Now SGI just has to port all its realtime video compositing and reality simulation to high-density linux on Blue Gene (didn't they port Performer to linux?) and we'll have three gods of Open Hardware as horsemen to start this century off with a bang!
I'm just scared M$ will get a clue somewhere along the line (they may be evil but surely the richest man in the world ain't dumb), but so far they play a good Darth Vader to our rebel army. It looks like we've finally got momentum, now turn on that exponential metric thing! Now we need a company like Eazel to polish it up, with a business plan.
Is it just me or does the Haskell one seem to have inverted parameters wrt the vanilla Perl one? Not that I know Haskell.. I was hoping to see some lexical wierdness in Perl and some lambdas and arrows all hanging out for the Haskell one. And wha-hey no smalltalk? well the world goes on..
Neat idea, I like to see how things are done in other languages. It would be an interesting and intuitive way to show people what other languages are like, a list of programs the same in each language, plus commentary about the highlights of the language that are used.
Just because the RIAA says something doesn't mean it's right. There just isn't anyone with as much clout arguing for relaxation of restrictions, or reduction of prices. However libraries and rental shops have for a long time rented audio/video media.
Libraries buy books and other items for lending, why couldn't a respectable outfit try to do this digitally? Are libraries illegal now?
Certainly there are differences between books and digital media, but there still remains a lot of unexploited potential between Fair Use and the Tradition of Libraries. There are even things called interlibrary loans. And there is value in promoting research and telling people about relevant authors which librarians also do. I'm curious about whether there would be any problem with software client that would let a repository track files, and provide the services that libraries generally do. This service would cover print media, audio and video. It might be free, for profit, nonprofit, tax-supported, members only, a coop, or something else. Hopefully it would be international.
The library would certainly have to pay for its own copies, and it would have to handle only items for which electronic distribution is allowed. But it is also conceivable that through fair use people could register their own music with the library and sign something which says they will not play a given song when it is being played by someone else. Some cryptography may be appropriate, but this is not really so important (except for some signature to identify each file uniquely).
It may sound like there are lots of loopholes for copying but this is true in many other realms. The public does not bear the burden of inventing ways to enable corporations to infringe on their rights. I am not interested in promoting illegal copying but the kinds of money the RIAA talks about are not part of reality as we know it. I believe that if we want to own our own past and future, we must take steps to do so ourselves. This means discussing the subject with professional librarians, publishers, broadcasters, authors and artists. Then bring on the developers and lastly the business people. It doesn't have to make a billion dollars but it should be a useful service for people of all ages.
As some people here suggest, UserLinux probably requires some rethinking and a lot of work to get to a system that is more useful than Windows for the general public. I think though that a lot of exciting things can be done along this path which also make UserLinux the most interesting distro for developers, scientists and other tech-savvy people.
organize development objectives and set priorities via a single collaboration site so that individual developers can lay some of the bricks in this warm, well-designed, efficient and delightful to live in building.
a GUI unification layer which will attempt to transform the chaos of the linux desktop into something like the ease of the MacOS desktop.
AI services such as getting some computational linguist nerds together and building desktop computing domain-specific tools.
The goal of AI services would be that linux would understand itself and how all the programs and hardware act, can do some reasoning and enable things like automatic configuration, automatic software search/installation/interworking, agent-assisted development and solution of user-initiated goals, development of personal assistants, and in general making linux something a magnitude beyond other operating systems. In fact parsing and syntactic recognition of written and spoken natural language is here today, but for now still requires experts to spend time creating a domain-specific system.
I have a feeling that the three projects above taken in order could launch linux and open source in general into a totally different strata compared to other software. Just setting up a single website as a clearinghouse for gathering and organizing ideas would be a major step. With such a tool then many people's work could be coordinated within a common map, and the next step (rebuilding the linux user experience, through GUI unification and also using the smarts of experts in this field) is possible.
At the same time a serious effort (step 3) can be made to acheive a linux desktop-centric set of complementary AI services. Here are some goals I can see might make it to the front page:
- We need a place where we can get experts in AI to provide insight. A little goes a long way.
- We need to build a knowledge representation of the world of user computing, from file copying to driver installation problems to word processing. This is one of the resources that will enable machine reasoning to make complicated computers easy to use.
- We need to provide a software framework for various artificial intelligence and natural language processing software tools to plug into.
- We need to find a way to introduce large numbers of non-linux-nerds (designers, technical writers, psychologists, PhDs, manufacturers) into the dialogue for concrete projects. For example a widget manufacturer could provide funding to a small group that would create entries in the knowledge base and so on for their product if it would guarantee plug-and-play compatibility with the reasoning software.
- We need a pilot project which will help solidify things we need to have done, for example we could try to incorporate NLP into the shell. A project oriented to the general public might be an onscreen "notebook". Maybe we should be using Squeak? Think about how we can get rid of the infinite number of control panels and buttons, and move toward an interface that would make it fun or easy to use.
- We need to create modular scripts that solve certain user cases and yet can be applied to many different situations. Open source desktop intelligence.
- We need to provide ways allow reasoning engines to be able to do things for the user.
It may sound like this is "silly science" but this is not at all impossible. First, the minimum technology is there already, and what isn't. Second,
Felt good scanning the news for the first time in a while!
Usually I have to lurch past interminable murders and battlefield pics to get to some maybe-already-read science story at the bottom of the page (on cnn).
