If a well-known investigative journalist appears to turn into a stttering moron, blipverts cause people to explode, and pirate analog TV becomes fashionable, then you are five and a half seconds into the future in a Max Headroom-like Dark Future where all "donor tissue" is from pirated sources. No questions asked.
...that the EU and South Korean courts may have already considered that possibility. South Korea is allowing Microsoft until August to produce a version of Windows that already exists in Europe, for example. A little on the generous side, given that even snail mail isn't that slow. Nor is the EU being lightning-fast in enforcing their penalties. The EU has every economic, legal and political incentive to act quickly but seems to be more content with mere threats.
I'm not saying Microsoft would take, ummm, direct action - although very rich organizations and organized crime undoubtedly have done so in the past, so the precedent exists. However, I do not believe anyone can be quite sure WHAT Microsoft will do at this point. The CEO is a loose cannon and there's no shortage of extremists out there. My suspicion is that the Koreans and the Europeans trying to stir up enough anti-trust action that the shareholders will take action before they have to.
The Enigma code was broken only in the trivial sense that it was possible to brute-force decrypt the messages, once the algorithm, prng and seed value were known. It was not "broken" in the purist sense of the term, in that there is no shorter method of cracking the messages other than by brute-force.
The full Enigma code is extremely difficult to break. The machine used by Alan Turing (Colossus) was massively parallel and highly optimized for the task - so much so that it is actually able to compute something like ten times as many keys per second as a modern Pentium 4 using the same algorithm. Not bad, for a machine of that era.
The Enigma suffered from numerous weaknesses - almost all of them operator error. The encryption mechanism itself was damn good and, if used correctly each time, every time, it would have been horribly difficult for the Bletchley Park team to break.
The one event that turned Enigma transparent was the re-transmission of a message without the cogs being randomized first. Because a machine had already been recovered, Turing knew what the cogs were, just not where they should be in relation to each other. By having the same message sent twice without change and without a prior reset, it was possible to overlay the two messages and thereby infer virtually everything else.
This only allows you to crack messages which use the same prng for initialization and identical cogs. Since the cogs were designed to be swappable, non-standard configurations would have been possible. These would not have been crackable - and would likely not be crackable today, if non-standard enough. (The number of arrangements you would need to test increases with the factorial of the number of ways the cogs could be designed, as well as the factorial of the number of ways the cogs could be inserted into the machine.)
The possibility exists that certain units may have used non-standard Enigma codes, but if that is the case, those codes will NOT be broken by this effort. The groups that spirited high-ranking Germans to South America and other "secure" locations must have had a communication system that the Allies had not yet deciphered, as they must have been able to operate over extremely large distances very quickly, making the use of radio a certainty.
It is also likely that some units within the German military adopted their own "extra secure" practices when using the Enigma system internally. These may or may not be crackable, depending on how paranoid the commanders were.
The judge is going over the case with a fine tooth comb, making absolutely certain everything is 1000% in order, every i dotted, every t crossed, every aspect examined. I would guess she's aware that Microsoft has abused the wording of previous legal decisions against them in order to carry on with assorted malpractices and she would likely want to avoid that. Further, a case like this WILL go to appeal, and she presumably wants to make utterly certain that such an appeal will fail, utterly. Finally, I would guess that she's so sick of SCO that ahe wants to exact the worst punishment that can hold up to scrutiny, so extra care would be essential for that.
There may also be aspects to the case that she believes SCO guilty as all hell, but feels that there may be technicalities which would limit the punishment she can give, or prevent one being given at all. Those parts of the trial are likely to drag on the longest, as bankrupting the bloodsuckers would be preferable to letting them go.
Why not embed a signed Java applet (so it gets the privs of an application) in a web page? The problem with AJAX and Javascript is that you are severely limited in what you can do. A signed/secure Java applet can spawn independent threads, establish connections to machines other than the originating server, gives you all the graphics of Swing and AWT, isn't prone to the differences between JScript and Javascript, can do grid computing via remote method invocations, etc.
Of course, if you want something to run at a decent speed, you do it the other way round - embed the web in your application. On Windows, this is actually quite easy - and the reason it's so vulnerable. You can run procedures in one program from another program, so allowing your primary application to control (or embed) a web browser very easily. Linux supports RPC, there are DCL and OLE extensions for it, and I've lost count of the messaging libraries, but you'd probably be better linking to the Mozilla libraries or the W3C's libwww, and handle things at a lower level.
SGI used to support the VRML standard. Well, until nobody added support in their browsers and everyone realised that it was slow and sucked. In comparison to all the application-level 3D stuff, that is, as there was nothing else 3D you could do on the web that was any good. (I seem to remember some crappy 3D Java applets, but that was about it.)
I never use something that has a version number ending in.0! That's always the buggy release. Besides, I've yet to hear of a single "feature" of this purported 2.0 that wasn't being done with HyperCard and couldn't have been done on Ted Nielson's Xanadu (if anyone had developed it). I see no reason to dignify bugfixes with a change in the major version number.
