The WiMax hype comes from conjunction and conjecture right now. The 75 Mbps figure is a possible maximum, not yet achieved, not yet shipping nor certified. The 75 Mbps, 30 miles issue is 75 Mbps *or* 30 miles. Not 75 Mbps at 30 miles. The 30 miles figure isn't yet fully graven in stone, either.
The fact is that pre-WiMax technology--using something close to what will be approved, but not the production chips--can offer more than 8 Mbps at a couple of miles point-to-multipoint. This is very very good, and much cheaper than comparable multiple T1s bound together for cost and simplicity.
Eventually, WiMax will probably offer hundreds of K to a few Mbps for distant line-of-sight locations and tens of Mbps for close by businesses as a fractional T3 alternative.
This is where you hit the culture issue: this coffeeshop wants its mini-society to live by implicit rules and not become one of those places that has all kinds of explicit rules one must follow. Their baristas don't want to time customers and evict them. They want people to linger--but they want lingerers to be a little convivial.
When I wrote up this story, I tried to make it clear that Victrola is in a unique position: the majority of coffeeshops have tons of transient business, and many of them see most traffic between 5 and 9 am. They want to fill seats after that. Victrola is more of a community center masquerading as a coffeeshop in the sense that it's a place that community forms, and thus they have a lot of dwell traffic all day. This is quite rare outside of libraries.
A few errors in this item's text. First, it's not Wi-Fi replacement. This version of WiMax (technically pre-WiMax at the moment) is point-to-multipoint high-speed T-1-plus replacement.
It's $800 per month for 6 Mbps aggregate bandwidth in either 3 up/3 down, 4 up/2 down, or 2 up/4 down configurations. It's intended for businesses that need more than T1 (about $500 per month in Seattle) and don't want to simply double their costs and increase their complxity.
You didn't read the article only the summary. No city money is going into it. The city is just granting access to allow a private company to offer the service. They're providing utility services via an RFP and will purchase services from whatever company builds the network.
My favorite part of this debate is Rep. King stating that you wouldn't want government to go into business by opening a grocery to compete with private enterprise groceries.
I agree. But if there was a single grocery chain in town and they refused to sell to people who lived in certain parts of town and set prices arbitrarily high compared to similar nearby towns that had more than one grocery, I would expect the government to try to defends its citizens basic right to eat.
They could encourage competition by helping other groceries open and defending those new groceries, or they could supply food to people who couldn't afford usurious prices.
But I wouldn't expect my city government to let people starve on the basis of competition.
Unfortunately, he and Steve have been mad at each other since 1982. But we'll see. There will be plenty of accolades for a man who changed the face of personal computing from the inside. (Steve Jobs changed it from the outside.)
I don't like WAPI nor do I think it's a good standard. But the Chinese have long held a desire to have their own self-determination on technology and not be beholden to patents (even ones they refuse to enforce) or manufacturing they don't control.
It doesn't matter whether we think it's reasonable or not; I strongly believe both the political side of the government and scientists within the government think that the U.S. and other nations are just as likely to use their technology domination to steal secrets (business or otherwise) as we think they are of ours.
Very smart point. But I think the Chinese are concerned that most people would use just the 128-bit AES key length that's mandatory. I don't say it's reasonable, but it's part of the justification they're employing.
The Chinese distrust 802.11i because it includes as a mandatory element (and the strongest method of encryption) a 128-bit AES key that the NSA doesn't certify as highest security. They also wanted both authentication and encryption in a single standard for some reason.
Now, the flip side is that 802.11i is great, and so is 802.1X, and the Chinese government clearly want a back door in their standard to allow simple eavesdropping. I cannot believe that WAPI doesn't have a back door. If it did, there would be no reason not to open it to scrutiny.
If there's a back door, someone else will discover it and WAPI will be rendered useless, anyway.
Your points are completely worth thinking about, but the issue here is that a group that is pretending to be independent is funded by telecom and incumbent interests to keep municipalities from even trying to build their own networks. This report will be waved in the face of every city and town and county council before they can fairly evaluate whether municipal broadband would work.
You can't have WiMax without a plan for certification. WiMax, like Wi-Fi, is a sticker that says, "this device has been tested and passed." A recent plugfest to check out interoperability and other characteristics for pre-WiMax gear was canceled, hence part of the delay.
Basically, the 802.16 standards on which WiMax will be a strong subset are done, and it's about figuring out what needs to be in final shipping silicon. No two WiMax devices may be identical, either. There will be profiles so that different features may be enabled in different devices.
