It's not just advertising Wifi availability - it's advertising that they probably charge money for it, especially since they paid $100 for three window decals, and that they might accept one-time charges or might not accept them, and might accept other members' service plans or not. It tells you that you're not going to get free Wifi service with your latte, so you might as well see if the other coffee shop on the block/mall/airport has free wifi or a neatly printed warchalk logo in _their_ window.
The performance certification isn't stunning, either - it's nice to get 2Mbps radio, but they're only insisting on 128kbps of actual wired bandwidth behind it, so who cares how fast the radio link is? (I'd initially read it as requiring 2Mbps of _bandwidth_, which means a European E1 can almost do it, unless they use 1984kbps or 1920kbps framing, but an American/Japanese/Taiwanese T1 can't. But it's just talking about radio speed. 128kbps is just ISDN/IDSL or slow-DSL-uplink speed, though it certainly beats modems.)
The no-interference-with-VPNs part is good, though it's not clear whether it's banning NAT or not (many IPSEC implementations support UDP tunnels for NAT, but it's bandwidth-intensive and not universally supported.)
Or if it was, it at least didn't have the same cooler-than-you appeal that Graffiti did, probably because it was done by lower-paid-than-techies typists using dead trees. It did actually have some coolness factor, but that was made up for by the fact that 99% of the people using it were using it to do things for somebody else, rather than using it to do things for themselves, and it was usually followed by typing the stuff. And while some of them were secretaries, many of them were just typists.
Macworld was in SF this week, so I went there for an hour or so, which was enough:-) There was a surprising amount of furniture. I don't remember the name of it, but somebody had a line of white and chrome stuff that went with the half-sphere iMacs, which provided a bottom half sphere and some bent chrome tubes and a keyboard/mouse holder on another extension arm, kind of like having your desk replaced with a white spider.
Hey, I've got just the obvious, childish anti-Microsoft joke you're looking for, but actually Win2K does seem to crash much much less often than previous products of theirs:-)
More seriously, though, another reason to put the power button on the keyboard is so you can put the computer under the desk or somewhere else that's not very accessible, and you only need to access the machine when you're feeding it CDs (or not even then, if you've got a USB CD drive.) This also lets you make the room quieter by hiding the PC in an accoustic-shielding box, or in a closet or in the next room with long cables.
The War on Drugs already killed pay-phone availability in areas where those "economically disadvantaged" live. One of the other commenters said that they only see drug dealers using them, but not only _do_ they use them, replying to their beepers from phone numbers that aren't easily traced to them, but police and cities have discouraged phone companies from providing them there because they want to discourage drug dealers, and this is most common in poorer urban areas. Pay phones used to be able to support incoming calls as well, but the telephone deregulation changes in the 80s and early 90s that let them be privately operated instead of only run by telcos killed that, because private operators didn't get any revenue from receiving calls (and also, the War On Politically Incorrect Drugs also meant that drug dealers would use them to receive calls.)
A few years ago, before I got a cell phone, I was trying to hunt for an apartment or house to rent in the San Francisco Bay Area. This involved a lot of trying to contact landlords and property managers to get in to see places that were advertised, but they're never in their offices - you call their beeper or answering machine and leave your phone number. I did have a beeper, but of course with no PAY PHONES around, it was hard to call them back. In some areas, there'd either be a 7-11 or else a restaurant that had a phone in the back, so if we'd left enough calls in a given area, we'd get coffee and more quarters and wait. Really frustrating....
Of course, pay phone usage in poor areas also went down because of low-cost lifeline phone rates, and because deregulation meant that the prices of pay phone calls went way up, and in high-crime neighborhoods, a coin-operated pay phone looks a lot like a parking meter - it's a box of money sitting there for any teenager with a spare metal pipe, unless it's in a well-lit high-traffic area.
I look at all of these as similar phenomenon. There's a certain amount of quality filtration that probably happens with Dojinshi, just because the costs of paper reduces distribution for works that aren't somewhat good or at least somewhat popular, while the Internet makes it easier to distribute bad stuff. And yes, slash is a subgenre of fan fiction, not the whole thing, just as hentai is a subgenre of manga (and probably of Dojinshi as well?.)
And fan comics are a subgenre of fan fiction as well; most of it that I've seen in the US has been from those evil Furries, but there's a lot of other artwork in the science fiction world.
Fanfic aka FanFiction aka Slash is not just a Japanese Anime phenomenon; it's wildly popular with Science Fiction / Fantasy / Horror book, movie, and TV genres as well, and is especially easy to find on the net. Some large fraction of it may be obsessively devoted to obsessively describing Scully's sexual relationships with Mulder or the Cigarette Smoking Man, or Kirk's with Spock, or Buffy's with whomever or whatever, but that's just Donaldson's Law ("Sturgeon was an optimist"). Some of it's actually quite good, and some of it's really really bad but still funny.