"But what about blogs?" What about them? People were writing diaries on USENET long before the CERN webserver ever came out. (Was CERN Web 0.0? And would NCSA or Apache be considered 1.0?) Cross-referencing and searches existed in Gopher and WAIS.
"Dynamic HTML?" There were perl scripts for emedding msql queries (not MySQL - msql) into web pages long before anyone had imagined you'd be doing anything other than CGI and many years before HTML 3 came out. Indeed, if you want merely programmable web pages (not database-generated pages) then the mere existance of CGI is enough.
"User-defined web pages" Oracle's "Powerbrowser" included a built-in web server which could serve a limited number of pages to external users. That was back in 1996, if I recall correctly.
Let me know when something worthy of a "Web 2.0" comes out, and THEN I'll pay attention.
Pathe News is already available. The BBC has some excellent high-definition TV-over-IP software. Whether this will actually be used by them is another question, but it would be nice. As I see it, the British have the technology and - God knows, after the BBC junked many of their archives in the 1970s - they have seen how easy it is to irretrievably lose key historical data.
Other very useful archives, such as Pathe News, have already been released over the Internet. (Frankly, I think Pathe will prove much more useful for many world events - at least, those outside of England. For British history, the National Archives might be better.)
Interesting: If it won't hold a kid's interest, then it'll be forgotten when the next toy comes along. The best way for a device to do this is to be re-usable in many ways. One specific game won't last for very long, no matter how good it is.
Modular: This builds off the interest. The more modular a device is, the more ways it can be assembled and the more games the kid can make up as they go along. Later on, modular becomes good for developing experiments, trying to see what works, what doesn't, and what produces the Magic Blue Smoke.
Fun: Intellectual interest is great, but it'll need to hold a high level of emotional interest, too - kids aren't known for having vast reservoirs of intellectual interest. Too few adults do, either, but that's beside the point. Besides, they can always become Talk Radio hosts.
Some examples of what is good:
Lego Mindstorms or any other controllable electronic Lego systems
Mecchano / Erector Sets
K'Nex - you'd want to drive the motors via the computer
Some examples of what would work for SOME kids, especially if older:
Great Egg Race Eggmobile
S-Deck or other solderless electronics kit, using the computer to supply an input or output
Computer-steerable telescope, where telescope eyepiece is rigged to a webcam with output to the computer. Put books giving an introduction to programming and an introduction to image processing next to the computer.
Stuff that is useless:
Any single-function electronic toy
Any single-function computer project or kit
Anything where practical experimentation would be too hard (home-made sugar-based rockets might be a great occasional bit of fun, but I can think of no practical way they can do more than entertain until they're large enough to require special licenses - and even then, research would be extremely limited, for safety reasons)
Anything a furious or distracted kid could turn into an expensive repair project (transistors, capacitors, LEDs - these are dirt cheap, and it takes a fair amount to break lego or mecchano pieces)
By this, I don't just mean 2D, although that is true too. I mean that they use string searches for keywords (so the only useful searches are the ones that just dump ALL the records), there are often multiple categories that mean the same thing (which means you have to search multiple places, to find anything at all), jobs that cross between multiple categories aren't guaranteed to be listed in all of them - and sometimes are listed somewhere else entirely, there is no quantification and job listings are often just place-holders for job listings on some other job sites (which, in turn, may also be just place-holders).
Then there is the resume side. In the era of relational databases, XML, context-sensitive information, semantic wikis and heirarchical data types, I think we can do rather better than a static file. Many sites require you to use a builder anyway, so I utterly fail to see the point in a system that regurgitates information that is of no interest and no relevance to the position to the corporation involved. It wastes the time of the applicant and the Human Resources department and could be avoided completely by allowing data to be grouped and related, then have a dynamic resume that is meaningful for the position.
Finally, there is responsibility. Jobs advertised aren't always genuine (the job doesn't exist, never has, the company or recruiter is merely collecting resumes or otherwise data-mining), are frequently inaccurate (the description doesn't match reality to any degree whatsoever, so it goes beyond mere artistic license), pay-scales are bogus for some careers, and locations can sometimes be 40-50 miles wide of the mark. Job-seekers generally don't have the resouces of the NSA and FBI to track down the inaccuracies. If someone needs a job, chances are they're not going to have the money to hold anyone accontable if the advert does turn out to be utter bullshit. Besides, although some (not all) States protect goods and services through lemon laws and accurate advertising (and even then, protections are so minimal as to barely exist), I have never heard of any such law covering job adverts.
Job sites are little more than online versions of a card index file, with the onus 100% on the reader of the card to know what the writer would have written, had they been honest. This takes no advantage of the technology, but DOES take advantage of job-seekers.
I apologize to those archaeologists who DO do all (or as much as realistic) of the study work that can be done. I guess some of my cynicism comes from the fiasco of Seahenge and also from the archaeology being done in the area I used to live in England - not the fault of those doing the work, they're just so massively underfunded and underequipt, with such gigantic time pressures, that I'm quite convinced they're missing at least as much as they're finding, and they're nowhere near doing all of the tests that could be being done.