And don't expect WiMax mobile to be any time soon, unlike the description that leads into this story -- that's 2006 or 2007, more likely. The first WiMax flavor is fixed point-to-multipoint.
When 802.11n is ratified possibly as late as Nov. 2006, it will likely include speeds at least twice as high as the Belkin product. There are a few different MIMO proposals under consideration, and they will likely be merged with faster speeds and options as optional and a lower speed and tech as mandatory.
But the problem with Belkin and other MIMO solutions is that even if they turn out to be fully 802.11n compatible, they won't do the highest possible speeds. Those highest speeds will likely not cost any more (and maybe less) than the pre-N/MIMO stuff costs today.
So EVEN if you'll be able to upgrade MIMO now to 802.11n in 2006--and that's a huge if and no companies are promising this even in the slighest--you won't get the real speed bump that 802.11n promises.
If you don't need 50 to 70 Mbps of real throughput on your network today, stick with cheap, interoperable 802.11g.
The description of the FCC decision in the lead-in to this topic is incorrect.
The FCC voted to auction off 4 MHz of spectrum for ground-to-air commercial communications including voice, data, and Internet access.
It's already legal to use Wi-Fi on a plane. Connexion by Boeing just yesterday dramatically expanded its satellite-to-plane service that uses Wi-Fi for distributing it on board. It's now available on some SAS flights, on Lufthansa, and a few other airlines with a number more coming online next year.
The real issue with this decision was to make it affordable to roll out cheaper voice and Internet access on domestic carriers. The satellite-based services have much higher costs and thus less flexibility in pricing.
eBooks have turned out to be a great way to fill the gap between articles on a Web site of a few thousand words and the increasingly large exhaustive tomes of 1,000 pages in print. I and a bunch of other authors have written books in the TidBITS' Take Control series for Adam and Tonya Engst, and we've sold collectively in the tens of thousands of "copies."
The books are all about 50 to 150 pages, running $5 to $10 each. They're in PDF form without any DRM enabled. We've turned six of them into print books (four in one volume and two others as single volumes). We use the eBook in part as a way to mature the books: buyers get a subscription to the edition and keep getting updates as we add, correct, and update the books.
For instance, I wrote Take Control of Your AirPort Network focusd on Mac Wi-Fi networks. The first edition was about 70 pages. The 1.1 release ballooned to well over 100 pages because I listened to what readers want and added it in. All the buyers of 1.0 got 1.1 for free.
More recently I spend a couple of hours incorporating all of the changes that Apple introduced with the AirPort 4.1 software update (fairly extensive small fixes and improvements). All the 1.0 to 1.1.2 buyers get 1.1.3 for free, too.
It's rewarding for me as an author to get the kind of quick and precise feedback from readers to write better books and then be able to shoot out those books to the original buyers and all new buyers. It's all a good financial return.
I'm the original poster. You're reading what I wrote with the lack of knowledge necessary to comment on it.
1. All WEP keys are susceptible to nearly the same degree of being broken by collecting enough data passively. Thus, they are all weak. From a definition of weak keys at an online dictionary: "In the extreme, a poor cipher design is simply one with a very large number of weak keys."
2. No, you're misreading this, too. Moskowitz (see his paper) is talking about the seed data, not the resulting way in which it's represented. The lack of randomness in seed data is the problem. So if you take 16 bits of data and turn them into a hex WPA key, it doesn't matter whether it's represented as 256 bits. The whole problem is the algorithm by which it's processed. You need to start with at least 128 bits of data (into hex) that are non-dictionary, non-weak. (In this sense, weak is much more limited.)
3. Sigh. Each user gets a key that has a full 256 bits of randomness.
Re: One: This is sort of the Glenn Gould argument, that an artist holds the audience in a tyranny, but he was talking about performance. Really, the only answer is the market response. If what we charged for our time and efforts weren't repaid by sales of the book and positive response, we'd have to change our position. But the market (so far) has spoken.
Re: Two: Elsewhere in this thread, I and Adam Engst, one of the two people responsible for the Take Control series, have talked about how the books we're producing are special niche items that are short enough to not exhaust readers and long enough to comprehensively cover one aspect of a topic that's typically not enough for a print book. We're not finding price elasticity for those items: that is, charging $1 wouldn't 10 times as many books as charging $10. Instead, we're pricing against the utility in the books we write versus the utility in print books.
Now, as far as global equity, we're not selling these books to the large number of people in the world who live on a dollar a day or less. For the people we write for, $10 is 10 minutes to 1 hour of work.
