There are several calendaring systems in the open-source world, either as open source applications or as commercially supported products that work well along-side open-source mail systems. Calendar features don't have to be directly integrated into a mail server or mail client to be useful, especially if the mail client can hand URLs to a web browser by clicking on them (which almost anything more GUI than pine can do.)
As far as "why you'd implement them using open source", the answers include "one size doesn't fit all" and "so you're not stuck with the annoying limitations of Outlook" or "so you can customize behaviour or fix things". Those are all reasons for using standards-based applications instead of proprietary.
My experience with the MS Outlook calendar functions are that they're *much* more useful for people who only use always-connected desktop computers than for laptop users who might or might not be connected to a work or home network and might be carrying the machine around. Outlook is a lot more reliable than it was a couple of years ago, but it's still not totally happy with that environment. It's what I use at work, but otherwise I'd simply use my Palm's calendaring.
Shared Folders are mostly a hack to make groupware fit into an email-like environment. If I wanted a hack like that (:-) I'd use IMAP, and some open-source IMAP server to do that. Otherwise, I'd use Web-based discussion boards - there are a wide range of them, open-source and commercial. Or you could even use Usenet technology, and you'll find that Netscape and Mozilla are friendly clients that look much like their email interfaces, plus there are others that are more customized for usenetting.
There are a whole range of things that can go wrong with networks, and part of Business Continuity Planning is being paranoid about the right problems. Too many East Coast companies got surprised by 9/11 - here on the West Coast we have buildings fall over every decade or so, though fortunately the loss of life has been kept low by good building technology. You do have to spend a lot of time looking at what your critical resources are, including your people, your computers, your data, your telecomm networks, your phone system, etc.
On the other hand, we're not so good at planning for the "ice storm kept everybody home" problem. I have once had a meeting cancelled because of snow in Silicon Valley - the critical customer lived in Santa Cruz, and there were six inches of snow on the mountain top, on a road which is unpredictable in good weather.
Fiber has lots more bandwidth than wireless, and in parts of the network this is critical. It's also much more reliable on a bit-error-rate basis. There are lots of good reasons that telcos mostly dumped Microwave a decade ago. A medium-end point-to-point radio-based system can get an OC3 (155 Mbps bidirectional) a few kilometers, typically 5-10 depending on how rainy your climate is. You can put multiples of these in an area, but basically you get short hops and it's hard to get more than a gigabit per second or so. High-end systems with big dishes instead of little ones can get a T3 a lot farther, but still less than 50-100km. By contrast, a wavelength on a fiber pair can run gigabit ethernet, or OC48 (2.5Gbps) or OC192 (10Gbps), and you can put lots of wavelengths on a fiber, and lots of fibers in a bundle, and run them 100km or more without repeaters. Radio has some advantages for some applications, but mainly lower capacity short-distance connections. Free-space optics shows some promise for fatter short-haul connections, but it's not perfect either.
Network backbones really need fiber or something like it. You can't run a modern telco without fiber, and there are huge operational costs wins by using fat pipes and routing or switching rather than using very large numbers of small pipes and patching the things together and trying to keep them load-balanced.
Edge distribution is an area where wireless *can* replace some traditional services. Telephony-speed wiring to your house could do fine over radio - a big reason we didn't develop that technology in the US is that the Roosevelt Administration's pro-monopoly policies (yes, I really mean that!) gave radio to the broadcast entertainment industry and telephony to the phone companies and prevented anybody from competing in the others' space, but now that we're mostly over that, it may be possible to adapt wireless to replace some of the traditional wired applications. On the other hand, it'll be tough to get to DSL-like speeds without a reasonably amount of distribution networks - you can replace the last mile, but the last 10 miles is harder because of aggregate bandwidth demands.
Reliability is another huge issue - if you get very much packet loss, TCP spends all its time retransmitting and can't reach high throughputs, and it becomes tempting to reinvent X.25-like link retransmissions because radio acts too much like barbed-wire if you're not careful. Rain fade is a real problem - there are technologies that work really well in Phoenix that you simply can't sell to Microsoft because Seattle is too rainly.