If you could suggest a good archaeological journal that actually covers the discoveries that DON'T make the mainstream news (online and free would be ideal, but I somehow doubt there are many of those) then I would truly appreciate it.
On the one hand, one must question the sanity of a person who hands in a large (for the time) solid gold, 6,500 year old artifact. Especially in a country well-known for artifacts being plundered. In England, the Government pays for finds classed as "Treasure Trove", but there is no evidence the unnamed woman got so much as a free can of soft drink for a find that will likely end up in a museum and earn the archaeological team involved in analyzing the find a good few million.
On the other hand, it is precisely because there are people who do hand in such amazing discoveries that so much is known about the ancient world. There are many sites, throughout Europe, which were discovered precisely because of a reported find leading to a study and finally an excavation.
I have often been critical of archaeologists, and the current state of Italy's archaeological remains doesn't give me much confidence in the competence of world heritage organizations either. Many of the major sites are at the point of collapse, one section of wall at a major site DID collapse last year and would have killed a few hundred tourists if it had happened during the day. Emergency repairs, required within the next year or two, will require between ten to twenty times the money budgetted for ALL Itallian archaeology and maintenance for the next decade, simply in order to prevent massive casualties.
Discoveries are of the utmost importance, proper excavation and documentation are vital, but all of that is useless if proper preservation of finds is ignored. The exceptionally fine ancient monument returned from Italy - a massive obelisk that had been plundered during World War II and was in exceptionally good condition, was smashed into three pieces in order to return it on the cheap. If this is the way things are going to happen in future, the Rosetta Stone will be returned to Egypt as a fine powder - the Egyptians can always glue the grains together again, after all.
Sorry if I sound cynical - well, maybe not entirely sorry. I have a very hard time reconciling demonstrable gross incompetence and money hoarding with any kind of respect for heritage or history. As I've said often enough before, we have many possible futures. Futures are a dime a dozen. We can take our pick of those. However, we only ever have one past. Lose that, and it's gone. You don't get another go. Whatever is destroyed is lost and can never be replaced.
Hey, for some things, that probably doesn't matter too much, and there's just too much history to preserve everything 100% from the information level through to the artifacts themselves. The world is only so big and we're running out of room as it is. Besides which, it is really the information that matters anyway, provided you have gathered as much as is practical and lose as little as possible.
In the "perfect world" (at least, perfect in my highly opinionated world view) no effort would be spared to gather all the information that technology can extract, with that information distributed as widely and as freely as the available technology supports. After that, artifacts become relatively unimportant and sites become more useful for tourism than for study. Provided they don't fall down.
I'm not seeing that kind of study going on, though. The new burial site that has been found, for example - there should be plenty of DNA and mDNA that can be extracted for testing to get an idea of the ethnic makeup of the people of the time. They could even put the mDNA markers up on one of the numerous DNA family history sites, to see if living relatives exist and to encourage a greater participation by average folk in the whole archaeology thing. People will be far more willing to invest a little extra time and money on a project if they feel involved - even if only highly superficially - than they will if it is purely seen as the idle musings of some University types with a trowel fetish.
The bounty for finding a critical bug in Windows seems to be about $10,000. Microsoft never gets new technology right, the first time. Vista won't cost $10,000 a seat. How can you possibly fail to make a decent income off the bounties alone, for the remainder of your working life?
Fred Hoyle, who rejected the Bing Bang theory and developed the rival "Continuous Creation" model, in which the Universe expanded and matter was continuously condensing and filling in the gaps, would have loved the current Open Source model.
If you liken Open Source to the raw, new matter being formed, and the corporate sector as being the older, "stable" matter, the current buying up of Open Source, and the community re-filling those gaps with yet more raw stuff, really does fit his model very well. Far better than the physical Universe did!
Adapting Sir Hoyle's model to the software world, it should be possible to make predictions on how well such a system can thrive, what adjustments would be required to keep it functional and keep the creation of new software going, and what the long-term consequences of such an environment would be. If the model is blatantly unstable, we would benefit from knowing that NOW, so we can deal with the commercial sector before it becomes a problem.
On the other hand, if the model is actually very stable and prone to accelerating, we should expect to see the corporate interest fuelling an ever-growing true F/L/OSS community, which would be no bad thing.
Instead of waxing philosophical about the whole deal, it is possible to apply abstract models that depict precisely these sorts of situations, so we can see what the longer-term results would be. Once we know what we're facing, THEN we can wax philosophical all we like, as we'll have something more solid to talk about.
They're now locked in a massive lawsuit in something like a dozen US States over potential port security, right now, owing to fears that they're going to utterly demolish the (non-existant) port security of America. This space tourism venture is interesting, but there are considerable risks that political hostilities could seriously impact any high-tech operation in the UAE. Also, we are talking about a region that is troubled and definitely outside of the US' "tier 1" for ITAR purposes. It will be interesting to see what happens, but until things improve, I would recommend seeing at a suitably LONG distance.