But with this it was easy. I clicked on Top Stories more.. and skip the photos which I don't know what they are. Right away I see my two choices, what seems like a gorgeous tanned piece of royalty in a crimson and silver dress, or a stressed out techie on the phone. Hmmm, which should I pick? It's over in a microsecond and obviously everyone else here is making the same decision since the story (Halle on her Disastrous Love Life) is slashdotted. But the theory works. I don't know who the heck Halle is but now I want to know and save her from a bad boyfriend too!
I would even go for fewer thumbnails about 5 times the size of these and scrap the ones with bad pictures. That way we could see the news before it gets slashdotted. Next we'll evolve to networked torrents of femmes fatales (girls you pick hommes fatals or whatever you like). It is so much easier to make a decision without all those pesky letters they give me so much eyestrain anyway.
Sure in Japan Love Getty has been around for a while, and phones all have email or cameras now. But I had a bunch of trouble finding cheap terminals for computer-assisted meetings. If you have a laptop for each participant that's fine but it would have been too much money for a 3000 person meeting for which I proposed CAM a year ago.
I couldn't find any information about the hardware on the site, but what I really want is a very cheap tag that can be used by the wearer to respond to multiple choice questions asked by a speaker. Presumably you could do this by using a button to activate one of a few printed RFID elements at a minimum.
The other use cases I wanted to support are inputting free text answers to questions, and also as a way of meeting people who can answer your questions one-on-one at a conference. The nTag seems to be close but too expensive as is, it would be more worth it if they had a membrane keyboard in there.
I'm sorry, even Microsoft apologists can go back and forth about capitalism until the sun goes down but the reality is that no government that is scrounging for $18/month per family for food rations has any business paying a U.S. monopoly 25 million dollars a year. THAT would be plain irresponsible.
I would like to have heard more about how using linux would help accelerate education, technology development, and communication. Or about how it is superior to Windows in many ways. Or about what open source really means, or about how governments have certain obligations which can be best met with open source.
But the clanging, steel hard bottom of the pot truth is, Brazil and most of the states considering linux are absolutely correct to FUCK Micro$oft and their double-dealing ways. It just so happens that South Americans seem to have bigger cojones AND clearer heads about this, but most likely every local or national government in this economy would do better to steer away from megacorporations and spend less money on developing maintainable systems of their own which leverage the work of other states as well.
Of course it will cost money, but on the order of the first $20 which after passed through the economy hundreds of times has created an exponential amount of wealth. This will also create jobs! THERE IS NO REASONABLE ARGUMENT FOR BRAZIL OR ANY OTHER GOVERNMENT TO PAY THE RICHEST MAN IN THE WORLD. So praise Brazil and Peru, and do your best to get people who understand what this is about - MONEY, JOBS, EFFICIENCY and FREEDOM FROM CUTTHROAT MONOPOLIES - into office where they can make similar decisions.
..if they include all encompassing airbags so when you hit a rock and go flying, you can bounce and tumble across the landscape like the Mars rover.. and walk away.
I think the title of the post is more interesting than the unconvincing article. With natural language recognition people might be able to use computers more efficiently and powerfully if they could explain things to it in an English-like language. I am interested in natural language recognition and have been having some trouble finding anything useable (though I have positively scoured everything I can find from darpa and nlp sites).
The issue of being controlled by "the windows" is just voicing a simple truth of media as recognized by many media theorists. However the issue of being controlled by corporate entities through subtle manipulation of the fundamental interfaces and protocols of our information infrastructure (including popular software) is a big one that was barely suggested.
There are plenty of examples in fiction and it would seem also in the real world (maybe in research, and defense) of natural language or pseudo language being used. However in the open source world it is very hard to find command line tools which enable the technical user or developer to easily incorporate elements of artificial intelligence in one's daily work. There are a number of projects in c/c++, java, and perl (I personally want to focus on perl and eventually perl6 with c/c++ extensions) in natural language processing, knowledge engineering, and artificial intelligence but it is not something you can find in any linux distro yet. I'd love to hear if anybody has good information in this area.
1. The parts of linux which are "not ready for the desktop" are the things needed on the enterprise desktop too. Things like cut and paste, and printing to name a couple. If RedHat does not approach these issues its enterprise desktops will not be polished in the near future.
2. RedHat is not big enough to take on Microsoft on its own, as plenty of companies have found. Its latest announcements have given weapons to Microsoft which can be used on decision makers and what's more, has alienated the community of individuals and companies which have created nearly all of RedHat's products.
3. There currently are in fact a lot of people using linux on the desktop, and RedHat seems to forget that ALL current RedHat customers started by using the RedHat desktop!! Does RedHat imagine they can throw ALL of their current customers into the trash and start from scratch?
4. Mac OS X is unix-based, unlike Microsoft Windows. It would make far more sense for RedHat to recommend that home users use Macs, which are easier to use, more advanced, and could also run unix applications, than to recommend Windows, which you can only trust to further muddy the waters of incompatibility. The less people use Windows in any sense, the better for RedHat.
5. RedHat has now officially alienated its customers, the producers of most of its products, the competitors which up to now had been in league with RH against Microsoft, and everybody smart enough or interested enough to recommend RedHat desktop software and use it. If they thought it would be ready in a couple years, they would not be dropping it now. After all the recent announcements about product life cycles and not selling desktop software, and now this total backstabbing, I am losing interest in RedHat and see no reason to recommend their enterprise line anymore though I would have done so in an instant a couple months ago. RedHat has forgotten how it got started and why they have gotten this far. Even in large companies and governments, the decision makers need to be on their side and this guy is RedHat's own worst enemy.
