence of malware seems to indicate that people can't control or trust the programs on their own hard drives. If that's the case, they can't trust any of their online interactions. Since Knoppix kills your harddrive and all its flexibility, it's much more secure.
What would be funny is if more and more institutions started demanding the use of bootable OS's. Our PC's would be reduced to a BIOS, monitor, and keyboard
THe Lookout program was mentioned a bit back on Joel on Software and interestingly enough, it looks like they used the Apache Foundation's open-source Lucene full-text indexer to create a good search plugin for Outlook.
1) Read/write is transparent. The burning step for CDRW is terrible; you should be able to directly open, save, and erase files just like any other drive. Then you don't need to copy files to your harddrive to work on them and then back again when you're done; that eventually invites confusion. The most prevalent network transfer protocols require separate download/upload steps.
2) The media is physically robust. Unlike CD's, a protective case isn't critical for floppies. Floppies do not start flaking out after being scratched a number of times. They're easier to transport and share -- I can put them in a backpack and run around all day without the flimsy plastic case breaking. And the fact you can write on them with a normal pen increases usefulness too: labelling is really helpful for yourself and essential for sharing.
And unlike USB drives, floppies have a standardized size, so you can stack them and store them in standardized cases.
3) The media is cheap, which facilitates sharing. USB drives cost lots of money; to give your data to someone you can't just hand them a spare drive. Floppies, even the older high quality ones, are cheap enough to give away.
With cheap media, you can afford to use a labelled disk as a unit of classification -- you don't need to fill up the disk to get your money's worth. USB drives can't do this (yet).
Expensive drives inside computers paired with cheap disks is much better than expensive combined disk+drives that can be swapped between computers. A good universal physical medium should be usable on all computers; it's not like the act of transferring files is something that only the rare person with a usb stick wants to do. You should only have to have a cheap disk to transfer files; you should not have to invest in a special drive.
To transfer files I once had to go around knocking on doors, looking for someone with a USB drive. This is ridiculous. (I am more likely to have a spare floppy, or only have to go knocking around for a floppy!)
4) Media reading/writing is (was) universal. CD drives are universal, but not always for writing. USB is pretty good now, but it can be a pain to find the plug in the back of the machine; I've also had weird OS hangups on certain systems (esp. older windows). Networks aren't always available in all environments -- especially figuring out which server or transfer protocol to use that will work for your particular situation.
Universality was definitely a bane of Zip drives and other floppy replacements -- a media type is useful only if everyone else has it.
5) They're dead easy to use. The CD burn step and usb issues were mentioned above. Further, network transfers are a pain. I've had the most annoying experiences just figuring out how to network transfer a file from one computer to another. Maybe you can upload/download via ftp -- if you have a server around, and you even know what ftp is? Maybe use email -- which requires extra space in someone's mailbox, and through web interfaces is often even clunkier than ftp? And the login steps are definitely extraneous. Store on a network drive -- if you have a server available nearby? Computers still can't universally detect each other's presence and sling around individual files without depending on some remote server. The easiest and most common way to transfer files I've observed on campus is to have an AOL IM signon on each computer, then use its file transfer mechanism. This is ridiculous. If files still fit on floppies this situation would be so much easier.
Obviously, it's possible to solve the peer-to-peer transfer problem via better and more universal pr
You can see Donald knuth's fvwm desktop and also his fvwm2rc on the bottom of this page
Wrong: it doesn't imply hopelessness, but rather encourages us to take action to change. Do you think that Kofi Annan wants us to throw up our hands and not care about the rate of improvement? No! By recognizing the magnitude of the problem, we can realize how important <i>more</i> improvement is. Just because things are improving doesn't mean we shouldn't be concerned about the huge inequalities that exist.
Um, .deb's do this too -- each has a signature inside. In fact, Debian has an entire organizational system for maintaining GPG keys and signatures.
What websites currently use OpenACS? I couldn't find a list on openacs.org.
Well, the success of an online community is more important than its size. The fact that it can work in either a small place like Blacksburg or a big one like Seattle says something about the entire idea. I'm wondering what elements are common to these different examples. Thanks for the comments.