...it assumes volatile RAM. Duh. Further, unmapping RAM doesn't wipe it - it just deletes the pointer. If you remove the assumption of volatile memory, you remove the need to transfer anything. All that would be required would be to suspend the thread when you hit the point in memory that has been hot-swapped out.
This is child's stuff and not even novel. Mainframes using mag-tape for virtual memory have allowed hot-swapped live processes for something like 20-30 years. Moving the capability to non-volatile hot-swappable main memory would not be hard. You couldn't use the patches as they stand, but I never said you could. You'd have to extend a few concepts first, but it IS just an extension and the method ALREADY exists in other systems. The novelty factor, which I believe would attract attention, is that it's never been done for main memory (only virtual memory) and it's never been done for home computers (only Really Big Iron). Neither of these are big problems, they barely qualify as small problems, but precisely because they have never been tried, you would get a lot of attention.
Honestly, sometimes I feel that 99% of the computing world's problems is a total lack of imagination on using existing methods. This is not the first time I've noted you can use pre-existing technology for more than it was originally designed for. I'm not even unique in that. But invariably, such uses are panned viciously - until they become the norm, at which point it becomes blindingly obvious to even the densest that it's remarkably doable. All that happens is that those who pointed out the possibility in the first place get panned for the next thing they speculate about for precisely the same reason.
Gah. I've been putting up with this for the past 28 years, I dunno why I'm expecting things to be any different now.
The question is basically a re-statement of the original ARPAnet design, you are correct. However, to be absolutely true to the question, you'd need two additional stipulations.
First, to be effective, all network connections would need to be fairly fat. A tiered Internet is designed along the same sort of design philosophy as a "fat tree" - low bandwidth at the work-node level, massive bandwidth in the middle. A tierless Internet, particularly one that supported enough multiple paths to be useful for robustness and decentralization of control, would need ALL connections to be much fatter than they currently are. You'd need gigabit to ten gigabit pipes between the majority of machines to be useful.
Second, you can't use the design strategy of bordered autonomous clouds, linked by a backbone, because you'd have no backbone. With no borders, you can't use internal and external routing protocols, as there would be no "internal" or "external". Besides which, they mostly suck when it comes to massively meshed networks where individual connections are unreliable and potentially mobile. BGP, OSPF - you'd need to RIP (yeah, bad pun) them out and replace them with an ad-hoc mesh routing protocol that supported mobile IP and NEMO. The complexity would be much higher, particularly as software packet switching and software routing are CPU and bus killers, which means an optimal path would need to figure in the density of traffic in a fairly sizable part of the mesh. Modern architectures just aren't built to handle such a design, but that would not stop you from building an architecture that COULD support it.
So, (1) yes it is possible, but (2) not effectively with the existing infrastructure or existing PC designs, though (3) both of those problems are solvable.
...because the memory is non-volatile, you could swap it out with other memory devices, restoring the machine to whatever you happened to feel like at the time. Thus, if a computer broke down, you could swap the memory card to a new laptop and "restore" the new laptop to the point at which the old machine left off. (You could also use this to handle failover systems where the machine was more likely to fail than an application.)
Linux now supports hot-swapping of actual RAM, which introduces the possibility of physically transferring live, active threads between machines without the need for a physical network. Not sure how useful this would be in practice, but the pure geek value in such a demonstration would surely turn a LOT of heads in the IT profession.
I was talking strictly non-volatile. If you want to talk about volatile RAM, like DRAM, where you are going to refresh the contents every few nanoseconds, degradation of contents - provided it is slower than your refresh rate - is completely unimportant. In fact, degradation of content is precisely WHY you have to refresh the content. In fact, fast degradation is a GOOD thing for volatile RAM. It means you can change the contents extremely quickly. Completely the opposite requirement of non-volatile storage, where retention is the key consideration.
Volatile RAM also has to remain powered at all times. Again, this is a GOOD thing. Old-fashioned "core" memories could retain data for a hundred years plus, which made rebooting somewhat of a lengthy process. You would not, for example, build a CPU where the internal registers used "core" memory or any other form of non-volatile memory. At least, not unless you were very drunk.
On the other hand, if you wanted to replace a hard drive, DRAM is next to useless. Sure, you can have a stack of NiCad batteries in parallel to keep the memory going, provided you remember to replace/recharge them as needed. Wouldn't help you, though, if you had a short. For mass storage, where the contents absolutely needs to be retained for a long period of time, you absolutely do NOT want to use DRAM.
When you get right down to it, though, if the CPU had a gig or four of register-speed RAM on board, you wouldn't really want DRAM for anything. Main memory is only useful because it's substantially cheaper than register-speed RAM and it wouldn't be trivial to build a processor big enough to hold that much memory. Main memory, for a long time now, has been treated as little more than a cache for virtual memory, where all the real storage is on disk, and as a dumping ground for what memory the processor does have. If CPUs held enough, and/or mass storage was fast enough, main memory would go the way of the dodo. It's a relic that persists only because the alternatives are too limited right now.