Per the high prices for the Mac column: that canard has been refuted many times. You can't buy a cheap Mac for as little as a cheap PC, but the cheapest Macs outperform PCs of the same price, all the way up to the most expensive Macs.
On the Windows side, add to that the fact that virus software, firewall software, and other protections haven't been needed, and that you don't lose you productivity or data when some new virus takes over your computer, and I think the dollar balance goes entirely away.
Mac compared to Linux: different purposes, apples and oranges.
Please let me join your glorious writers' paradise in which writers can produce anything and still eat! Ah, to live in Switzerland in the 1920s again...
I wrote the file sharing book, so I'll respond to you directly. It's not an electronic pamphlet. It's over 100 pages of focused advice. It's $10 because that was the optimum price that allows us to sell a relatively small number of books (about 2,000 so far) while compensating me in a reasonable manner for the time it took to write it, and the ongoing time I spend in answering email and revising it. It's actually worked out perfectly.
The book isn't (as I noted in another Slashdot post), select this menu item, click start, next task. I explain how to modify Apache to set up WebDAV under Mac OS X. I have details on creating custom Samba shares. I explain the bugs in Apple's implementation of lukemftpd which prevents proper use of chroot and how to get around it.
Low-level topics these ain't if you've seen the book. This thread on Slashdot has given us a lot of good feedback, but the critique is all coming from people who are IMAGINING what's in the books, not actually looking at the site, downloading the free samples, and then responding.
The economics of publishing are really weird, too. The DVD Studio Pro book you bought for $20 gives a royalty of between $1.50 and $3 to the author or authors per copy sold. The book has to sell over 10,000 copies at that price and size to really make any money for the publisher. If the author worked alone, they might wind up making between $30 and $50 an hour for their time. Not bad at all, but not a massive return.
We're producing these niche -- not low-level -- books for intermediate users who need specific information and don't want to buy $40 and $50 exhaustive books. The exhaustive books are great for general reference, but my file sharing book has details that I was unable to find in any of the giant Panther books: they perversely don't have the space to cover every scenario in each topic because they have to cover EVERY topic.
I can't confirm your numbers (that would be telling), but there are two great factors that accompany your analysis: no marketing budget -- all work of mouth or existing content channels -- and only a few months' sales history on most books!
I'm the author of two of these books and have been using a Mac since 1985. I'm not going to pump up my own effort, but I can tell you how much of these books arise specifically from the fact that we, as authors and experienced Mac users, couldn't find complete and/or accurate answers to the questions that the books address, nor could we find the comprehensive start to finish advice that we needed.
Our books aren't "here's menu A, here's menu B." The whole point is that they're not exhaustive, but they focus in on specific details. The books try to solve problems and to do it in finite space.
It would also be another thing if you could spend a few minutes and find the answer on Google for everything in the 50 to 100 pages in the books. But you can't. It might take you a few minutes per page to find what's in the book. So if you spent, say, 2 to 4 hours, you might save $5 to $10 -- if you could find the information.
My first book on file sharing took me about 60 hours to write on top of my experience with Unix (1994 to present), Linux (1997 to present), and Mac OS X (10.0.0 to present). The AirPort book that I just released a few weeks ago took less time in the first edition, but we commit to releasing updates with new and updated material--version 1.0 was about 90 pages; 1.1 (a free update for 1.0 book buyers) will be about 160.
Another interesting interaction with the ebooks is that we hear from readers and can practically immediately make changes. People who bought my AirPort books first version gave me great feedback. I incorporated almost all of it into new information for the 1.1 release, which all of these readers will get for free. I love that.
I hope this clears up a few of the issues. Almost all of the writers involved to date are freelancers, and it's really quite difficult to make a good living writing about using technology, which, I hope, helps other people. These ebooks make it financially possible for me to write books on topics that people are asking us for but that aren't available in a few minutes of Google searching, and that aren't cost effective for a print book, which has to sell 5,000 to 10,000 copies (depending on size) to be even a reasonable success.
Imagine, for instance, a 50-page book on regular expression pattern matching for Mac OS X users. It's a possibility, and would be highly useful. But you can't write a print book like that. (Although O'Reilly has a more generalized book on the topic in print!)
Slashdot Mirror
The WiMax hype comes from conjunction and conjecture right now. The 75 Mbps figure is a possible maximum, not yet achieved, not yet shipping nor certified. The 75 Mbps, 30 miles issue is 75 Mbps *or* 30 miles. Not 75 Mbps at 30 miles. The 30 miles figure isn't yet fully graven in stone, either.