6. The linux desktop has made many strides recently and will continue to get better, no thanks to RedHat. Possibly this and the Novell news will help other distros become more unified and give some more impetus to fixing the desktop. But this is mainly going to happen because people want to invest their time/money in making it happen and RedHat's lack of interest in such development indicates it is not interested in supporting linux's future. Possibly Fedora will take off, but something tells me it is not going to be because of this suit. U.S. corporate culture has breeded a whole generation of smiling, heartless executives who imagine that getting the numbers right will spell success. I feel ashamed.
7. Slashdot, arguably one of RedHat's best free PR outlets, has turned from RedHat lovers to FSCK RedHat (that's FUCK REDHAT! for nontechies). RedHat can kiss their allies goodbye. FUCK EM!
This is the last straw. I was already having trouble getting pushed into upgrades of turnkey systems I had based on RedHat, due to their discovery of the ignominous phrase "end of life cycle".
I switched 1-2 years ago to RedHat from SuSE due mainly to Japanese support, for my laptop. But also RedHat has taken over in Japan. And I recommend/install RedHat boxes all the time. Recently I upgraded one from pre-7 to RH9. This would be a prime client for an enterprise liscense perhaps, and I recommended joining the RedHat program. I have done so for other clients as well.
But here's the thing. I don't want to pay RedHat for my development machines, but I want the same OS and directory layout as the for-sale servers just to keep sane. I also want easy security updates for them. But RedHat keeps making it more difficult to answer my clients' questions about how safe RedHat is (yes lots of companies still only really trust Sun for servers) and this is just another thing they can point to.
If I can have a system that handles Japanese okay, runs a desktop and server edition, and has a for-pay security update service I'm there. I was just about to purchase a linux server to host a new domain and figured of course RedHat, but the reality of it is I don't need anything more than the standard GPL stuff that is in all the distros. It's not like they invented everything in it themselves right? Why should they get money for the kernel, perl, apache, sendmail, iptables? It's not like RedHat has made my life much easier except with regard to providing signed binary builds, and that is the only thing I want to pay for now.
Another thing: I am willing to pay RedHat for security patches, but the shortness of their life cycles is so horribly cynical I want to scream. Do they think they have got enough critical mass to take over the world and drag us all along with them? I think they are shooting themselves in the foot. And I am now for the first time since my conversion to RedHat, considering something - anything - else! I'll vote with my money and the money of my clients. Hope you do too.
By the way I recently installed two IBM xSeries turnkey servers for a client. The servers came with RH7 which is all IBM would guarantee, but I tested and put RH9 in - only because of security updates - that darned "life cycle" they invented. I recommended they purchase RHN. Now tell me RedHat, because I trusted you am I going to have to tell them to purchase a fat liscense for you some day in the future and then hire me to migrate them? What in fact will IBM or other companies do? I could live with RH9 on those machines because I wasn't using the software that broke with RH9, and I decided I didn't need IBM's patch on RH7 needed to make the autoreboot feature work. Do these kind of issues make RedHat sound like the kind of a company for which you are willing to put your name on the line? This is utter bullshit, and the funny thing is if they just provided a simple service which kept those servers securely updated for me I would not hesitate to tell all my clients to pay TEN OR TWENTY TIMES what RHN would have cost them. RedHat is going to discover that someone else is going to do it the right way and leverage the off-hours work of RedHat's employees in a very lucrative way. That's where I hope this goes. Or I may just change distros; gentoo, slack, and even FreeBSD and Debian are looking mighty good right now.
It would be useful to be able to keyword search the captions of all these cartoons so you could find appropriate ones for any given situation. Perhaps someone would provide a reference on line? Otherwise it seems utterly impossible to use unless you set up a whole learning infrastructure along the lines of Bible/Torah studies and have scholars dedicate their lives to studying it. Arguably profitable but..
This is great but it took a long time to figure out the rules since the description is so horrible. Well I got this far so good luck!
For 104015 as drawn here is the list of decisions. Basically it seems that every time the worm comes across a point with an eaten segment attached to it, the worm takes the next rule as a new decision, then draws a segment, and then goes back to the first rule in the list (which makes it try to start drawing a hexagon again). However I have to say that Pegg's coloring is rather arbitrary, that lone green segment in the middle stumped me for a while! Also the way decisions are made is not intuitive.
111111 black hexagon Hit hexagon starting point, pick rule #2 which is a "0". 0 go straight one segment in red. Go back to start of program, now use rule #1: 111 draw three more red segments using that rule. Hit the black hexagon again, pick rule #3 (a "4") unclear yet whether it is picked because it is next in order or because the rules #1 and #2 can't be used. But considering next decision, I believe it is because it is next in order: 4 turn two hours to left and draw first blue segment. Then go back to beginning of program. 1111 draw four more blue segments until black hexagon is hit again. Hit hexagon, pick rule #4 (a "0"). 0 draw first yellow segment straight ahead. Then go back to beginning of program 1 draw second yellow segment. Hit a vertex of the black hexagon, but we can crash through for some reason. So we keep drawing segments with rule #1 which should be in yellow but is green in the drawing: 11 draw two more segments Hit the blue hexagon so pick rule #6 (a "5"). 5 turn two hours to left and draw another green segment. Then go back to beginning of program with rule #1: 1111 draw more green segments Hit blue hexagon again. Rule #1 doesn't do the job so rule #2 (a "0"): 0 go straight ahead drawing a single green segment. Then back to beginning of program with rule #1: 1 turn right and draw a single green segment. Hit black hexagon, so pick next rule. All the rules are useless but the last rule #6 (a "5"): 5 turn left and draw a single green segment. Then go back to beginning of program and rule #1: 1 turn right and draw a single green segment. Now we hit the black and red hexagons so another decision is to be made and we are into magenta which I won't follow.