Solid-state "disks" (such as the 1980's "solidisk" system) may be the future, but they're also very much the past too. Genuinely non-volatile solid-state memory date back to the earliest "core" memories, but have taken many forms (eeproms, bubble memory - there are even forms of static RAM that can hold data for significant periods of time with no power).
I would also question the usefulness of the proposed system. I am not confident you could change the spin of anything at that scale for any useful length of time. Too many variables and too much "noise". If you want to change a property, it needs to be a property that can "latch" in whatever state you place it and have no trivial way of unlatching itself without significant input. Otherwise, your data will degrade very rapidly.
There are two ways to "store" data - permanently or erasably. Permanent storage is much simpler, in that there need not be any way of reversing the process. It's better to do this in a mechanical form, because you can have a much higher density. Erasable storage is better as solid-state, because erasable mechanical storage will wear out rapidly, which means it's not particularly reliable or trustable over meaningful periods of time.
Permanent storage that is high density is relatively simple. You could have a mix of two molecules which are highly stable but, when energy is delivered, react to form something different. Since different molecules absorb energy at different wavelengths, the absorption pattern would give you your 1s and 0s. Molecules are extremely small, compared to magnetic fields or even to the "blisters" formed on CDROMs to store data. You can also look at multiple bits at the same time, with this method. Unlike conventional magnetic media, a read-head need not be serially streaming data but could read as much in parallel as you liked. This WOULD be permanent, though, so would only be useful as a means of replacing CDROMs or DVDs, but would be far more expensive per byte of data and would only offer an advantage where you needed such a system to be considerably faster and vastly more durable.
Erasable non-volatile storage is a tougher problem, as you need something that can be altered by an electric current in both directions and where the change could be read through some alteration in an electric current. This can get to be a problem, if you want extremely high densities of storage, as all the supporting electronics will take space and will likely take space for each and every single bit of data. (Pun intended.) Usually, there is some magnetic component to such systems (magnets are good at holding states) OR a battery backup, as transistors won't hold a state when there is no power to them. There are many ways of building such an arrangement, with different methods having different speeds for read and write and different densities of storage.
I would assume that one could (ab)use "electron migration" to store information, provided an easy way of resetting the electrons existed. This would have the benefit of not needing any magnetic mechanisms (which may mean you could get higher densities) but it would certainly be slower to write to, and likely to read from. I would suspect that something similar will offer much better opportunities for solid-state non-volatile storage in the future, precisely because it should be capable of far higher densities.
An electron has 720' rotational symmetry (see: Brief History of Time) so if they spin it too far, it'll become a positron. Since they've no way of detecting the rotation of an electron (it's a point charge) other than seeing if it explodes when it strikes another electron, this could definitely be an interesting - if short-lived - storage mechanism.
Slashdot Mirror
If a well-known investigative journalist appears to turn into a stttering moron, blipverts cause people to explode, and pirate analog TV becomes fashionable, then you are five and a half seconds into the future in a Max Headroom-like Dark Future where all "donor tissue" is from pirated sources. No questions asked.
I'm not saying Microsoft would take, ummm, direct action - although very rich organizations and organized crime undoubtedly have done so in the past, so the precedent exists. However, I do not believe anyone can be quite sure WHAT Microsoft will do at this point. The CEO is a loose cannon and there's no shortage of extremists out there. My suspicion is that the Koreans and the Europeans trying to stir up enough anti-trust action that the shareholders will take action before they have to.
The full Enigma code is extremely difficult to break. The machine used by Alan Turing (Colossus) was massively parallel and highly optimized for the task - so much so that it is actually able to compute something like ten times as many keys per second as a modern Pentium 4 using the same algorithm. Not bad, for a machine of that era.
The Enigma suffered from numerous weaknesses - almost all of them operator error. The encryption mechanism itself was damn good and, if used correctly each time, every time, it would have been horribly difficult for the Bletchley Park team to break.
The one event that turned Enigma transparent was the re-transmission of a message without the cogs being randomized first. Because a machine had already been recovered, Turing knew what the cogs were, just not where they should be in relation to each other. By having the same message sent twice without change and without a prior reset, it was possible to overlay the two messages and thereby infer virtually everything else.
This only allows you to crack messages which use the same prng for initialization and identical cogs. Since the cogs were designed to be swappable, non-standard configurations would have been possible. These would not have been crackable - and would likely not be crackable today, if non-standard enough. (The number of arrangements you would need to test increases with the factorial of the number of ways the cogs could be designed, as well as the factorial of the number of ways the cogs could be inserted into the machine.)
The possibility exists that certain units may have used non-standard Enigma codes, but if that is the case, those codes will NOT be broken by this effort. The groups that spirited high-ranking Germans to South America and other "secure" locations must have had a communication system that the Allies had not yet deciphered, as they must have been able to operate over extremely large distances very quickly, making the use of radio a certainty.
It is also likely that some units within the German military adopted their own "extra secure" practices when using the Enigma system internally. These may or may not be crackable, depending on how paranoid the commanders were.