It's not a big deal - obviously there will be important packages that nobody's really interested in, like boring-glue-stuff.rpm. So you use the voting process to pick the popular packages, and use your engineering judgement to add the other critical stuff, but you probably don't need more than 100-200MB of mandatory stuff, plus another 100MB of packages that are required to support the packages picked by the popular vote.
That link points to http://www.mandrakelinux.com/en/ftptmp/1042225920. 78796bc94d6342c90d73c6d7e2ec5baa.php#beta which gets the usual Mandrake web server message about "That page is missing. Go back to the main page". Probably some sort of auto-updating thing; I got that two days ago when trying to find the mirrors page to download 9.0:-). Alternatively, it's possible that my company's firewall is up to something.
Yeah, I remember those days too. We generally figured about 1 failure per year per drive. Of course, our most common drive type for a few years was removable-pack RM05s, which made nice pretty wall decorations after having a head crash.
There's a bit of a price difference, as well:-) In about 1984, we'd paid about $35000 for four RM05 drives, which gave us 1GB of total drive capacity, and removable 250MB disk packs were about $1100. So that was $4/MB for the media, and $35/MB for the drives. I forget what tape drives cost, but 6250bpi 9-track tapes cost about $25 and held about 160MB of data, so they were the obvious backup medium.
Searchking hasn't been the only attack on Google. There's also Daniel Brandt, the Google-watch guy, who thinks that Google ought to be regulated, because it's so important. Both of these people are wrong - the Searchking guy because he's a parasite, and the Google-watch guy because, as near as I can tell, he's one of those misguided people who wants to tell everybody what to do because he knows better but does not in fact understand what makes the Net work so well, which is that anybody can put any material they want out there.
Well the reason Google is so important is that so many people use it, and the reason so many people use it instead of Yahoo or Altavista or Northernlights or Hotbot or LongDefunct.Com or Excite.com or Teoma or some of the other search engines out there is that they do a really excellent job. I used to use Altavista, who were not only the original big search engine, they were one of the best in terms of coverage, as opposed to Yahoo who had much better indexing but nowhere near as many pages. If you wanted to find something obscure, you'd use Altavista, but if you wanted to find something common, it might be hard because Altavista would get 50,000 references that you could look through 10 at a time. I switched to using Google because their search engine did a really good job of usually having the information I wanted in the first page or two, often in the first one or two references, as well as because their pages were lean and mean and not cluttered with dancing broken Javascript ads, and I've occasionally found the cache to be valuable for finding information that was once on the web but isn't any more.
As far as Daniel Brandt's rants about how the government ought to be regulating Google and PageRank because so many people use it, that's purely backwards. The government could accomplish any positive aspects of his goals by building their own search engines with their own page ranking algorithms, but if they go messing with Google, they're not only likely to censor some content and artificially inflate things they want to propagandize, but they're likely to make it less likely to have the material I want near the top, destroying the Pagerank in order to save it.
Some search engines have tried to make money by letting people pay for good placement - the pundits yell at them for it, and the public tends to use those engines less because they're better at finding advertising drivel than interesting content. Lots of web sites try to game the page ranking systems on all the major search engines, typically by including relevant keywords many many times in comments or meta-things, or by including them in small print at the bottom of the page, and the main reason the system doesn't get swamped by this is that the better algorithms try to detect this manipulation and neutralize it or seriously downrate for it. Otherwise the search engines would have a high proportion of uninteresting material near the top, mostly pages that are really just spam. If Google's PageRank didn't protect itself against whatever techniques SearchKing is using, he'd be doing the same thing, making it much easier to find pages people pay to promote than pages that are rated high because they're actually interesting. (I've got slightly mixed feelings about that, because his stuff seems to look less obnoxious than banner ads or dancing javascipts, and is usually on pages I don't ever read...)
Cell phones, except possibly a few new ones, don't _have_ GPS-type features. GPS technology is getting smaller and lower in power consumption, but still takes too much space and power for a small phone. What actually goes on is that cell phones talk to cell antenna towers so that they've always got the best signal they can get, and this lets them locate you because (in the dumb version) the system knows what cell site you're using, and (in the newer smarter versions) the nearest N cell sites know you're nearby so they can triangulate. If you turn the phone off, and then turn it on again, it sends out a broadcast letting the system know it's there.
On the other hand, there are a number of international carriers starting to use VOIP, because it makes it convenient to use compression and save a lot of bandiwidth in spite of the IP packet overhead, plus of course the Net2Phone-type-people do VOIP.