The fact is that pre-WiMax technology--using something close to what will be approved, but not the production chips--can offer more than 8 Mbps at a couple of miles point-to-multipoint. This is very very good, and much cheaper than comparable multiple T1s bound together for cost and simplicity.
Eventually, WiMax will probably offer hundreds of K to a few Mbps for distant line-of-sight locations and tens of Mbps for close by businesses as a fractional T3 alternative.
This is where you hit the culture issue: this coffeeshop wants its mini-society to live by implicit rules and not become one of those places that has all kinds of explicit rules one must follow. Their baristas don't want to time customers and evict them. They want people to linger--but they want lingerers to be a little convivial.
When I wrote up this story, I tried to make it clear that Victrola is in a unique position: the majority of coffeeshops have tons of transient business, and many of them see most traffic between 5 and 9 am. They want to fill seats after that. Victrola is more of a community center masquerading as a coffeeshop in the sense that it's a place that community forms, and thus they have a lot of dwell traffic all day. This is quite rare outside of libraries.
A few errors in this item's text. First, it's not Wi-Fi replacement. This version of WiMax (technically pre-WiMax at the moment) is point-to-multipoint high-speed T-1-plus replacement.
It's $800 per month for 6 Mbps aggregate bandwidth in either 3 up/3 down, 4 up/2 down, or 2 up/4 down configurations. It's intended for businesses that need more than T1 (about $500 per month in Seattle) and don't want to simply double their costs and increase their complxity.
You didn't read the article only the summary. No city money is going into it. The city is just granting access to allow a private company to offer the service. They're providing utility services via an RFP and will purchase services from whatever company builds the network.
My favorite part of this debate is Rep. King stating that you wouldn't want government to go into business by opening a grocery to compete with private enterprise groceries.
I agree. But if there was a single grocery chain in town and they refused to sell to people who lived in certain parts of town and set prices arbitrarily high compared to similar nearby towns that had more than one grocery, I would expect the government to try to defends its citizens basic right to eat.
They could encourage competition by helping other groceries open and defending those new groceries, or they could supply food to people who couldn't afford usurious prices.
But I wouldn't expect my city government to let people starve on the basis of competition.
Unfortunately, he and Steve have been mad at each other since 1982. But we'll see. There will be plenty of accolades for a man who changed the face of personal computing from the inside. (Steve Jobs changed it from the outside.)
I don't like WAPI nor do I think it's a good standard. But the Chinese have long held a desire to have their own self-determination on technology and not be beholden to patents (even ones they refuse to enforce) or manufacturing they don't control.
It doesn't matter whether we think it's reasonable or not; I strongly believe both the political side of the government and scientists within the government think that the U.S. and other nations are just as likely to use their technology domination to steal secrets (business or otherwise) as we think they are of ours.
Very smart point. But I think the Chinese are concerned that most people would use just the 128-bit AES key length that's mandatory. I don't say it's reasonable, but it's part of the justification they're employing.
The Chinese distrust 802.11i because it includes as a mandatory element (and the strongest method of encryption) a 128-bit AES key that the NSA doesn't certify as highest security. They also wanted both authentication and encryption in a single standard for some reason.
Now, the flip side is that 802.11i is great, and so is 802.1X, and the Chinese government clearly want a back door in their standard to allow simple eavesdropping. I cannot believe that WAPI doesn't have a back door. If it did, there would be no reason not to open it to scrutiny.
If there's a back door, someone else will discover it and WAPI will be rendered useless, anyway.
Your points are completely worth thinking about, but the issue here is that a group that is pretending to be independent is funded by telecom and incumbent interests to keep municipalities from even trying to build their own networks. This report will be waved in the face of every city and town and county council before they can fairly evaluate whether municipal broadband would work.
You can't have WiMax without a plan for certification. WiMax, like Wi-Fi, is a sticker that says, "this device has been tested and passed." A recent plugfest to check out interoperability and other characteristics for pre-WiMax gear was canceled, hence part of the delay.
Basically, the 802.16 standards on which WiMax will be a strong subset are done, and it's about figuring out what needs to be in final shipping silicon. No two WiMax devices may be identical, either. There will be profiles so that different features may be enabled in different devices.
And don't expect WiMax mobile to be any time soon, unlike the description that leads into this story -- that's 2006 or 2007, more likely. The first WiMax flavor is fixed point-to-multipoint.