There may also be aspects to the case that she believes SCO guilty as all hell, but feels that there may be technicalities which would limit the punishment she can give, or prevent one being given at all. Those parts of the trial are likely to drag on the longest, as bankrupting the bloodsuckers would be preferable to letting them go.
Law firms will now invest large sums in Linux companies, in the hope that they'll get more business.
Of course, if you want something to run at a decent speed, you do it the other way round - embed the web in your application. On Windows, this is actually quite easy - and the reason it's so vulnerable. You can run procedures in one program from another program, so allowing your primary application to control (or embed) a web browser very easily. Linux supports RPC, there are DCL and OLE extensions for it, and I've lost count of the messaging libraries, but you'd probably be better linking to the Mozilla libraries or the W3C's libwww, and handle things at a lower level.
SGI used to support the VRML standard. Well, until nobody added support in their browsers and everyone realised that it was slow and sucked. In comparison to all the application-level 3D stuff, that is, as there was nothing else 3D you could do on the web that was any good. (I seem to remember some crappy 3D Java applets, but that was about it.)
"But what about blogs?" What about them? People were writing diaries on USENET long before the CERN webserver ever came out. (Was CERN Web 0.0? And would NCSA or Apache be considered 1.0?) Cross-referencing and searches existed in Gopher and WAIS.
"Dynamic HTML?" There were perl scripts for emedding msql queries (not MySQL - msql) into web pages long before anyone had imagined you'd be doing anything other than CGI and many years before HTML 3 came out. Indeed, if you want merely programmable web pages (not database-generated pages) then the mere existance of CGI is enough.
"User-defined web pages" Oracle's "Powerbrowser" included a built-in web server which could serve a limited number of pages to external users. That was back in 1996, if I recall correctly.
Let me know when something worthy of a "Web 2.0" comes out, and THEN I'll pay attention.
Pathe News is already available. The BBC has some excellent high-definition TV-over-IP software. Whether this will actually be used by them is another question, but it would be nice. As I see it, the British have the technology and - God knows, after the BBC junked many of their archives in the 1970s - they have seen how easy it is to irretrievably lose key historical data.
I'd always thought the moon had formed a few billion years ago.
Other very useful archives, such as Pathe News, have already been released over the Internet. (Frankly, I think Pathe will prove much more useful for many world events - at least, those outside of England. For British history, the National Archives might be better.)
Modular: This builds off the interest. The more modular a device is, the more ways it can be assembled and the more games the kid can make up as they go along. Later on, modular becomes good for developing experiments, trying to see what works, what doesn't, and what produces the Magic Blue Smoke.
Fun: Intellectual interest is great, but it'll need to hold a high level of emotional interest, too - kids aren't known for having vast reservoirs of intellectual interest. Too few adults do, either, but that's beside the point. Besides, they can always become Talk Radio hosts.
Some examples of what is good:
Some examples of what would work for SOME kids, especially if older:
Stuff that is useless:
Then there is the resume side. In the era of relational databases, XML, context-sensitive information, semantic wikis and heirarchical data types, I think we can do rather better than a static file. Many sites require you to use a builder anyway, so I utterly fail to see the point in a system that regurgitates information that is of no interest and no relevance to the position to the corporation involved. It wastes the time of the applicant and the Human Resources department and could be avoided completely by allowing data to be grouped and related, then have a dynamic resume that is meaningful for the position.
Finally, there is responsibility. Jobs advertised aren't always genuine (the job doesn't exist, never has, the company or recruiter is merely collecting resumes or otherwise data-mining), are frequently inaccurate (the description doesn't match reality to any degree whatsoever, so it goes beyond mere artistic license), pay-scales are bogus for some careers, and locations can sometimes be 40-50 miles wide of the mark. Job-seekers generally don't have the resouces of the NSA and FBI to track down the inaccuracies. If someone needs a job, chances are they're not going to have the money to hold anyone accontable if the advert does turn out to be utter bullshit. Besides, although some (not all) States protect goods and services through lemon laws and accurate advertising (and even then, protections are so minimal as to barely exist), I have never heard of any such law covering job adverts.
Job sites are little more than online versions of a card index file, with the onus 100% on the reader of the card to know what the writer would have written, had they been honest. This takes no advantage of the technology, but DOES take advantage of job-seekers.
If you could suggest a good archaeological journal that actually covers the discoveries that DON'T make the mainstream news (online and free would be ideal, but I somehow doubt there are many of those) then I would truly appreciate it.
On the other hand, it is precisely because there are people who do hand in such amazing discoveries that so much is known about the ancient world. There are many sites, throughout Europe, which were discovered precisely because of a reported find leading to a study and finally an excavation.
I have often been critical of archaeologists, and the current state of Italy's archaeological remains doesn't give me much confidence in the competence of world heritage organizations either. Many of the major sites are at the point of collapse, one section of wall at a major site DID collapse last year and would have killed a few hundred tourists if it had happened during the day. Emergency repairs, required within the next year or two, will require between ten to twenty times the money budgetted for ALL Itallian archaeology and maintenance for the next decade, simply in order to prevent massive casualties.