Residential 911 is relatively easy - the phone company installed your phone at your house and knows where it is, so they can program the 911 computer to know that the call from +1-202-456-2121 is from 1600 Pennsylvania Ave. Cellphone 911 is obviously hard - the phone can be almost anywhere, and the easiest thing the phone company can tell is that it's near the cell site it's using now; some digital cell phone standards can do a reasonable job of triangulation on distances to multiple cell towers to figure out locations. At least the wireless people have some clue.
For VOIP, though, the way the phone call reaches the phone company is that somebody has a box that translates between phone lines (usually T1 trunks) and IP addresses, so the only thing the phone company knows is that it got a call from 202-456-2121, which terminates on a box in the basement at 1600 Pennsylvania Ave. PBXs already cause some trouble with this, but they're often bright enough to know that the call is from extension 1234, and somebody can run a database that knows that x1234 is on the 12th floor. With VOIP, it's even worse - the VOIP-to-POTS gateway is some router or PBX that knows it got a connection from 10.32.11.1, and it's possible that somebody has a database that shows that that address belongs to a DHCP server on the 12th floor, not that the 911-police know how to find your data network management staff in a hurry, but in most of the VOIP standards, there's really no information beyond the IP address.
And of course your IP address might be anywhere in the world - did you dial in today from home, or a hotel, or an airplane, or your corporate office 3000 miles away? And think about the VOIP phones themselves - some people use telephone-oriented software applications on their PCs (it's 10.01.01.23 - do you know where your laptop is?), while other people use desktop VOIP phones from a variety of vendors, which you program to know that you're Linus Torvalds on +1-202-456-2121, and if you plug them in anywhere in your company's network, they'll find the gateway server, let it know your current IP address, and be ready to pick up your voicemail and incoming phone calls. Anywhere. So if you dial 911, the town your company's main office is in knows that there's an emergency somewhere near you, but it doesn't know where you are. And if you're not using a coporate VOIP system, you could really be almost anywhere. And if you're going through a NAT firewall or VPN gateway, you could be even farther anywhere.
So what kinds of approaches can people take to fix this? The two obvious first steps are either to get the phone company out of the way (give the 911 people VOIP so they can at least try to traceroute you, though that still has all the IP-vs-location uncertainties), or else to make sure that the VOIP standards are updated to do a better job of passing location information (for people who want to pass it) and that the VOIP-to-POTS gateway standards provide some mechanism for passing that along, whether it's starting the call with a 300-baud beepstream or using a separate internet or modem channel to pass on the VOIP as packets rather than translating to audio. That's still not enough - your laptop or portable voip phone only knows what you've told it, and unless GPS becomes much much cheaper, lower electric power, and better at working inside, it's can't use GPS to find out for itself.
Somebody could develop standards and implementations for some kind of where-am-I beacon, which probably would be better to run on a router but could be run on a PC, which you could program with your location, so a device can check with the net to get at least some advice about where it's located physically, though obviously that information could be misadministered or forged or just blinking 12:00. And if there's more than one of them that you can see, obviously you'd want some kind of decision-making process to find the closer one....
Then there's the whole privacy issue. Usually if you're making a 911 call, you probably want the police to be able to find you. But not always, and you certainly wouldn't want them to be able to find you when you haven't asked....
In my case, I was an American attending grad school the University of California at Berkeley. My usual rant about immigration is that if some guy from Ensenada moves 500 miles to San Francisco, people get all huffy because he's a foreigner, but when I moved 2500 miles to get there from New Jersey, nobody complains, even though it's five times as far and they speak different languages there, and the only reason I needed to show citizenship documents the second time I moved here was because the governor at the time thought it was unsafe for people to drive while speaking Spanish.
In case you've never been an out-of-state college student, it costs a *lot* more than being an in-state student. State universities make a profit on this. At the time, California students didn't pay tuition at all, just a couple hundred dollars in "fees", while out-of-state tuition was close to Ivy League prices, and in New York State, out-of-state tuition at the state colleges was about twice in-state tuition, though probably about 2/3 the price of Ivy League schools. In my case, the cost wasn't a problem, because it was the late 70s high-tech boom and my employer was paying for my degree, under a program where they kicked in extra money beyond the tuition so the school would let them send people after the normal admission dates and short-cut the admissions paperwork for qualified students. (Sadly, those days are gone... They were The Phone Company, and without their scholarship program they could have easily hired about 90% of the electrical engineering masters' degree students in those years, as could several other high-tech companies.)
ICANN (and the whole political flap around them) isn't fundamentally interested in managing DNS as a technical aspect of the Internet, or the related semi-social issues of "how to fund stuff on the net" - they're interests are in the other IP, Intellectual Property, particularly in letting big trademark holders protect their big trademarks. Much of the Uniform Dispute Resolution Policy deals with how to decide who gets a name if there's a conflict about it, which is obvious enough, but you really see this in policies about "whois" registration information - instead of a pro-privacy policy, they go insisting that every domain registrant provide their True Name and True Street Address so that they can serve court papers on any registrant, and yet they don't seem to care so much if the administrative and technical and spam-abuse email addresses go somewhere that works.