It still seems wierd how decisions are made and maybe following magenta for yourself will tell you how the recursion is supposed to be done. Personally I would rather see some pseudo code since it is difficult to tell which of the two "0" rules is being chosen.
It could be a short, but it would make more sense (in that I doubt anybody who does shorts is going to be so sure the GPL will win) if Microsoft is behind them. Especially if GPL code is in Windows. This would allow M$ to die in a truly spectacular fashion. Dream on..
Actually, no. I have been making some ad hoc inquiries into this matter with the help of the company microwave, a bunch of free pasta, and a mighty hunger.
Now I am not saying you should microwave water by itself in an unmoving tray or even a moving one for that matter, but (though I think somewhat inferior to ordinary pasta cooking) it works fine if you put cook pasta as I have several times in a plastic microwave safe deep tray with about equal parts water and spaghetti. The starch that initially comes off the pasta perhaps stabilizes it, anyway about 6 minutes at 500 watts on a rotating base and you have lots of small bubbles perking vigorously though I have not had the courage to attempt a rolling boil with a lot of water. At first it was a little scary and you have to wonder what the temperature is going to be when you stop it..
The presentation and report were very interesting. I kept wondering why the BeOS was not covered and what differences there might be. I seem to remember the BeOS had a completely object-oriented windowing system, and its messaging system while not designed for the net, actually worked when someone tried to pack up a bmessage, email it, and unpack it on his computer (to do a drag and drop across the net).
I was surprised but believed his analysis of the impossibility of using C++ as the basic language, but maybe BeOS was like Y, really C with a C++ api.
Actually I was also thinking that I would not like to have the miniscule memory on my laptop being chewed up by buffers but..
If a new integrated windowing system is being designed though I wonder what other things could be done at the same time. For example, could a module be added to Y which subscribed to all events, listened to all data, and was able to index and learn about your activities. It would seem that an object oriented windowing system would be a big ally in this. Also, if a system providing live queries like on the BeOS was implemented, it could answer queries based on the content of open applications, not just static files. Of course I seem to remember IRIX would let you browse windows as files too..
Well thanks to slashdot I got my brain back in gear on this question after several years. I am pretty sure that the thing which looks like a rock is in fact a lithop, which is a type of succulent from South Africa often called a living stone, of the the plant family Mesembryanthemaceae (now called Aizoaceae) or "Mesembs" for short (google that and go nuts!).
Specifically it must have beenL. olivacea which I guess means olive colored, since as in the photo it had no markings, it just looked like a beautiful hunk of chalky, greenish colored velvety living stone. Can't believe I found it. Some really bizarre, ugly, and beautiful pics on this page. Also more interesting photos here>/a> and here.
I also am thinking of throwing out the pencil plant (Euphorbia tirucalli) stem which will certainly take root by itself, but apparently causes cancer! I wouldn't want a cat to eat it.
This was really interesting so I googled. Cavet: IANA Paleobotanist.
Apparently ginkos are also extremely old and resemeble a Jurassic variety. And Cycads, which are woody plants that create seeds. They also seem to be quite poisonous although they are eaten as "beach tucker" after processing in the jungle. (link)
Anyway here are some links.
Finally I there are also the extremely visually (and biochemically?) wierd Gymnopsperms like Welwitschia And Ephedra, which seem ancient, maybe same era..
All this because I was trying to figure out if the inch-long stem/leaf in my pocket which I snapped off a huge pencil plant was one of those. Not sure yet.. I remember my mother also has some kind of ancient plant which looks like a gray rock and does nothing, but then one day suddenly splits in half, and then each half will continue to split in the same way recursively. A very cool plant if anyone can figure out what it is!
If you know Mac OSX you might apply to a biotech firm, a design company, a film company, an architecture firm(well maybe), an advertising firm (the design side), or an institution doing simulations if they have a Mac OSX cluster like that mentioned recently. But yeah I know what you're saying.
According to one recent/. article (terascale storage), systems and database administrators will be increasingly in need as storage increases, and also increasingly automated, to solve that problem. But if you focus on the applications/integration side, or if you can find a neat job at a lab you may ride a wave.. Good luck.
John Walker is also the author of a fascinating set of free programs and book on dieting called the Hacker's Diet in which he discusses a realistic way to lose weight while remaining healthy, using techniques of project management which have served him well in other fields of endeavor. Check it out! (I have the Palm version running myself and have lost 6kg)
Would taint-checking addresses in a CGI program cover this hole in unpatched sendmails?
Certainly you want to patch your own machine, and get the admins you know to fix theirs. But I am thinking now about people running on a virtual hosting server which is perhaps getting most hits from Perl CGI program, and which are using sendmail only internally, for example to send email to automatically registered users. Would it be sufficient to (as usual) check length of address and remove nulls?
I got hold of an IMSAI when I was younger, it went well with my apple II and teletype.. anyway boy did it have LEDs. You had to enter programs into it by flipping toggle switches, each LED was a bit.
Unfortunately I got talked into my folks giving it to a guy when I was away who said he was making something for handicapped, which I don't think every materialized. That was dumb!
I wouldn't mind having some LEDs to show what each key on the keyboard was, it would probably be relaxing. Maybe a software version is not that hard but not sure it would be as pleasing.
I don't know about you but I just got a major urge to buy a few tins of thinkgeek's smartmass, and send some of them to members of my family.
I tried to make some kind of silly putty when I had a pretty good chemistry set for a kid, like 25 years ago. (a store-bought totally un-understandable chemistry kit and a ton of antique chemistry glassware my parents found in shops!).