Many of these sites actually have started with a newspaper site, or later get taken up by a newspaper for funding reasons. It's definitely a good idea, except that sometimes they could need to do things beyond the scope of a newspaper -- like host websites for community groups, provide a forum to criticize media coverage of a particular event, etc.
Good point. One of the questions I'm researching is, can it be successful among poorer communities as well? You then have to deal with digital divide issues at the same time -- very tricky.
No, it's the immature obsession with sex so prevalent among male, adolescent-minded geeky community that's offensive. Only an overwhelmingly male subculture will be so disgusting across the board. While making dirty sex jokes isn't always bad, the fact that EVERYONE here is making them, and laughing at them (all +5 funny), the environment becomes intimidating and hostile towards women. The all-male reinforces the sexism of the computer science world.
What's an anonymous coward doing calling me a moron? And you certainly have _your_ facts wrong...
1) There are huge differences between the source code for each. Did you try running a diff? IT's huge! The GUI was changed, etc.
2) There is no zlib directory. THere is one in the Gnucleus source, but _not_ Morpheus's. In fact, there are no subdirectories at all.
Sigh.
(I just had to post this again...)
.ico's or .bmp's.... No explanation how or where to get them, no instructions how to make the thing go. If all the "Source Code" link on a website does is give you raw, instruction-less code without data and descriptor files critical to actually _building_ something, the GPL hasn't been fulfilled.
$ grep 'res\\' *|wc -l
grep: Debug: Is a directory
75
Unfortunately, there is no 'res' directory for these 75 references to it. The project descriptor files refer to dozens of files THAT DO NOT EXIST.
There are no makefiles, there are no
For the sake of everyone, Streamcast, get this together!
..is AllegroHack, for Windows and Linux. It's really sweet, and IMHO more complete than Falcon's Eye (the last time I played it a monkey's picuture was that of a guard!)
Also, Allegrohack works with SLASH'EM a really fun version of nethack (if a wee bit unbalanced) where you get more abilities and fun stuff like TNT.
Take a look at the comments to the osnews.com article. Every frickin' one is incredibly positive and written in the same buzzword-rich, press-release style as the interview itself, extolling the virtues of REBOL. The only non-exuberantly-positive posts are ones asking about negative aspects to which there's no answer.. Is a response like that possible on a fairly moderated messageboard? It looks like the posters are selling the product instead of commenting on their experiences with it...
A sufficient battery backup system with lots of warning indicators when the battery is running low should solve the problem. Maybe it's a bad example, but I remember the solid state memory devices in those old NES cartridges lasting for years and years.
But seriously, it shouldn't be too hard to hack this together. Anyone up for writing the drivers?
I recently finished reading Ray Kurzweil's The Age of Spritual Machines in which he says that humans will start using their computers more and more, with better and better interfaces (speech recognition, then neural implants, etc.) and more and more convincing computer personalities. Slowly but surely, we would merge into the machines and become one as we naturally use the computer to enhance our memory, reasoning and finally thought processes. Cyc could fit nicely into such a seemingly far-fetched scenario: we start using it to get data or suggestions, and as it grows more and more intelligent and useful, we set up a "direct connection" for even better response time... combined with a proliferation of other AI programs, the blur between biological humans and our computer-based intelligent offspring will blur as to become indistinguishable. Just a thought: Perhaps, one day all the stuff going on right now will be in the history books. Cyc will be known as one of the great pioneers of the new age of humanity. Now THAT would be cool.
How is this any different from what humans do? We're fed an incredible amount of facts and assumptions about life and the world that we take for granted -- as demonstrated by how hard it is to program common sense into cyc. The reason our thoughts seem "creative" or "spontaneous" is because of the vast, interconnecting and rich set of ideas, facts and experiences we've accumulated and cross-relate, without even thinking about it.
What differentiates The Linux Developers Group from Eazel? Eazel's only product of value was given away for free, as is GnuCash, which from the LDG's website seems to be its only product. I absolutely applaud your efforts for Free Software, but what is your plan to be profitable?