It's been a long time since I studied German in school.... Modern German printing type resembles American or British, but the older Fraktur type had a more decorative ornate letter forms. It probably started getting replaced around 1900, plus or minus 50 years. This change also shows up in cursive handwriting - textbooks on German from about ~1900 still show the Fraktur forms. It's also somewhat visible in older English writing - The most notable example is lower-case F which looks like a tall narrow S.
So I don't find it surprising that this shows up in the Germanic Scandinavian languages as well, though I had no guesses what the Finns did.
Australia's no farther from civilization than Hong Kong or Korea or Singapore. OK, it's perhaps a bit farther from the US than Hong Kong, but some of the cable routes from the US to SG go through Australia rather than HK. There are some differences - being English-speaking means their traffic is largely outside-world-focused, as opposed to Korea, where most people are getting content from Korea or Japan. But Hong Kong and Singapore don't have that problem. One of the big reasons Australians aren't getting lots of Australian content is that Telstra's pricing policies make it unattractive to run home web servers there.
The problem is that Telstra has always been the First World's most clueless-about-data telco. Yes, India's VSNL has always been worse, but they have real infrastructure and economic difficulties as well as being clueless, as do places like Thailand or Mongolia. Australia doesn't have that excuse. The data cap business is set at such a low level that it also kills couch-potato applications like downloading videos, as well as killing web servers. And worse, US cable companies are starting to pick up on the broadband-cap idea - many of them have also been suicidally clueless, but at least they had competition, and there were stupid things they hadn't thought of.
Back in 1990, Telstra was ravingly clueless about ISDN, worse than 90% of the US local telcos - they thought that the only reason to use a PRI was to get 30 B channels of small connections from BRIs, and weren't useful about that; friends of mine wanted to use PRIs to dial bigger connections (e.g. 384 video or fatter data connections), but Telstra not only didn't support it, they didn't know why anyone might want such a thing. They haven't acquired any new clues since then, as near as I can tell.
A bit earlier than that, friends of mine worked at a database company that was bought by a large company that buys other companies and drains the money out of them, killing off most of the development and milking the product for what it's worth as a maintenance product. The buyer basically gave them a "Be our slaves or leave by the end of the week" offer. So they pretty much all left. One event that happened in the process was that there was a list of a dozen or so key developers that they didn't want to lose, and somebody "found" it. So those dozen developers walked in together to all resign.
This was after the late-80s boom and the early-90s slump, but before the internet boom made Silicon Valley totally silly. It was about as close to a normal market as things get around here. Most of them did pretty well.
Sure, it's not the same everywhere, but go read "The Firm" by John Grisham. (no, no, no, don't see the lame-oid Tom Cruise movie about it, read the *book*, or listen to the audio book...)
Lawyers who are in their own practices can control their hours like most other professionals, and lawyers for governments or corporations generally get treated like other employees there, but junior partners in law firms pretty much get treated like harassed overworked programmers except they've also got to wear suits and aren't supposed to shoot each other with Nerf-foam rockets over their cubicle tops.
China actually has all the space they need for now, because their censorship-happy government and several quasi-monopolistic telecom providers have kept a pretty tight control on the internet's growth there. The "Great Firewall of China" that enforces web and email censorship can keep most internet users (particularly home and small business users) behind NAT or make them use IPv6 space or whatever, and most of the people who need real Internet access are businesses that don't need much space for the outside of their firewalls, which can be efficiently aggregated by the small number of ISPs.
Japan and especially Korea are more interesting cases, because they don't have the censorship problem, they've got a much much higher fraction of their population wired, and their telecom infrastructure is much more liberalized. And besides, you don't have to sell spammers Korean address space to M4K3 M0N3Y Fa$$T!! - you can sell them lists of broken relays and proxies:-)
Public Key signatures are really helpful for some applications, and they do cut out lots of the anklebiters, which is worthwhile in itself, though almost any password mechanism can do that. They're not a 100% solution, though - anybody who knows the password or private key can still steal the address space, like that disgruntled ex-sysadmin, and trust in a public-key mechanism is likely to reduce the amount of human oversight that goes along with accepting requests, making it easier for someone who has the key to social-engineer changes, especially things like "loans" of subnets of your address space that fell off a truck.