Alas the putty (which was probably more the borax type I guess) was extremely hard to make and it seemed very sensitive to the ratio of ingredients.. I do remember that washing with I guess alcohol maybe, I ended up with a stringy white mass. After a couple tries I got something like putty but kinda nasty..
Anyway my question to offset these sickening legal posts. Do you think the stuff would conduct electricity, or be susceptible to heat from wires? I would guess no on electricity but yes on heat.. anyway wouldn't this be kind of cool material for an embedded linux project (if it is not conducting)?
You could wrap some circuit boards and leds together and bind the whole thing in a mass of putty (maybe thinkgeek's glowing green one, though the blue, white, and black ones are intriguing and the copper one looks good wrapped around buzz' leg..)
presumably the green one would afford a persistence of vision kind of effect like tv phosphors that glow for a long time after being stimulated. I'm a little worried about whether heat or time would melt the stuff, or could it cause a fire.
Could putty be our best friend? I mean you could drop the stuff and even if the putty shatters, that's energy that went into breaking putty bonds and not your hardware! I wouldn't mind buying a 2 1/2 inch hard drive encased in a globe of black smartmass putty, with a black USB wire coming out the back (or wireless!). Well we can dream anyway.
Slashdot Mirror
So Sun has found their niche!
Now SGI just has to port all its realtime video compositing and reality simulation to high-density linux on Blue Gene (didn't they port Performer to linux?) and we'll have three gods of Open Hardware as horsemen to start this century off with a bang!
I'm just scared M$ will get a clue somewhere along the line (they may be evil but surely the richest man in the world ain't dumb), but so far they play a good Darth Vader to our rebel army. It looks like we've finally got momentum, now turn on that exponential metric thing!
Now we need a company like Eazel to polish it up, with a business plan.
Is it just me or does the Haskell one seem to have inverted parameters wrt the vanilla Perl one? Not that I know Haskell.. I was hoping to see some lexical wierdness in Perl and some lambdas and arrows all hanging out for the Haskell one. And wha-hey no smalltalk? well the world goes on..
Neat idea, I like to see how things are done in other languages. It would be an interesting and intuitive way to show people what other languages are like, a list of programs the same in each language, plus commentary about the highlights of the language that are used.
Just because the RIAA says something doesn't mean it's right. There just isn't anyone with as much clout arguing for relaxation of restrictions, or reduction of prices. However libraries and rental shops have for a long time rented audio/video media.
Libraries buy books and other items for lending, why couldn't a respectable outfit try to do this digitally? Are libraries illegal now?
Certainly there are differences between books and digital media, but there still remains a lot of unexploited potential between Fair Use and the Tradition of Libraries. There are even things called interlibrary loans. And there is value in promoting research and telling people about relevant authors which librarians also do. I'm curious about whether there would be any problem with software client that would let a repository track files, and provide the services that libraries generally do. This service would cover print media, audio and video. It might be free, for profit, nonprofit, tax-supported, members only, a coop, or something else. Hopefully it would be international.
The library would certainly have to pay for its own copies, and it would have to handle only items for which electronic distribution is allowed. But it is also conceivable that through fair use people could register their own music with the library and sign something which says they will not play a given song when it is being played by someone else. Some cryptography may be appropriate, but this is not really so important (except for some signature to identify each file uniquely).
It may sound like there are lots of loopholes for copying but this is true in many other realms. The public does not bear the burden of inventing ways to enable corporations to infringe on their rights. I am not interested in promoting illegal copying but the kinds of money the RIAA talks about are not part of reality as we know it. I believe that if we want to own our own past and future, we must take steps to do so ourselves. This means discussing the subject with professional librarians, publishers, broadcasters, authors and artists. Then bring on the developers and lastly the business people. It doesn't have to make a billion dollars but it should be a useful service for people of all ages.
The goal of AI services would be that linux would understand itself and how all the programs and hardware act, can do some reasoning and enable things like automatic configuration, automatic software search/installation/interworking, agent-assisted development and solution of user-initiated goals, development of personal assistants, and in general making linux something a magnitude beyond other operating systems. In fact parsing and syntactic recognition of written and spoken natural language is here today, but for now still requires experts to spend time creating a domain-specific system.
I have a feeling that the three projects above taken in order could launch linux and open source in general into a totally different strata compared to other software. Just setting up a single website as a clearinghouse for gathering and organizing ideas would be a major step. With such a tool then many people's work could be coordinated within a common map, and the next step (rebuilding the linux user experience, through GUI unification and also using the smarts of experts in this field) is possible.
At the same time a serious effort (step 3) can be made to acheive a linux desktop-centric set of complementary AI services. Here are some goals I can see might make it to the front page:
- We need a place where we can get experts in AI to provide insight. A little goes a long way.
- We need to build a knowledge representation of the world of user computing, from file copying to driver installation problems to word processing. This is one of the resources that will enable machine reasoning to make complicated computers easy to use.
- We need to provide a software framework for various artificial intelligence and natural language processing software tools to plug into.
- We need to find a way to introduce large numbers of non-linux-nerds (designers, technical writers, psychologists, PhDs, manufacturers) into the dialogue for concrete projects. For example a widget manufacturer could provide funding to a small group that would create entries in the knowledge base and so on for their product if it would guarantee plug-and-play compatibility with the reasoning software. - We need a pilot project which will help solidify things we need to have done, for example we could try to incorporate NLP into the shell. A project oriented to the general public might be an onscreen "notebook". Maybe we should be using Squeak? Think about how we can get rid of the infinite number of control panels and buttons, and move toward an interface that would make it fun or easy to use.
- We need to create modular scripts that solve certain user cases and yet can be applied to many different situations. Open source desktop intelligence.