In the Marks-and-Spencer.co.uk case, that was a clear attempt at ripping off somebody's name, and there was no legitimacy about it. But there are lots of cases where Example Industries in New York and Example Pizza Shoppe in California both want to be example.com, or where two different unrelated Example Pizza Shoppes in different countries want to be example.com, or where Fred Example has been using example.com for his home website since 1992 and Example MegaConglomerate wants to take the name, and one of their arguments for why it should belong to them was that Fred's web page has a picture of his dog on it and They Sell Dogfood so he's obviously trying to infringe their trademark.
And then of course there was EToys trying to steal EToy.com's web site, claiming it was trademark dilution and was a mean nasty offensive art-project site that might upset people who were looking for etoys.com, when in fact EToy.com had existed for several years before EToys was even founded as a company.
In the sex.com case, standard law has mostly covered the situation, but two of the Federal Appeals Court judges think that there may be some issues in Califonia law that affect whether Verisign is liable for damages in this case or whether they only have to give the original owner the name back and only the guy who ripped it off is liable.
Alex Kozinski's a pretty good judge, and he writes good rants - if you haven't read the case, it's a fun read. One oddity here is that the appeals court has three judges, and two of them thought that they should refer it to the California Supreme Court, because the remaining critical issue depends on California law, not Federal, and Kozinski disagreed - but because he's head judge, he has to write the opinion sending it to the California Supremes, as well as writing his dissenting opinion.
Common law on conversion originally only applied to tangible property - not just real estate, but horses and cooking pots and later cars and computers - but not to intangibles. As Werdna says, in most states that's changed. Kozinski argues, and IANAL but it looked convincing to me, that California law has really settled the issue solidly that conversion can apply to intangibles, or at least to intangibles that have some connection to tangibles like paper, so there's no issue here.
The big deal for Verisign is that conversion is a tort, and since they lost, they have to pay up, and since Cohen, who actually ripped off the domain name, has skipped out of the country and hidden or spent most of the money, they may have to pay the whole $65 million to Kremen and try to get it back from Cohen if they can. They already lost the part about giving back the domain name, but that didn't bother them so much.
The amount of money actually is a bit surprising - this case started in 1995, before the Boom, before business.com sold for $150K in 1997, and before Altavista.com sold for $3M in 1998. The plaintiff Kremen had registered the name, but hadn't actually used it for anything yet and didn't know it was gone for about 8 months. Meanwhile, Cohen, who had ripped it off, had figured out how to turn it into a highly profitable business, and after several years of litigation, he'd made a big enough pile of money on it that the court awarded the plaintiff $65M as well as giving him the name back (that was either ~$20M in actual damages and ~$40M in punitive damages, or maybe the other way around.) He and Verisign appealed, in Cohen's case partly because he said the judgement was way out of line for an asset was only worth so much because he'd built it up himself, and in September 2002, the court handed him his ass for even trying to appeal because he'd fled and hidden the money, and they told Verisign to give Kremen the name back, but the conversion issue is still not decided, so Verisign doesn't know how much of it they'll have to pay.
I doubt the legislation would pass, and particularly that it would pass in the clean, simple form he recommends without getting lots of gunk added to it. Even if it does, it won't be too effective unless the _bounty_ is available not only to Americans, but to _anybody_, anywhere in the world, who succeeds in tracking down the spammer, which I consider to be unlikely.
Some of the non-US spam you get is really sent by non-Americans, but lots of it is sent by Americans abusing non-US machines (either by abusing open relays, or by buying cheap services.) US law can't touch the non-Americans effectively, but it can touch Americans using non-US ISPs. The entertaining thing that would happen if the bill were to pass and non-Americans could collect would be an instant market in Korea and China for mail servers that simultaneously forward mail, track down the sender, and log the recipients so that they can document it for the US authorities. Pretty soon, everybody in Korea with a broadband connection (which appears to be just about everybody) will start getting email ads for servers like this, because for a little while, it'll actually be possible to M4K3 M0n3Y F4$$7 on the Net by tracking American spammers. And $10K per successful event, minus US lawyer commissions, is pretty nice for something that doesn't take too much work.
Back when they were Network Solutions, everybody always hated them. Being bought by Verisign didn't help that. By keeping the brand names and marketing separate, they can use one name for "Stuff Everybody Hates" and another for "Stuff Everybody Doesn't Hate Yet", and move products back and forth as needed.