They also don't help the old-defunct-company problem - any address-space owner who didn't have a public key N years ago and isn't easy to find now can still be hijacked with the fake letterhead request for a public key, which is now the obvious first step before using the fake letterhead to social engineer the ISP. Pretending to own a Class A owner is hard to fake credibly - pretending to own a Class B or/19 owner is a lot easier.
I've had one friend of a friend who at least temporarily was the last-registered technical and administrative contact for a Class B that was the remains of a long-defunct technology company, and they were thinking about selling it on the legitimate market (not the spammer market), but decided that their chain of ownership through the various bankruptcy settlements was too dubious - I forget whether the space eventually got recycled by ARIN or whether the somewhat more legitimate owners of the remaining assets got it.
There are probably three main reasons hijackers steal addresses like this
To get a big space free/cheap, given IPv4 address space's semi-artificial scarcity. IPv6 really takes care of this - a/48 is big enough for almost anybody, and a/64 is enough for almost any subnet - the 2**64 addresses you get in your/64 let you use 48-bit MAC addresses to automatically address everything and still give you 16 more bits to play with.
To target a specific address space owner for nefarious purposes. Yeah, fine, IPv6 isn't going to prevent somebody who wants to hijack Bill Gates's House's IP address or remap all of Korea's IPv4 address space through Spamcop's T1. That's a problem for other mechanisms.
To imitate somebody random other than yourself to make tracking you down or blocking your resources harder. IPv6 isn't much help for that, but that's also a case where hijacking subnets can be more fun than hijacking whole networks (e.g. don't steal the whole/16, just announce some/19 or/24 subnets that they weren't using.)
There may be occasional games that you can play where hijacking the whole/16 is useful, e.g. doing a more credible imitation of a mis-sized ISP with a customer who's a victim of a spammer relay trojan or something, but mostly that's pretty far-edge cases.
It's really an early-adopter bias, from back when 32 bits was enough for everybody, especially because Internet-connected computers were big things that supported lots of users per machine, not PCs on home networks or PDAs and cellphones on Personal Area Networks.
There weren't firewalls or NATs to prevent local machines' addresses from being reachable by the Whole Internet, and
there wasn't RFC1918 private address space until after the ARPANET was shut down, and
Networks were always Class A, B, or C, and even if they were subnetted, it was still on class boundaries, and
supernetting and CIDR didn't exist.
The Class A allocations are basically a pile of dinosaur bones, and most of the dinosaurs were either native to North America or else ate other dinosaurs that were.
But yes, the early-adopter bias is a US bias, because before the work of people like CIX, the Commercial Internet Exchange, the ARPANET was a thing run by the US government, and you could only get on it if you were a US defense contractor doing appropriate kinds of work or a University that had some appropriate government-funded research, and there was an Acceptable Use Policy that said you couldn't do commercial activities that weren't related to the Government Work you were doing (though much of the interestingness of the Internet culture evolved because there was deliberately slack enforcement, especially on universities and non-commercial-related discussions.) The rest of us had UUCP, and Usenet, and X.25, and it wasn't until ~1990 that you could reliably use email for outside-your-company business without having to worry about whether you were violating the AUP.
Currently? Looks like Stanford gave theirs back in ~2000. About 60% of the Class A space is unused now.
AT&T and BBN are ISPs, so they've got legitimate uses for large amounts of address space. (In AT&T's case, they got lucky, because while they were late getting into the ISP business, the Class A was a leftover from the Bell Labs Cray's Hyperchannel LAN, which for some reason had insisted on having a Class A network and couldn't be subnetted:-)
The Interop Show Network has always been special. For you young folks out there (:-), Interop used to be an engineering conference where vendors actually tested interoperability and worked on implementation bugs, as opposed to being primarily marketing-related, and back in ~1990, not everything knew how to do variable-length subnetting or CIDR or whatever, and the show needed real internet addresses, not just RFC1918, because it was connected to the Real Internet.