- We need to provide ways allow reasoning engines to be able to do things for the user.
It may sound like this is "silly science" but this is not at all impossible. First, the minimum technology is there already, and what isn't. Second,
Usually I have to lurch past interminable murders and battlefield pics to get to some maybe-already-read science story at the bottom of the page (on cnn).
But with this it was easy. I clicked on Top Stories more.. and skip the photos which I don't know what they are. Right away I see my two choices, what seems like a gorgeous tanned piece of royalty in a crimson and silver dress, or a stressed out techie on the phone. Hmmm, which should I pick? It's over in a microsecond and obviously everyone else here is making the same decision since the story (Halle on her Disastrous Love Life) is slashdotted. But the theory works. I don't know who the heck Halle is but now I want to know and save her from a bad boyfriend too!
I would even go for fewer thumbnails about 5 times the size of these and scrap the ones with bad pictures. That way we could see the news before it gets slashdotted. Next we'll evolve to networked torrents of femmes fatales (girls you pick hommes fatals or whatever you like). It is so much easier to make a decision without all those pesky letters they give me so much eyestrain anyway.
I couldn't find any information about the hardware on the site, but what I really want is a very cheap tag that can be used by the wearer to respond to multiple choice questions asked by a speaker. Presumably you could do this by using a button to activate one of a few printed RFID elements at a minimum.
The other use cases I wanted to support are inputting free text answers to questions, and also as a way of meeting people who can answer your questions one-on-one at a conference. The nTag seems to be close but too expensive as is, it would be more worth it if they had a membrane keyboard in there.
I would like to have heard more about how using linux would help accelerate education, technology development, and communication. Or about how it is superior to Windows in many ways. Or about what open source really means, or about how governments have certain obligations which can be best met with open source.
But the clanging, steel hard bottom of the pot truth is, Brazil and most of the states considering linux are absolutely correct to FUCK Micro$oft and their double-dealing ways. It just so happens that South Americans seem to have bigger cojones AND clearer heads about this, but most likely every local or national government in this economy would do better to steer away from megacorporations and spend less money on developing maintainable systems of their own which leverage the work of other states as well.
Of course it will cost money, but on the order of the first $20 which after passed through the economy hundreds of times has created an exponential amount of wealth. This will also create jobs! THERE IS NO REASONABLE ARGUMENT FOR BRAZIL OR ANY OTHER GOVERNMENT TO PAY THE RICHEST MAN IN THE WORLD. So praise Brazil and Peru, and do your best to get people who understand what this is about - MONEY, JOBS, EFFICIENCY and FREEDOM FROM CUTTHROAT MONOPOLIES - into office where they can make similar decisions.
..if they include all encompassing airbags so when you hit a rock and go flying, you can bounce and tumble across the landscape like the Mars rover.. and walk away.
I think the title of the post is more interesting than the unconvincing article. With natural language recognition people might be able to use computers more efficiently and powerfully if they could explain things to it in an English-like language. I am interested in natural language recognition and have been having some trouble finding anything useable (though I have positively scoured everything I can find from darpa and nlp sites).
The issue of being controlled by "the windows" is just voicing a simple truth of media as recognized by many media theorists. However the issue of being controlled by corporate entities through subtle manipulation of the fundamental interfaces and protocols of our information infrastructure (including popular software) is a big one that was barely suggested.
There are plenty of examples in fiction and it would seem also in the real world (maybe in research, and defense) of natural language or pseudo language being used. However in the open source world it is very hard to find command line tools which enable the technical user or developer to easily incorporate elements of artificial intelligence in one's daily work. There are a number of projects in c/c++, java, and perl (I personally want to focus on perl and eventually perl6 with c/c++ extensions) in natural language processing, knowledge engineering, and artificial intelligence but it is not something you can find in any linux distro yet. I'd love to hear if anybody has good information in this area.
This guy ignores these realities:
1. The parts of linux which are "not ready for the desktop" are the things needed on the enterprise desktop too. Things like cut and paste, and printing to name a couple. If RedHat does not approach these issues its enterprise desktops will not be polished in the near future.
2. RedHat is not big enough to take on Microsoft on its own, as plenty of companies have found. Its latest announcements have given weapons to Microsoft which can be used on decision makers and what's more, has alienated the community of individuals and companies which have created nearly all of RedHat's products.
3. There currently are in fact a lot of people using linux on the desktop, and RedHat seems to forget that ALL current RedHat customers started by using the RedHat desktop!! Does RedHat imagine they can throw ALL of their current customers into the trash and start from scratch?
4. Mac OS X is unix-based, unlike Microsoft Windows. It would make far more sense for RedHat to recommend that home users use Macs, which are easier to use, more advanced, and could also run unix applications, than to recommend Windows, which you can only trust to further muddy the waters of incompatibility. The less people use Windows in any sense, the better for RedHat.
5. RedHat has now officially alienated its customers, the producers of most of its products, the competitors which up to now had been in league with RH against Microsoft, and everybody smart enough or interested enough to recommend RedHat desktop software and use it. If they thought it would be ready in a couple years, they would not be dropping it now. After all the recent announcements about product life cycles and not selling desktop software, and now this total backstabbing, I am losing interest in RedHat and see no reason to recommend their enterprise line anymore though I would have done so in an instant a couple months ago. RedHat has forgotten how it got started and why they have gotten this far. Even in large companies and governments, the decision makers need to be on their side and this guy is RedHat's own worst enemy.