It's about the Joneses, and keeping ahead of them
on
Moore's Law Disputed
·
· Score: 2
It's partly a prediction about the limits of technology, but partly about the nature of the business - there are a lot of Joneses out there, and you not only have to keep up with them, you have to keep ahead of them, and if you do, you can make enough money to do the next step of dealing with them.
Sometimes you make the next step by pushing the same limits you did before (either on your own or using equipment other people are making to keep up with the Joneses in their part of the industry), and sometimes by finding a different way to do things. For instance, even if you can't make transistors much smaller in two dimensions, but at some point people will figure out ways to go 3D, stacking a bunch of transistors vertically. That turns out to have annoying manufacturing and heat dissipation issues, but huge speed gains if you do it right, because distances become much shorter.
The fun change that's happened recently has been the recent jumps in disk drive price and performance - it's been more like quadrupling instead of doubling the last couple of years, and neither Bill Gates nor Linus have managed to fill it (though Linux can build file systems well enough, and Grandson-of-Napster can give you things to fill it with...)
Sometime in the mid-90s I saw a catalog advertising electrically conductive hats - really nice professional quality stuff, not just rattly tinfoil. It was one of those NewAgey things advertising some local big Wellness fair, and full ads for vitamins and healing magnets and Tachyon bracelets (I forget it they were tachyon-generating bracelets or tachyon-absorbing bracelets, but it was aimed at the kind of people for whom simple magnets just weren't high-tech-sounding enough.) The hats looked really good, sort of basic hunting cap made of aluminized cloth (unless it was really just spraypainted:-)
Why was I reading this tripe? Well, it was on the bench at the train station while I was waiting for the train, and the newspaper headlines had looked boring...
If everybody else on your street bought HomePlug data-over-electricity, you'd not only have the security issues, but also the issue of bandwidth. Is this the kind of thing that gets you 5-14Mbps if you're the only home on your block using it, but shares those same 5-14Mbps with everybody on your block, so get much lower effective bandwidth if the system becomes popular? What's the distance at which it communicates, or interferes with communication? Is this something that could wire an entire apartment building? A "typical" city block? Everybody behind a given power transformer?
I looked at some of the homeplug.org web sites and member sites like Asoka. Homeplug runs natively at 14Mbps (USB devices are limited to USB's 12 Mbps speed), though effective speeds are often lower, depending on how noisy your environment is (one site said 80% of their tests got 5Mbps or better), and it's good for up to 1km, as long as there aren't power transformers in the way. You can only put 16 devices on the network; I assume that's 16 devices per 56-bit-DES security key, but I could be wrong. That does mean that you're not going to wire everybody in your neighborhood together in the same LAN. Nothing I saw talked about the throughput effects of having your neighbors sharing the network, only the security effects.
Slashdot Mirror
The performance certification isn't stunning, either - it's nice to get 2Mbps radio, but they're only insisting on 128kbps of actual wired bandwidth behind it, so who cares how fast the radio link is? (I'd initially read it as requiring 2Mbps of _bandwidth_, which means a European E1 can almost do it, unless they use 1984kbps or 1920kbps framing, but an American/Japanese/Taiwanese T1 can't. But it's just talking about radio speed. 128kbps is just ISDN/IDSL or slow-DSL-uplink speed, though it certainly beats modems.)
The no-interference-with-VPNs part is good, though it's not clear whether it's banning NAT or not (many IPSEC implementations support UDP tunnels for NAT, but it's bandwidth-intensive and not universally supported.)
Or if it was, it at least didn't have the same cooler-than-you appeal that Graffiti did, probably because it was done by lower-paid-than-techies typists using dead trees. It did actually have some coolness factor, but that was made up for by the fact that 99% of the people using it were using it to do things for somebody else, rather than using it to do things for themselves, and it was usually followed by typing the stuff. And while some of them were secretaries, many of them were just typists.
Macworld was in SF this week, so I went there for an hour or so, which was enough :-) There was a surprising amount of furniture. I don't remember the name of it, but somebody had a line of white and chrome stuff that went with the half-sphere iMacs, which provided a bottom half sphere and some bent chrome tubes and a keyboard/mouse holder on another extension arm, kind of like having your desk replaced with a white spider.
I think Jimi actually used lighter fluid to do that, but don't try this at home...
More seriously, though, another reason to put the power button on the keyboard is so you can put the computer under the desk or somewhere else that's not very accessible, and you only need to access the machine when you're feeding it CDs (or not even then, if you've got a USB CD drive.) This also lets you make the room quieter by hiding the PC in an accoustic-shielding box, or in a closet or in the next room with long cables.
but not only _do_ they use them, replying to their beepers from phone numbers that aren't easily traced to them, but police and cities have discouraged phone companies from providing them there because they want to discourage drug dealers, and this is most common in poorer urban areas. Pay phones used to be able to support incoming calls as well, but the telephone deregulation changes in the 80s and early 90s that let them be privately operated instead of only run by telcos killed that, because private operators didn't get any revenue from receiving calls (and also, the War On Politically Incorrect Drugs also meant that drug dealers would use them to receive calls.)