Discoveries are of the utmost importance, proper excavation and documentation are vital, but all of that is useless if proper preservation of finds is ignored. The exceptionally fine ancient monument returned from Italy - a massive obelisk that had been plundered during World War II and was in exceptionally good condition, was smashed into three pieces in order to return it on the cheap. If this is the way things are going to happen in future, the Rosetta Stone will be returned to Egypt as a fine powder - the Egyptians can always glue the grains together again, after all.
Sorry if I sound cynical - well, maybe not entirely sorry. I have a very hard time reconciling demonstrable gross incompetence and money hoarding with any kind of respect for heritage or history. As I've said often enough before, we have many possible futures. Futures are a dime a dozen. We can take our pick of those. However, we only ever have one past. Lose that, and it's gone. You don't get another go. Whatever is destroyed is lost and can never be replaced.
Hey, for some things, that probably doesn't matter too much, and there's just too much history to preserve everything 100% from the information level through to the artifacts themselves. The world is only so big and we're running out of room as it is. Besides which, it is really the information that matters anyway, provided you have gathered as much as is practical and lose as little as possible.
In the "perfect world" (at least, perfect in my highly opinionated world view) no effort would be spared to gather all the information that technology can extract, with that information distributed as widely and as freely as the available technology supports. After that, artifacts become relatively unimportant and sites become more useful for tourism than for study. Provided they don't fall down.
I'm not seeing that kind of study going on, though. The new burial site that has been found, for example - there should be plenty of DNA and mDNA that can be extracted for testing to get an idea of the ethnic makeup of the people of the time. They could even put the mDNA markers up on one of the numerous DNA family history sites, to see if living relatives exist and to encourage a greater participation by average folk in the whole archaeology thing. People will be far more willing to invest a little extra time and money on a project if they feel involved - even if only highly superficially - than they will if it is purely seen as the idle musings of some University types with a trowel fetish.
The pendant is another good example. Gold i
...I'm inclined to think the correct title should be "Stir Fried". :)
The bounty for finding a critical bug in Windows seems to be about $10,000. Microsoft never gets new technology right, the first time. Vista won't cost $10,000 a seat. How can you possibly fail to make a decent income off the bounties alone, for the remainder of your working life?
If you liken Open Source to the raw, new matter being formed, and the corporate sector as being the older, "stable" matter, the current buying up of Open Source, and the community re-filling those gaps with yet more raw stuff, really does fit his model very well. Far better than the physical Universe did!
Adapting Sir Hoyle's model to the software world, it should be possible to make predictions on how well such a system can thrive, what adjustments would be required to keep it functional and keep the creation of new software going, and what the long-term consequences of such an environment would be. If the model is blatantly unstable, we would benefit from knowing that NOW, so we can deal with the commercial sector before it becomes a problem.
On the other hand, if the model is actually very stable and prone to accelerating, we should expect to see the corporate interest fuelling an ever-growing true F/L/OSS community, which would be no bad thing.
Instead of waxing philosophical about the whole deal, it is possible to apply abstract models that depict precisely these sorts of situations, so we can see what the longer-term results would be. Once we know what we're facing, THEN we can wax philosophical all we like, as we'll have something more solid to talk about.
They're now locked in a massive lawsuit in something like a dozen US States over potential port security, right now, owing to fears that they're going to utterly demolish the (non-existant) port security of America. This space tourism venture is interesting, but there are considerable risks that political hostilities could seriously impact any high-tech operation in the UAE. Also, we are talking about a region that is troubled and definitely outside of the US' "tier 1" for ITAR purposes. It will be interesting to see what happens, but until things improve, I would recommend seeing at a suitably LONG distance.
This is child's stuff and not even novel. Mainframes using mag-tape for virtual memory have allowed hot-swapped live processes for something like 20-30 years. Moving the capability to non-volatile hot-swappable main memory would not be hard. You couldn't use the patches as they stand, but I never said you could. You'd have to extend a few concepts first, but it IS just an extension and the method ALREADY exists in other systems. The novelty factor, which I believe would attract attention, is that it's never been done for main memory (only virtual memory) and it's never been done for home computers (only Really Big Iron). Neither of these are big problems, they barely qualify as small problems, but precisely because they have never been tried, you would get a lot of attention.
Honestly, sometimes I feel that 99% of the computing world's problems is a total lack of imagination on using existing methods. This is not the first time I've noted you can use pre-existing technology for more than it was originally designed for. I'm not even unique in that. But invariably, such uses are panned viciously - until they become the norm, at which point it becomes blindingly obvious to even the densest that it's remarkably doable. All that happens is that those who pointed out the possibility in the first place get panned for the next thing they speculate about for precisely the same reason.
Gah. I've been putting up with this for the past 28 years, I dunno why I'm expecting things to be any different now.