6. The linux desktop has made many strides recently and will continue to get better, no thanks to RedHat. Possibly this and the Novell news will help other distros become more unified and give some more impetus to fixing the desktop. But this is mainly going to happen because people want to invest their time/money in making it happen and RedHat's lack of interest in such development indicates it is not interested in supporting linux's future. Possibly Fedora will take off, but something tells me it is not going to be because of this suit. U.S. corporate culture has breeded a whole generation of smiling, heartless executives who imagine that getting the numbers right will spell success. I feel ashamed.
7. Slashdot, arguably one of RedHat's best free PR outlets, has turned from RedHat lovers to FSCK RedHat (that's FUCK REDHAT! for nontechies). RedHat can kiss their allies goodbye. FUCK EM!
I switched 1-2 years ago to RedHat from SuSE due mainly to Japanese support, for my laptop. But also RedHat has taken over in Japan. And I recommend/install RedHat boxes all the time. Recently I upgraded one from pre-7 to RH9. This would be a prime client for an enterprise liscense perhaps, and I recommended joining the RedHat program. I have done so for other clients as well.
But here's the thing. I don't want to pay RedHat for my development machines, but I want the same OS and directory layout as the for-sale servers just to keep sane. I also want easy security updates for them. But RedHat keeps making it more difficult to answer my clients' questions about how safe RedHat is (yes lots of companies still only really trust Sun for servers) and this is just another thing they can point to.
If I can have a system that handles Japanese okay, runs a desktop and server edition, and has a for-pay security update service I'm there. I was just about to purchase a linux server to host a new domain and figured of course RedHat, but the reality of it is I don't need anything more than the standard GPL stuff that is in all the distros. It's not like they invented everything in it themselves right? Why should they get money for the kernel, perl, apache, sendmail, iptables? It's not like RedHat has made my life much easier except with regard to providing signed binary builds, and that is the only thing I want to pay for now.
Another thing: I am willing to pay RedHat for security patches, but the shortness of their life cycles is so horribly cynical I want to scream. Do they think they have got enough critical mass to take over the world and drag us all along with them? I think they are shooting themselves in the foot. And I am now for the first time since my conversion to RedHat, considering something - anything - else! I'll vote with my money and the money of my clients. Hope you do too.
By the way I recently installed two IBM xSeries turnkey servers for a client. The servers came with RH7 which is all IBM would guarantee, but I tested and put RH9 in - only because of security updates - that darned "life cycle" they invented. I recommended they purchase RHN. Now tell me RedHat, because I trusted you am I going to have to tell them to purchase a fat liscense for you some day in the future and then hire me to migrate them? What in fact will IBM or other companies do? I could live with RH9 on those machines because I wasn't using the software that broke with RH9, and I decided I didn't need IBM's patch on RH7 needed to make the autoreboot feature work. Do these kind of issues make RedHat sound like the kind of a company for which you are willing to put your name on the line? This is utter bullshit, and the funny thing is if they just provided a simple service which kept those servers securely updated for me I would not hesitate to tell all my clients to pay TEN OR TWENTY TIMES what RHN would have cost them. RedHat is going to discover that someone else is going to do it the right way and leverage the off-hours work of RedHat's employees in a very lucrative way. That's where I hope this goes. Or I may just change distros; gentoo, slack, and even FreeBSD and Debian are looking mighty good right now.
It would be useful to be able to keyword search the captions of all these cartoons so you could find appropriate ones for any given situation. Perhaps someone would provide a reference on line? Otherwise it seems utterly impossible to use unless you set up a whole learning infrastructure along the lines of Bible/Torah studies and have scholars dedicate their lives to studying it. Arguably profitable but..
This is great but it took a long time to figure out the rules since the description is so horrible. Well I got this far so good luck!
For 104015 as drawn here is the list of decisions. Basically it seems that every time the worm comes across a point with an eaten segment attached to it, the worm takes the next rule as a new decision, then draws a segment, and then goes back to the first rule in the list (which makes it try to start drawing a hexagon again). However I have to say that Pegg's coloring is rather arbitrary, that lone green segment in the middle stumped me for a while! Also the way decisions are made is not intuitive.
111111 black hexagon
Hit hexagon starting point, pick rule #2 which is a "0".
0 go straight one segment in red.
Go back to start of program, now use rule #1:
111 draw three more red segments using that rule.
Hit the black hexagon again, pick rule #3 (a "4") unclear yet whether it is picked because it is next in order or because the rules #1 and #2 can't be used. But considering next decision, I believe it is because it is next in order:
4 turn two hours to left and draw first blue segment. Then go back to beginning of program.
1111 draw four more blue segments until black hexagon is hit again.
Hit hexagon, pick rule #4 (a "0").
0 draw first yellow segment straight ahead. Then go back to beginning of program
1 draw second yellow segment.
Hit a vertex of the black hexagon, but we can crash through for some reason. So we keep drawing segments with rule #1 which should be in yellow but is green in the drawing:
11 draw two more segments
Hit the blue hexagon so pick rule #6 (a "5").
5 turn two hours to left and draw another green segment. Then go back to beginning of program with rule #1:
1111 draw more green segments
Hit blue hexagon again. Rule #1 doesn't do the job so rule #2 (a "0"):
0 go straight ahead drawing a single green segment. Then back to beginning of program with rule #1:
1 turn right and draw a single green segment.
Hit black hexagon, so pick next rule. All the rules are useless but the last rule #6 (a "5"):
5 turn left and draw a single green segment. Then go back to beginning of program and rule #1:
1 turn right and draw a single green segment.
Now we hit the black and red hexagons so another decision is to be made and we are into magenta which I won't follow.