Auto companies have been an early developer of networking technology - there was all that ISO MAP/TOP stuff in the Mid-80s, and they were one of the big players in getting IPSEC to be a practical technology where equipment from multiple vendors actually interoperated as opposed to a custom thing for spooks and occasional banks. (That also affected the Crypto Export Regulations Wars of the 90s.) At least in the US, automobile manufacturing isn't really done by big monolithic integrated companies which could use 10.x intranets - it's done by a wide mesh of manufacturers of parts, subassemblies, components, random little job shops, etc., as well as the big companies that stamp out metal and assemble it into cars, rather like the computer and software industry except with a lot more metal shipped around, and they need registered address space to be able to talk to each other cleanly. I'm not sure that Mercedes needs all that space, but the industry certainly does.
As of December 2001, the biggest hog of Class A addresses was the US government, including the military and its friends like Halliburton. Also Eli Lilly had a Class A then...
Memory works in funny ways - I thought the following reference was about the EFF, but apparently it was an EFF person referring to the CIX, but in any case:
One of Jon Bentley's best pieces of advice about "efficiency" is to avoid premature optimization - don't try to speed things up if they're not the slow part. The performance bottleneck for bittorrent isn't how much CPU it burns keeping track of the parts people have downloaded, it's the amount of bandwidth it spends carrying the messages between clients and the tracker. Don't try to fix the "python is slow" problem until you know that a) you've got CPU constraints and b) python is implementing the critical functions inefficiently and c) you can speed the system up a lot.
In particular, Bram said in his responses that the current bandwidth ratio is 1000:1, and that he might be able to push it to 10000:1 without fundamentally disabling critical functions. Is there a CPU bottleneck now on a fast server? Would supporting 10x as many users by buying a faster pipe for the torrent server without changing the software make CPU the bottleneck? Would supporting 10 times as many users by changing data structures or algorithms like that increase the CPU load by 10x? 2x? 20x? Decrease to 0.8x? (Remember that sending fewer or smaller messages often means using less CPU to manage them, though sometimes it means more CPU to handle bit-twiddly compression.) If you want the thing changed around, those are directions to look before you start programming.
Also, if you want to change the bittorrent environment by *putting* the tracker for some interesting product on a big server, e.g. because you're Red Hat or Some Big Music/Movie Company, you also want to check these things out.
*Real* Karma mostly works by rewarding you in your next trip around the wheel, or by letting you get off and not have to reincarnate again. Bram would like to get rewarded now, not after he's dead:-)
On the other hand, if you interpret "the wheel" as merely the computer industry boom-and-bust cycle, next time will do, but this time would still be better. And the industry version of nerdvana is to sell your business or cash out your stock options during the boom and retire at 35 or 23 as a Mozillionaire; if you feel like coming back as an avatar, that either means doing yet another startup when you've already made your money, or else only showing up as MMORPG characters.
"P2P" doesn't mean "Illegal Filesharing", even if that's a popular application. P2P means a system that works based on Peer-to-Peer communications. That's in fact how BitTorrent works, and it's the right thing to call it. You're mixing up the applications of some P2P technologies with the technology itself, though, sure, some of them were designed to make illegal file sharing convenient.
Hi, Bram - What scaling lessons did you learn from the Matrix ReDownloaded? How well did BT work, what broke, do the things that broke look fixable? Or did the DDOS attacks on the main trackers get in the way of learning much? How many simultaneous downloads were happening?
(Note to Slashdot readers:) Bram's done a lot of work on testing BitTorrent with different levels of scaling as he developed it, and that's led to tweaking a lot of the algorithms and parameters to make it scale well and distribute files efficiently. But there's only some much you can fake before you need to try it with real users. The jam band music distributions and Linux ISO distributions have given some good advice, but there's always another order of magnitude possible in the number of people using something simultaneously that might break it.