First, to be effective, all network connections would need to be fairly fat. A tiered Internet is designed along the same sort of design philosophy as a "fat tree" - low bandwidth at the work-node level, massive bandwidth in the middle. A tierless Internet, particularly one that supported enough multiple paths to be useful for robustness and decentralization of control, would need ALL connections to be much fatter than they currently are. You'd need gigabit to ten gigabit pipes between the majority of machines to be useful.
Second, you can't use the design strategy of bordered autonomous clouds, linked by a backbone, because you'd have no backbone. With no borders, you can't use internal and external routing protocols, as there would be no "internal" or "external". Besides which, they mostly suck when it comes to massively meshed networks where individual connections are unreliable and potentially mobile. BGP, OSPF - you'd need to RIP (yeah, bad pun) them out and replace them with an ad-hoc mesh routing protocol that supported mobile IP and NEMO. The complexity would be much higher, particularly as software packet switching and software routing are CPU and bus killers, which means an optimal path would need to figure in the density of traffic in a fairly sizable part of the mesh. Modern architectures just aren't built to handle such a design, but that would not stop you from building an architecture that COULD support it.
So, (1) yes it is possible, but (2) not effectively with the existing infrastructure or existing PC designs, though (3) both of those problems are solvable.
Linux now supports hot-swapping of actual RAM, which introduces the possibility of physically transferring live, active threads between machines without the need for a physical network. Not sure how useful this would be in practice, but the pure geek value in such a demonstration would surely turn a LOT of heads in the IT profession.
Volatile RAM also has to remain powered at all times. Again, this is a GOOD thing. Old-fashioned "core" memories could retain data for a hundred years plus, which made rebooting somewhat of a lengthy process. You would not, for example, build a CPU where the internal registers used "core" memory or any other form of non-volatile memory. At least, not unless you were very drunk.
On the other hand, if you wanted to replace a hard drive, DRAM is next to useless. Sure, you can have a stack of NiCad batteries in parallel to keep the memory going, provided you remember to replace/recharge them as needed. Wouldn't help you, though, if you had a short. For mass storage, where the contents absolutely needs to be retained for a long period of time, you absolutely do NOT want to use DRAM.
When you get right down to it, though, if the CPU had a gig or four of register-speed RAM on board, you wouldn't really want DRAM for anything. Main memory is only useful because it's substantially cheaper than register-speed RAM and it wouldn't be trivial to build a processor big enough to hold that much memory. Main memory, for a long time now, has been treated as little more than a cache for virtual memory, where all the real storage is on disk, and as a dumping ground for what memory the processor does have. If CPUs held enough, and/or mass storage was fast enough, main memory would go the way of the dodo. It's a relic that persists only because the alternatives are too limited right now.
I would also question the usefulness of the proposed system. I am not confident you could change the spin of anything at that scale for any useful length of time. Too many variables and too much "noise". If you want to change a property, it needs to be a property that can "latch" in whatever state you place it and have no trivial way of unlatching itself without significant input. Otherwise, your data will degrade very rapidly.
There are two ways to "store" data - permanently or erasably. Permanent storage is much simpler, in that there need not be any way of reversing the process. It's better to do this in a mechanical form, because you can have a much higher density. Erasable storage is better as solid-state, because erasable mechanical storage will wear out rapidly, which means it's not particularly reliable or trustable over meaningful periods of time.
Permanent storage that is high density is relatively simple. You could have a mix of two molecules which are highly stable but, when energy is delivered, react to form something different. Since different molecules absorb energy at different wavelengths, the absorption pattern would give you your 1s and 0s. Molecules are extremely small, compared to magnetic fields or even to the "blisters" formed on CDROMs to store data. You can also look at multiple bits at the same time, with this method. Unlike conventional magnetic media, a read-head need not be serially streaming data but could read as much in parallel as you liked. This WOULD be permanent, though, so would only be useful as a means of replacing CDROMs or DVDs, but would be far more expensive per byte of data and would only offer an advantage where you needed such a system to be considerably faster and vastly more durable.
Erasable non-volatile storage is a tougher problem, as you need something that can be altered by an electric current in both directions and where the change could be read through some alteration in an electric current. This can get to be a problem, if you want extremely high densities of storage, as all the supporting electronics will take space and will likely take space for each and every single bit of data. (Pun intended.) Usually, there is some magnetic component to such systems (magnets are good at holding states) OR a battery backup, as transistors won't hold a state when there is no power to them. There are many ways of building such an arrangement, with different methods having different speeds for read and write and different densities of storage.
I would assume that one could (ab)use "electron migration" to store information, provided an easy way of resetting the electrons existed. This would have the benefit of not needing any magnetic mechanisms (which may mean you could get higher densities) but it would certainly be slower to write to, and likely to read from. I would suspect that something similar will offer much better opportunities for solid-state non-volatile storage in the future, precisely because it should be capable of far higher densities.
An electron has 720' rotational symmetry (see: Brief History of Time) so if they spin it too far, it'll become a positron. Since they've no way of detecting the rotation of an electron (it's a point charge) other than seeing if it explodes when it strikes another electron, this could definitely be an interesting - if short-lived - storage mechanism.