A few years ago, before I got a cell phone, I was trying to hunt for an apartment or house to rent in the San Francisco Bay Area. This involved a lot of trying to contact landlords and property managers to get in to see places that were advertised, but they're never in their offices - you call their beeper or answering machine and leave your phone number. I did have a beeper, but of course with no PAY PHONES around, it was hard to call them back. In some areas, there'd either be a 7-11 or else a restaurant that had a phone in the back, so if we'd left enough calls in a given area, we'd get coffee and more quarters and wait. Really frustrating....
Of course, pay phone usage in poor areas also went down because of low-cost lifeline phone rates, and because deregulation meant that the prices of pay phone calls went way up, and in high-crime neighborhoods, a coin-operated pay phone looks a lot like a parking meter - it's a box of money sitting there for any teenager with a spare metal pipe, unless it's in a well-lit high-traffic area.
And fan comics are a subgenre of fan fiction as well; most of it that I've seen in the US has been from those evil Furries, but there's a lot of other artwork in the science fiction world.
And then there's Filk .
It's not a big deal - obviously there will be important packages that nobody's really interested in, like boring-glue-stuff.rpm. So you use the voting process to pick the popular packages, and use your engineering judgement to add the other critical stuff, but you probably don't need more than 100-200MB of mandatory stuff, plus another 100MB of packages that are required to support the packages picked by the popular vote.
That link points to http://www.mandrakelinux.com/en/ftptmp/1042225920. 78796bc94d6342c90d73c6d7e2ec5baa.php#beta :-). Alternatively, it's possible that my company's firewall is up to something.
which gets the usual Mandrake web server message about "That page is missing. Go back to the main page". Probably some sort of auto-updating thing; I got that two days ago when trying to find the mirrors page to download 9.0
There's a bit of a price difference, as well :-) In about 1984, we'd paid about $35000 for four RM05 drives, which gave us 1GB of total drive capacity, and removable 250MB disk packs were about $1100. So that was $4/MB for the media, and $35/MB for the drives. I forget what tape drives cost, but 6250bpi 9-track tapes cost about $25 and held about 160MB of data, so they were the obvious backup medium.
Well the reason Google is so important is that so many people use it, and the reason so many people use it instead of Yahoo or Altavista or Northernlights or Hotbot or LongDefunct.Com or Excite.com or Teoma or some of the other search engines out there is that they do a really excellent job. I used to use Altavista, who were not only the original big search engine, they were one of the best in terms of coverage, as opposed to Yahoo who had much better indexing but nowhere near as many pages. If you wanted to find something obscure, you'd use Altavista, but if you wanted to find something common, it might be hard because Altavista would get 50,000 references that you could look through 10 at a time. I switched to using Google because their search engine did a really good job of usually having the information I wanted in the first page or two, often in the first one or two references, as well as because their pages were lean and mean and not cluttered with dancing broken Javascript ads, and I've occasionally found the cache to be valuable for finding information that was once on the web but isn't any more.
As far as Daniel Brandt's rants about how the government ought to be regulating Google and PageRank because so many people use it, that's purely backwards. The government could accomplish any positive aspects of his goals by building their own search engines with their own page ranking algorithms, but if they go messing with Google, they're not only likely to censor some content and artificially inflate things they want to propagandize, but they're likely to make it less likely to have the material I want near the top, destroying the Pagerank in order to save it.
Some search engines have tried to make money by letting people pay for good placement - the pundits yell at them for it, and the public tends to use those engines less because they're better at finding advertising drivel than interesting content. Lots of web sites try to game the page ranking systems on all the major search engines, typically by including relevant keywords many many times in comments or meta-things, or by including them in small print at the bottom of the page, and the main reason the system doesn't get swamped by this is that the better algorithms try to detect this manipulation and neutralize it or seriously downrate for it. Otherwise the search engines would have a high proportion of uninteresting material near the top, mostly pages that are really just spam. If Google's PageRank didn't protect itself against whatever techniques SearchKing is using, he'd be doing the same thing, making it much easier to find pages people pay to promote than pages that are rated high because they're actually interesting. (I've got slightly mixed feelings about that, because his stuff seems to look less obnoxious than banner ads or dancing javascipts, and is usually on pages I don't ever read...)
Searchking's Web Site has a page of News about their lawsuit and Searchking's Comments on Google's Response to the Court and their attempt to get the public to pay $20 for more details. I wasn't that impressed with it, but maybe you'd have a different opinion....