In "The Crying of Lot 49", which is a nice short fast spacy read, there's a plot thread about competing mail services and a conspiracy that conducts its private communications in a way that, if you refer to the name of the product as "waste" rather than "W A S T E", indicates you're clearly not part of their group. There are also email systems called "Trystero" for similar reasons, and it makes looking at post office boxes in Scandinavia quite silly even without sampling the local agricultural products.
I loved the Thomas Covenant series - some of the best fiction in any genre since Tolkien or at least Dune. He's also got Daughter of Regals, a collection of shorter stories. One's a Thomas Covenant outtake, but they basically stand alone, and the title story rocks. His Mirror series was also very good. (Especially with Thomas Covenant, you may need to keep your dictionary handy, because he unfortunately overused his obscure-words thesaurus while writing, but it's the characters and emotional states and world-creation that give it the depth and beauty it has.)
But, ummm, how shall I say this, the Gap series sucked rocks. Yes, fine, somebody has to take the Ring of the Niebelungen and do something deeper than Tom Holt's lightweight cheerful "Expecting Someone Taller", like transforming it into Space Opera, but for me, this one just failed badly, and the last book or two of it I bought in unfulfilled hopes that he'd finally get somewhere worthwhile with it.
I didn't find anything deep in Feist's work - it struck me as a commercial "I'm going to write a series of books in the Fantasy genre that sells well" effort, without the vast deepness of Tolkein or the grimness of George Were-his-middle-initials-really-R-R Martin.
On the other hand, that's ok if it's done well and you're not expecting more of it, and Feist's books really were done well. He's telling a bunch of stories, and telling them well, does an OK job with his characters and the worlds he sets them in and does a good job with the pacing, and does a good enough job of tying the books together. If you're going for basic summer read rather than a deeper literary experience, try the first one or two and see if you want to pick up the rest. It's read-it-once stuff, but it holds up well enough for that.
I'd recommend Steven Brust's books instead, or Donaldson's Thomas Covenant books if they work for you. For Brust, you REALLY REALLY need to start them with "Jhereg", and after that the order's less critical, except that you need to read Athyra before Orca, and The Phoenix Guards before Five Hundred Years After before The Viscount of Adrilankha.
Krakauer's Into Thin Air and 1996 Everest
on
A Good Summer Read?
·
· Score: 1
I haven't read it yet, but there are a bunch of books about the disastrous 1996 Everest climbing season. "The Climb" by Anatoli Boukreev and G. Weston DeWalt is another view of the same events - apparently Krakauer thinks Boukreev made some of the most critical mistakes on the mountain, while Boukreev thought other things led to it, particularly communications problems and oxygen provisioning problems. (The book was posthumous on Boukreev's part - he got killed in an avalanche on Annapurna a year or so later.)
David Brashears's book High Exposure is partly about that year, and partly about his experiences as a climber in general. He led the IMAX film-making expedition that was on Everest that year.
Sure, back in the dark ages, Michael Moorcock did the Elric fantasies, with a moody swordsman with a really nasty haunted sword, and did the "Dancers At the End Of Time" series which was much more interesting, and the didn't work-for-me mixture of the two in "Elric at the end of time". But he's also done some non-fantasy stuff set in the current time with modern human beings, such as King of the City, which is a totally different kind of thing and worth reading.
Slashdot Mirror
Better to do this with corn-based materials than cardboard, I suppose :-)
As far as "why you'd implement them using open source", the answers include "one size doesn't fit all" and "so you're not stuck with the annoying limitations of Outlook" or "so you can customize behaviour or fix things". Those are all reasons for using standards-based applications instead of proprietary.
My experience with the MS Outlook calendar functions are that they're *much* more useful for people who only use always-connected desktop computers than for laptop users who might or might not be connected to a work or home network and might be carrying the machine around. Outlook is a lot more reliable than it was a couple of years ago, but it's still not totally happy with that environment. It's what I use at work, but otherwise I'd simply use my Palm's calendaring.