When 802.11n is ratified possibly as late as Nov. 2006, it will likely include speeds at least twice as high as the Belkin product. There are a few different MIMO proposals under consideration, and they will likely be merged with faster speeds and options as optional and a lower speed and tech as mandatory.
But the problem with Belkin and other MIMO solutions is that even if they turn out to be fully 802.11n compatible, they won't do the highest possible speeds. Those highest speeds will likely not cost any more (and maybe less) than the pre-N/MIMO stuff costs today.
So EVEN if you'll be able to upgrade MIMO now to 802.11n in 2006--and that's a huge if and no companies are promising this even in the slighest--you won't get the real speed bump that 802.11n promises.
If you don't need 50 to 70 Mbps of real throughput on your network today, stick with cheap, interoperable 802.11g.
First at a consumer price level. Nikon's cameras are awesome and awesomely expensive.
MT 3.x has a Comments page that lets you review 20, 50, etc., comments at a time, select them all to delete, etc.
Much improve and appreciated. I also turn on comment moderation and this fixed the problems I had with comment spam.
The description of the FCC decision in the lead-in to this topic is incorrect.
The FCC voted to auction off 4 MHz of spectrum for ground-to-air commercial communications including voice, data, and Internet access.
It's already legal to use Wi-Fi on a plane. Connexion by Boeing just yesterday dramatically expanded its satellite-to-plane service that uses Wi-Fi for distributing it on board. It's now available on some SAS flights, on Lufthansa, and a few other airlines with a number more coming online next year.
The real issue with this decision was to make it affordable to roll out cheaper voice and Internet access on domestic carriers. The satellite-based services have much higher costs and thus less flexibility in pricing.
eBooks have turned out to be a great way to fill the gap between articles on a Web site of a few thousand words and the increasingly large exhaustive tomes of 1,000 pages in print. I and a bunch of other authors have written books in the TidBITS' Take Control series for Adam and Tonya Engst, and we've sold collectively in the tens of thousands of "copies."
The books are all about 50 to 150 pages, running $5 to $10 each. They're in PDF form without any DRM enabled. We've turned six of them into print books (four in one volume and two others as single volumes). We use the eBook in part as a way to mature the books: buyers get a subscription to the edition and keep getting updates as we add, correct, and update the books.
For instance, I wrote Take Control of Your AirPort Network focusd on Mac Wi-Fi networks. The first edition was about 70 pages. The 1.1 release ballooned to well over 100 pages because I listened to what readers want and added it in. All the buyers of 1.0 got 1.1 for free.
More recently I spend a couple of hours incorporating all of the changes that Apple introduced with the AirPort 4.1 software update (fairly extensive small fixes and improvements). All the 1.0 to 1.1.2 buyers get 1.1.3 for free, too.
It's rewarding for me as an author to get the kind of quick and precise feedback from readers to write better books and then be able to shoot out those books to the original buyers and all new buyers. It's all a good financial return.
WPA isn't easy to crack when you choose a good key. Just invent a passphrase like "My d0g!! has f6666%%&%__seas" and you're safe as houses.
I'm the original poster. You're reading what I wrote with the lack of knowledge necessary to comment on it.
1. All WEP keys are susceptible to nearly the same degree of being broken by collecting enough data passively. Thus, they are all weak. From a definition of weak keys at an online dictionary: "In the extreme, a poor cipher design is simply one with a very large number of weak keys."
2. No, you're misreading this, too. Moskowitz (see his paper) is talking about the seed data, not the resulting way in which it's represented. The lack of randomness in seed data is the problem. So if you take 16 bits of data and turn them into a hex WPA key, it doesn't matter whether it's represented as 256 bits. The whole problem is the algorithm by which it's processed. You need to start with at least 128 bits of data (into hex) that are non-dictionary, non-weak. (In this sense, weak is much more limited.)
3. Sigh. Each user gets a key that has a full 256 bits of randomness.
You are being picky about your words incorrectly.
Nah.
Good points.
Re: One: This is sort of the Glenn Gould argument, that an artist holds the audience in a tyranny, but he was talking about performance. Really, the only answer is the market response. If what we charged for our time and efforts weren't repaid by sales of the book and positive response, we'd have to change our position. But the market (so far) has spoken.
Re: Two: Elsewhere in this thread, I and Adam Engst, one of the two people responsible for the Take Control series, have talked about how the books we're producing are special niche items that are short enough to not exhaust readers and long enough to comprehensively cover one aspect of a topic that's typically not enough for a print book. We're not finding price elasticity for those items: that is, charging $1 wouldn't 10 times as many books as charging $10. Instead, we're pricing against the utility in the books we write versus the utility in print books.