That was a fun case; the judge was quite impressed that they'd tried a Third Amendment issue because it's basically never violated.
Cell phones, except possibly a few new ones, don't _have_ GPS-type features. GPS technology is getting smaller and lower in power consumption, but still takes too much space and power for a small phone. What actually goes on is that cell phones talk to cell antenna towers so that they've always got the best signal they can get, and this lets them locate you because (in the dumb version) the system knows what cell site you're using, and (in the newer smarter versions) the nearest N cell sites know you're nearby so they can triangulate. If you turn the phone off, and then turn it on again, it sends out a broadcast letting the system know it's there.
On the other hand, there are a number of international carriers starting to use VOIP, because it makes it convenient to use compression and save a lot of bandiwidth in spite of the IP packet overhead, plus of course the Net2Phone-type-people do VOIP.
For VOIP, though, the way the phone call reaches the phone company is that somebody has a box that translates between phone lines (usually T1 trunks) and IP addresses, so the only thing the phone company knows is that it got a call from 202-456-2121, which terminates on a box in the basement at 1600 Pennsylvania Ave. PBXs already cause some trouble with this, but they're often bright enough to know that the call is from extension 1234, and somebody can run a database that knows that x1234 is on the 12th floor. With VOIP, it's even worse - the VOIP-to-POTS gateway is some router or PBX that knows it got a connection from 10.32.11.1, and it's possible that somebody has a database that shows that that address belongs to a DHCP server on the 12th floor, not that the 911-police know how to find your data network management staff in a hurry, but in most of the VOIP standards, there's really no information beyond the IP address.
And of course your IP address might be anywhere in the world - did you dial in today from home, or a hotel, or an airplane, or your corporate office 3000 miles away? And think about the VOIP phones themselves - some people use telephone-oriented software applications on their PCs (it's 10.01.01.23 - do you know where your laptop is?), while other people use desktop VOIP phones from a variety of vendors, which you program to know that you're Linus Torvalds on +1-202-456-2121, and if you plug them in anywhere in your company's network, they'll find the gateway server, let it know your current IP address, and be ready to pick up your voicemail and incoming phone calls. Anywhere. So if you dial 911, the town your company's main office is in knows that there's an emergency somewhere near you, but it doesn't know where you are. And if you're not using a coporate VOIP system, you could really be almost anywhere. And if you're going through a NAT firewall or VPN gateway, you could be even farther anywhere.
So what kinds of approaches can people take to fix this? The two obvious first steps are either to get the phone company out of the way (give the 911 people VOIP so they can at least try to traceroute you, though that still has all the IP-vs-location uncertainties), or else to make sure that the VOIP standards are updated to do a better job of passing location information (for people who want to pass it) and that the VOIP-to-POTS gateway standards provide some mechanism for passing that along, whether it's starting the call with a 300-baud beepstream or using a separate internet or modem channel to pass on the VOIP as packets rather than translating to audio. That's still not enough - your laptop or portable voip phone only knows what you've told it, and unless GPS becomes much much cheaper, lower electric power, and better at working inside, it's can't use GPS to find out for itself.
Somebody could develop standards and implementations for some kind of where-am-I beacon, which probably would be better to run on a router but could be run on a PC, which you could program with your location, so a device can check with the net to get at least some advice about where it's located physically, though obviously that information could be misadministered or forged or just blinking 12:00. And if there's more than one of them that you can see, obviously you'd want some kind of decision-making process to find the closer one....
Then there's the whole privacy issue. Usually if you're making a 911 call, you probably want the police to be able to find you. But not always, and you certainly wouldn't want them to be able to find you when you haven't asked....
In case you've never been an out-of-state college student, it costs a *lot* more than being an in-state student. State universities make a profit on this. At the time, California students didn't pay tuition at all, just a couple hundred dollars in "fees", while out-of-state tuition was close to Ivy League prices, and in New York State, out-of-state tuition at the state colleges was about twice in-state tuition, though probably about 2/3 the price of Ivy League schools. In my case, the cost wasn't a problem, because it was the late 70s high-tech boom and my employer was paying for my degree, under a program where they kicked in extra money beyond the tuition so the school would let them send people after the normal admission dates and short-cut the admissions paperwork for qualified students. (Sadly, those days are gone... They were The Phone Company, and without their scholarship program they could have easily hired about 90% of the electrical engineering masters' degree students in those years, as could several other high-tech companies.)
In the Marks-and-Spencer.co.uk case, that was a clear attempt at ripping off somebody's name, and there was no legitimacy about it. But there are lots of cases where Example Industries in New York and Example Pizza Shoppe in California both want to be example.com, or where two different unrelated Example Pizza Shoppes in different countries want to be example.com, or where Fred Example has been using example.com for his home website since 1992 and Example MegaConglomerate wants to take the name, and one of their arguments for why it should belong to them was that Fred's web page has a picture of his dog on it and They Sell Dogfood so he's obviously trying to infringe their trademark.