Shared Folders are mostly a hack to make groupware fit into an email-like environment. If I wanted a hack like that (:-) I'd use IMAP, and some open-source IMAP server to do that. Otherwise, I'd use Web-based discussion boards - there are a wide range of them, open-source and commercial. Or you could even use Usenet technology, and you'll find that Netscape and Mozilla are friendly clients that look much like their email interfaces, plus there are others that are more customized for usenetting.
On the other hand, we're not so good at planning for the "ice storm kept everybody home" problem. I have once had a meeting cancelled because of snow in Silicon Valley - the critical customer lived in Santa Cruz, and there were six inches of snow on the mountain top, on a road which is unpredictable in good weather.
Network backbones really need fiber or something like it. You can't run a modern telco without fiber, and there are huge operational costs wins by using fat pipes and routing or switching rather than using very large numbers of small pipes and patching the things together and trying to keep them load-balanced.
Edge distribution is an area where wireless *can* replace some traditional services. Telephony-speed wiring to your house could do fine over radio - a big reason we didn't develop that technology in the US is that the Roosevelt Administration's pro-monopoly policies (yes, I really mean that!) gave radio to the broadcast entertainment industry and telephony to the phone companies and prevented anybody from competing in the others' space, but now that we're mostly over that, it may be possible to adapt wireless to replace some of the traditional wired applications. On the other hand, it'll be tough to get to DSL-like speeds without a reasonably amount of distribution networks - you can replace the last mile, but the last 10 miles is harder because of aggregate bandwidth demands.
Reliability is another huge issue - if you get very much packet loss, TCP spends all its time retransmitting and can't reach high throughputs, and it becomes tempting to reinvent X.25-like link retransmissions because radio acts too much like barbed-wire if you're not careful. Rain fade is a real problem - there are technologies that work really well in Phoenix that you simply can't sell to Microsoft because Seattle is too rainly.
So I don't find it surprising that this shows up in the Germanic Scandinavian languages as well, though I had no guesses what the Finns did.
The problem is that Telstra has always been the First World's most clueless-about-data telco. Yes, India's VSNL has always been worse, but they have real infrastructure and economic difficulties as well as being clueless, as do places like Thailand or Mongolia. Australia doesn't have that excuse. The data cap business is set at such a low level that it also kills couch-potato applications like downloading videos, as well as killing web servers. And worse, US cable companies are starting to pick up on the broadband-cap idea - many of them have also been suicidally clueless, but at least they had competition, and there were stupid things they hadn't thought of.
Back in 1990, Telstra was ravingly clueless about ISDN, worse than 90% of the US local telcos - they thought that the only reason to use a PRI was to get 30 B channels of small connections from BRIs, and weren't useful about that; friends of mine wanted to use PRIs to dial bigger connections (e.g. 384 video or fatter data connections), but Telstra not only didn't support it, they didn't know why anyone might want such a thing. They haven't acquired any new clues since then, as near as I can tell.
This was after the late-80s boom and the early-90s slump, but before the internet boom made Silicon Valley totally silly. It was about as close to a normal market as things get around here. Most of them did pretty well.
Lawyers who are in their own practices can control their hours like most other professionals, and lawyers for governments or corporations generally get treated like other employees there, but junior partners in law firms pretty much get treated like harassed overworked programmers except they've also got to wear suits and aren't supposed to shoot each other with Nerf-foam rockets over their cubicle tops.
Surely you'll want to *transmit* those clay tablets, won't you?
Better hook up with somebody whose college also had an Ag School....
Japan and especially Korea are more interesting cases, because they don't have the censorship problem, they've got a much much higher fraction of their population wired, and their telecom infrastructure is much more liberalized. And besides, you don't have to sell spammers Korean address space to M4K3 M0N3Y Fa$$T!! - you can sell them lists of broken relays and proxies :-)
They also don't help the old-defunct-company problem - any address-space owner who didn't have a public key N years ago and isn't easy to find now can still be hijacked with the fake letterhead request for a public key, which is now the obvious first step before using the fake letterhead to social engineer the ISP. Pretending to own a Class A owner is hard to fake credibly - pretending to own a Class B or /19 owner is a lot easier.