It still seems wierd how decisions are made and maybe following magenta for yourself will tell you how the recursion is supposed to be done. Personally I would rather see some pseudo code since it is difficult to tell which of the two "0" rules is being chosen.
It could be a short, but it would make more sense (in that I doubt anybody who does shorts is going to be so sure the GPL will win) if Microsoft is behind them. Especially if GPL code is in Windows. This would allow M$ to die in a truly spectacular fashion. Dream on..
Actually, no. I have been making some ad hoc inquiries into this matter with the help of the company microwave, a bunch of free pasta, and a mighty hunger.
Now I am not saying you should microwave water by itself in an unmoving tray or even a moving one for that matter, but (though I think somewhat inferior to ordinary pasta cooking) it works fine if you put cook pasta as I have several times in a plastic microwave safe deep tray with about equal parts water and spaghetti. The starch that initially comes off the pasta perhaps stabilizes it, anyway about 6 minutes at 500 watts on a rotating base and you have lots of small bubbles perking vigorously though I have not had the courage to attempt a rolling boil with a lot of water. At first it was a little scary and you have to wonder what the temperature is going to be when you stop it..
I was surprised but believed his analysis of the impossibility of using C++ as the basic language, but maybe BeOS was like Y, really C with a C++ api.
Actually I was also thinking that I would not like to have the miniscule memory on my laptop being chewed up by buffers but..
If a new integrated windowing system is being designed though I wonder what other things could be done at the same time. For example, could a module be added to Y which subscribed to all events, listened to all data, and was able to index and learn about your activities. It would seem that an object oriented windowing system would be a big ally in this. Also, if a system providing live queries like on the BeOS was implemented, it could answer queries based on the content of open applications, not just static files. Of course I seem to remember IRIX would let you browse windows as files too..
or maybe L. meyeri..
Specifically it must have beenL. olivacea which I guess means olive colored, since as in the photo it had no markings, it just looked like a beautiful hunk of chalky, greenish colored velvety living stone. Can't believe I found it. Some really bizarre, ugly, and beautiful pics on this page. Also more interesting photos here>/a> and here.
I also am thinking of throwing out the pencil plant (Euphorbia tirucalli) stem which will certainly take root by itself, but apparently causes cancer! I wouldn't want a cat to eat it.
Apparently ginkos are also extremely old and resemeble a Jurassic variety. And Cycads, which are woody plants that create seeds. They also seem to be quite poisonous although they are eaten as "beach tucker" after processing in the jungle. (link) Anyway here are some links.
Finally I there are also the extremely visually (and biochemically?) wierd Gymnopsperms like Welwitschia And Ephedra, which seem ancient, maybe same era..
All this because I was trying to figure out if the inch-long stem/leaf in my pocket which I snapped off a huge pencil plant was one of those. Not sure yet.. I remember my mother also has some kind of ancient plant which looks like a gray rock and does nothing, but then one day suddenly splits in half, and then each half will continue to split in the same way recursively. A very cool plant if anyone can figure out what it is!
According to one recent /. article (terascale storage), systems and database administrators will be increasingly in need as storage increases, and also increasingly automated, to solve that problem. But if you focus on the applications/integration side, or if you can find a neat job at a lab you may ride a wave.. Good luck.
..simply spellcheck their submissions before publication? Ispell catches $DIETY fine.
John Walker is also the author of a fascinating set of free programs and book on dieting called the Hacker's Diet in which he discusses a realistic way to lose weight while remaining healthy, using techniques of project management which have served him well in other fields of endeavor. Check it out! (I have the Palm version running myself and have lost 6kg)
Would taint-checking addresses in a CGI program cover this hole in unpatched sendmails?
Certainly you want to patch your own machine, and get the admins you know to fix theirs. But I am thinking now about people running on a virtual hosting server which is perhaps getting most hits from Perl CGI program, and which are using sendmail only internally, for example to send email to automatically registered users. Would it be sufficient to (as usual) check length of address and remove nulls?
Unfortunately I got talked into my folks giving it to a guy when I was away who said he was making something for handicapped, which I don't think every materialized. That was dumb!
I wouldn't mind having some LEDs to show what each key on the keyboard was, it would probably be relaxing. Maybe a software version is not that hard but not sure it would be as pleasing.
I tried to make some kind of silly putty when I had a pretty good chemistry set for a kid, like 25 years ago. (a store-bought totally un-understandable chemistry kit and a ton of antique chemistry glassware my parents found in shops!).
Alas the putty (which was probably more the borax type I guess) was extremely hard to make and it seemed very sensitive to the ratio of ingredients.. I do remember that washing with I guess alcohol maybe, I ended up with a stringy white mass. After a couple tries I got something like putty but kinda nasty..
Anyway my question to offset these sickening legal posts. Do you think the stuff would conduct electricity, or be susceptible to heat from wires? I would guess no on electricity but yes on heat.. anyway wouldn't this be kind of cool material for an embedded linux project (if it is not conducting)?
You could wrap some circuit boards and leds together and bind the whole thing in a mass of putty (maybe thinkgeek's glowing green one, though the blue, white, and black ones are intriguing and the copper one looks good wrapped around buzz' leg..)
presumably the green one would afford a persistence of vision kind of effect like tv phosphors that glow for a long time after being stimulated. I'm a little worried about whether heat or time would melt the stuff, or could it cause a fire.
Could putty be our best friend? I mean you could drop the stuff and even if the putty shatters, that's energy that went into breaking putty bonds and not your hardware! I wouldn't mind buying a 2 1/2 inch hard drive encased in a globe of black smartmass putty, with a black USB wire coming out the back (or wireless!). Well we can dream anyway.