Now, as far as global equity, we're not selling these books to the large number of people in the world who live on a dollar a day or less. For the people we write for, $10 is 10 minutes to 1 hour of work.
Per the high prices for the Mac column: that canard has been refuted many times. You can't buy a cheap Mac for as little as a cheap PC, but the cheapest Macs outperform PCs of the same price, all the way up to the most expensive Macs.
On the Windows side, add to that the fact that virus software, firewall software, and other protections haven't been needed, and that you don't lose you productivity or data when some new virus takes over your computer, and I think the dollar balance goes entirely away.
Mac compared to Linux: different purposes, apples and oranges.
Please let me join your glorious writers' paradise in which writers can produce anything and still eat! Ah, to live in Switzerland in the 1920s again...
I wrote the file sharing book, so I'll respond to you directly. It's not an electronic pamphlet. It's over 100 pages of focused advice. It's $10 because that was the optimum price that allows us to sell a relatively small number of books (about 2,000 so far) while compensating me in a reasonable manner for the time it took to write it, and the ongoing time I spend in answering email and revising it. It's actually worked out perfectly.
The book isn't (as I noted in another Slashdot post), select this menu item, click start, next task. I explain how to modify Apache to set up WebDAV under Mac OS X. I have details on creating custom Samba shares. I explain the bugs in Apple's implementation of lukemftpd which prevents proper use of chroot and how to get around it.
Low-level topics these ain't if you've seen the book. This thread on Slashdot has given us a lot of good feedback, but the critique is all coming from people who are IMAGINING what's in the books, not actually looking at the site, downloading the free samples, and then responding.
The economics of publishing are really weird, too. The DVD Studio Pro book you bought for $20 gives a royalty of between $1.50 and $3 to the author or authors per copy sold. The book has to sell over 10,000 copies at that price and size to really make any money for the publisher. If the author worked alone, they might wind up making between $30 and $50 an hour for their time. Not bad at all, but not a massive return.
We're producing these niche -- not low-level -- books for intermediate users who need specific information and don't want to buy $40 and $50 exhaustive books. The exhaustive books are great for general reference, but my file sharing book has details that I was unable to find in any of the giant Panther books: they perversely don't have the space to cover every scenario in each topic because they have to cover EVERY topic.
I can't confirm your numbers (that would be telling), but there are two great factors that accompany your analysis: no marketing budget -- all work of mouth or existing content channels -- and only a few months' sales history on most books!
I'm the author of two of these books and have been using a Mac since 1985. I'm not going to pump up my own effort, but I can tell you how much of these books arise specifically from the fact that we, as authors and experienced Mac users, couldn't find complete and/or accurate answers to the questions that the books address, nor could we find the comprehensive start to finish advice that we needed.
Our books aren't "here's menu A, here's menu B." The whole point is that they're not exhaustive, but they focus in on specific details. The books try to solve problems and to do it in finite space.
It would also be another thing if you could spend a few minutes and find the answer on Google for everything in the 50 to 100 pages in the books. But you can't. It might take you a few minutes per page to find what's in the book. So if you spent, say, 2 to 4 hours, you might save $5 to $10 -- if you could find the information.
My first book on file sharing took me about 60 hours to write on top of my experience with Unix (1994 to present), Linux (1997 to present), and Mac OS X (10.0.0 to present). The AirPort book that I just released a few weeks ago took less time in the first edition, but we commit to releasing updates with new and updated material--version 1.0 was about 90 pages; 1.1 (a free update for 1.0 book buyers) will be about 160.
Another interesting interaction with the ebooks is that we hear from readers and can practically immediately make changes. People who bought my AirPort books first version gave me great feedback. I incorporated almost all of it into new information for the 1.1 release, which all of these readers will get for free. I love that.
I hope this clears up a few of the issues. Almost all of the writers involved to date are freelancers, and it's really quite difficult to make a good living writing about using technology, which, I hope, helps other people. These ebooks make it financially possible for me to write books on topics that people are asking us for but that aren't available in a few minutes of Google searching, and that aren't cost effective for a print book, which has to sell 5,000 to 10,000 copies (depending on size) to be even a reasonable success.
Imagine, for instance, a 50-page book on regular expression pattern matching for Mac OS X users. It's a possibility, and would be highly useful. But you can't write a print book like that. (Although O'Reilly has a more generalized book on the topic in print!)