I've had one friend of a friend who at least temporarily was the last-registered technical and administrative contact for a Class B that was the remains of a long-defunct technology company, and they were thinking about selling it on the legitimate market (not the spammer market), but decided that their chain of ownership through the various bankruptcy settlements was too dubious - I forget whether the space eventually got recycled by ARIN or whether the somewhat more legitimate owners of the remaining assets got it.
- To get a big space free/cheap, given IPv4 address space's semi-artificial scarcity. IPv6 really takes care of this - a
/48 is big enough for almost anybody, and a /64 is enough for almost any subnet - the 2**64 addresses you get in your /64 let you use 48-bit MAC addresses to automatically address everything and still give you 16 more bits to play with. - To target a specific address space owner for nefarious purposes. Yeah, fine, IPv6 isn't going to prevent somebody who wants to hijack Bill Gates's House's IP address or remap all of Korea's IPv4 address space through Spamcop's T1. That's a problem for other mechanisms.
- To imitate somebody random other than yourself to make tracking you down or blocking your resources harder. IPv6 isn't much help for that, but that's also a case where hijacking subnets can be more fun than hijacking whole networks (e.g. don't steal the whole
/16, just announce some /19 or /24 subnets that they weren't using.)
There may be occasional games that you can play where hijacking the whole- There weren't firewalls or NATs to prevent local machines' addresses from being reachable by the Whole Internet, and
- there wasn't RFC1918 private address space until after the ARPANET was shut down, and
- Networks were always Class A, B, or C, and even if they were subnetted, it was still on class boundaries, and
- supernetting and CIDR didn't exist.
The Class A allocations are basically a pile of dinosaur bones, and most of the dinosaurs were either native to North America or else ate other dinosaurs that were.But yes, the early-adopter bias is a US bias, because before the work of people like CIX, the Commercial Internet Exchange, the ARPANET was a thing run by the US government, and you could only get on it if you were a US defense contractor doing appropriate kinds of work or a University that had some appropriate government-funded research, and there was an Acceptable Use Policy that said you couldn't do commercial activities that weren't related to the Government Work you were doing (though much of the interestingness of the Internet culture evolved because there was deliberately slack enforcement, especially on universities and non-commercial-related discussions.) The rest of us had UUCP, and Usenet, and X.25, and it wasn't until ~1990 that you could reliably use email for outside-your-company business without having to worry about whether you were violating the AUP.
AT&T and BBN are ISPs, so they've got legitimate uses for large amounts of address space. (In AT&T's case, they got lucky, because while they were late getting into the ISP business, the Class A was a leftover from the Bell Labs Cray's Hyperchannel LAN, which for some reason had insisted on having a Class A network and couldn't be subnetted
The Interop Show Network has always been special. For you young folks out there (:-), Interop used to be an engineering conference where vendors actually tested interoperability and worked on implementation bugs, as opposed to being primarily marketing-related, and back in ~1990, not everything knew how to do variable-length subnetting or CIDR or whatever, and the show needed real internet addresses, not just RFC1918, because it was connected to the Real Internet.
Auto companies have been an early developer of networking technology - there was all that ISO MAP/TOP stuff in the Mid-80s, and they were one of the big players in getting IPSEC to be a practical technology where equipment from multiple vendors actually interoperated as opposed to a custom thing for spooks and occasional banks. (That also affected the Crypto Export Regulations Wars of the 90s.) At least in the US, automobile manufacturing isn't really done by big monolithic integrated companies which could use 10.x intranets - it's done by a wide mesh of manufacturers of parts, subassemblies, components, random little job shops, etc., as well as the big companies that stamp out metal and assemble it into cars, rather like the computer and software industry except with a lot more metal shipped around, and they need registered address space to be able to talk to each other cleanly. I'm not sure that Mercedes needs all that space, but the industry certainly does.
As of December 2001, the biggest hog of Class A addresses was the US government, including the military and its friends like Halliburton. Also Eli Lilly had a Class A then...
[They] "might be a conspiracy of lizard-like aliens here to steal our water, but I doubt it."
In case you're too young to recognize the Subject line...