And then of course there was EToys trying to steal EToy.com's web site, claiming it was trademark dilution and was a mean nasty offensive art-project site that might upset people who were looking for etoys.com, when in fact EToy.com had existed for several years before EToys was even founded as a company.
In the sex.com case, standard law has mostly covered the situation, but two of the Federal Appeals Court judges think that there may be some issues in Califonia law that affect whether Verisign is liable for damages in this case or whether they only have to give the original owner the name back and only the guy who ripped it off is liable.
Common law on conversion originally only applied to tangible property - not just real estate, but horses and cooking pots and later cars and computers - but not to intangibles. As Werdna says, in most states that's changed. Kozinski argues, and IANAL but it looked convincing to me, that California law has really settled the issue solidly that conversion can apply to intangibles, or at least to intangibles that have some connection to tangibles like paper, so there's no issue here.
The big deal for Verisign is that conversion is a tort, and since they lost, they have to pay up, and since Cohen, who actually ripped off the domain name, has skipped out of the country and hidden or spent most of the money, they may have to pay the whole $65 million to Kremen and try to get it back from Cohen if they can. They already lost the part about giving back the domain name, but that didn't bother them so much.
The amount of money actually is a bit surprising - this case started in 1995, before the Boom, before business.com sold for $150K in 1997, and before Altavista.com sold for $3M in 1998. The plaintiff Kremen had registered the name, but hadn't actually used it for anything yet and didn't know it was gone for about 8 months. Meanwhile, Cohen, who had ripped it off, had figured out how to turn it into a highly profitable business, and after several years of litigation, he'd made a big enough pile of money on it that the court awarded the plaintiff $65M as well as giving him the name back (that was either ~$20M in actual damages and ~$40M in punitive damages, or maybe the other way around.) He and Verisign appealed, in Cohen's case partly because he said the judgement was way out of line for an asset was only worth so much because he'd built it up himself, and in September 2002, the court handed him his ass for even trying to appeal because he'd fled and hidden the money, and they told Verisign to give Kremen the name back, but the conversion issue is still not decided, so Verisign doesn't know how much of it they'll have to pay.
Some of the non-US spam you get is really sent by non-Americans, but lots of it is sent by Americans abusing non-US machines (either by abusing open relays, or by buying cheap services.) US law can't touch the non-Americans effectively, but it can touch Americans using non-US ISPs. The entertaining thing that would happen if the bill were to pass and non-Americans could collect would be an instant market in Korea and China for mail servers that simultaneously forward mail, track down the sender, and log the recipients so that they can document it for the US authorities. Pretty soon, everybody in Korea with a broadband connection (which appears to be just about everybody) will start getting email ads for servers like this, because for a little while, it'll actually be possible to M4K3 M0n3Y F4$$7 on the Net by tracking American spammers. And $10K per successful event, minus US lawyer commissions, is pretty nice for something that doesn't take too much work.
Back when they were Network Solutions, everybody always hated them. Being bought by Verisign didn't help that. By keeping the brand names and marketing separate, they can use one name for "Stuff Everybody Hates" and another for "Stuff Everybody Doesn't Hate Yet", and move products back and forth as needed.
Sometimes you make the next step by pushing the same limits you did before (either on your own or using equipment other people are making to keep up with the Joneses in their part of the industry), and sometimes by finding a different way to do things. For instance, even if you can't make transistors much smaller in two dimensions, but at some point people will figure out ways to go 3D, stacking a bunch of transistors vertically. That turns out to have annoying manufacturing and heat dissipation issues, but huge speed gains if you do it right, because distances become much shorter.
The fun change that's happened recently has been the recent jumps in disk drive price and performance - it's been more like quadrupling instead of doubling the last couple of years, and neither Bill Gates nor Linus have managed to fill it (though Linux can build file systems well enough, and Grandson-of-Napster can give you things to fill it with...)
Why was I reading this tripe? Well, it was on the bench at the train station while I was waiting for the train, and the newspaper headlines had looked boring...
I looked at some of the homeplug.org web sites and member sites like Asoka.
Homeplug runs natively at 14Mbps (USB devices are limited to USB's 12 Mbps speed), though effective speeds are often lower, depending on how noisy your environment is (one site said 80% of their tests got 5Mbps or better), and it's good for up to 1km, as long as there aren't power transformers in the way. You can only put 16 devices on the network; I assume that's 16 devices per 56-bit-DES security key, but I could be wrong. That does mean that you're not going to wire everybody in your neighborhood together in the same LAN. Nothing I saw talked about the throughput effects of having your neighbors sharing the network, only the security effects.