Does anybody still have a copy of the 4.2 > V poster? It showed a "4.2BSD" tie fighter speeding away from an exploding AT&T Death Star.
In particular, Bram said in his responses that the current bandwidth ratio is 1000:1, and that he might be able to push it to 10000:1 without fundamentally disabling critical functions. Is there a CPU bottleneck now on a fast server? Would supporting 10x as many users by buying a faster pipe for the torrent server without changing the software make CPU the bottleneck? Would supporting 10 times as many users by changing data structures or algorithms like that increase the CPU load by 10x? 2x? 20x? Decrease to 0.8x? (Remember that sending fewer or smaller messages often means using less CPU to manage them, though sometimes it means more CPU to handle bit-twiddly compression.) If you want the thing changed around, those are directions to look before you start programming.
Also, if you want to change the bittorrent environment by *putting* the tracker for some interesting product on a big server, e.g. because you're Red Hat or Some Big Music/Movie Company, you also want to check these things out.
On the other hand, if you interpret "the wheel" as merely the computer industry boom-and-bust cycle, next time will do, but this time would still be better. And the industry version of nerdvana is to sell your business or cash out your stock options during the boom and retire at 35 or 23 as a Mozillionaire; if you feel like coming back as an avatar, that either means doing yet another startup when you've already made your money, or else only showing up as MMORPG characters.
"P2P" doesn't mean "Illegal Filesharing", even if that's a popular application. P2P means a system that works based on Peer-to-Peer communications. That's in fact how BitTorrent works, and it's the right thing to call it. You're mixing up the applications of some P2P technologies with the technology itself, though, sure, some of them were designed to make illegal file sharing convenient.
(Note to Slashdot readers:) Bram's done a lot of work on testing BitTorrent with different levels of scaling as he developed it, and that's led to tweaking a lot of the algorithms and parameters to make it scale well and distribute files efficiently. But there's only some much you can fake before you need to try it with real users. The jam band music distributions and Linux ISO distributions have given some good advice, but there's always another order of magnitude possible in the number of people using something simultaneously that might break it.
In "The Crying of Lot 49", which is a nice short fast spacy read, there's a plot thread about competing mail services and a conspiracy that conducts its private communications in a way that, if you refer to the name of the product as "waste" rather than "W A S T E", indicates you're clearly not part of their group. There are also email systems called "Trystero" for similar reasons, and it makes looking at post office boxes in Scandinavia quite silly even without sampling the local agricultural products.
But, ummm, how shall I say this, the Gap series sucked rocks. Yes, fine, somebody has to take the Ring of the Niebelungen and do something deeper than Tom Holt's lightweight cheerful "Expecting Someone Taller", like transforming it into Space Opera, but for me, this one just failed badly, and the last book or two of it I bought in unfulfilled hopes that he'd finally get somewhere worthwhile with it.
On the other hand, that's ok if it's done well and you're not expecting more of it, and Feist's books really were done well. He's telling a bunch of stories, and telling them well, does an OK job with his characters and the worlds he sets them in and does a good job with the pacing, and does a good enough job of tying the books together. If you're going for basic summer read rather than a deeper literary experience, try the first one or two and see if you want to pick up the rest. It's read-it-once stuff, but it holds up well enough for that.
I'd recommend Steven Brust's books instead, or Donaldson's Thomas Covenant books if they work for you. For Brust, you REALLY REALLY need to start them with "Jhereg", and after that the order's less critical, except that you need to read Athyra before Orca, and The Phoenix Guards before Five Hundred Years After before The Viscount of Adrilankha.
David Brashears's book High Exposure is partly about that year, and partly about his experiences as a climber in general. He led the IMAX film-making expedition that was on Everest that year.
Sure, back in the dark ages, Michael Moorcock did the Elric fantasies, with a moody swordsman with a really nasty haunted sword, and did the "Dancers At the End Of Time" series which was much more interesting, and the didn't work-for-me mixture of the two in "Elric at the end of time". But he's also done some non-fantasy stuff set in the current time with modern human beings, such as King of the City, which is a totally different kind of thing and worth reading.