Well, you can use unique local addressing for internal stuff. NAT is called a big no-no, but pragmatically, it should be doable if required. Besides, *everything* was architected carefully to make renumbering easier if your subnet changes in IPv6, which was not done in IPv4. In short, theoretically no worse than a company having a route from an ISP and changing ISPs today, with internal addresses that remain constant and NAT to remap on the way out. It's in practice significantly better because renumbering is baked in all levels. If you are talking about large companies that have their own allocations in v4 that do *not* vary on their 'ISP', I'll wager they will get/32s like an ISP and get to keep their address no matter who they partner with. In terms of security, it's no more work than maintaining the NAT rules even if not NATing. Replace the configuration required to do NAT with an equivalent amount of effort on firewalling and you are there.
I don't see why IPv6 makes it an urgent need for medium and small companies to get PI space. If they had no urgent need in IPv4, they can either accept the renumbering risk like they do today, or use ULA and NAT like they do today.
Now, they still have a few kinks to work out before DHCPv6 is even theoretically as workable as DHCPv4, and even with all the pending drafts going through, operators would have to rethink how they do things entirely, though they may be able to get to the same general goal.
More like they *want* the future to be vendor-controlled. They always hoped that, but never thought the consumers to be *that* self-destructive until Apple essentially did it. Now they hope they can ape Apple's success on that front.
'The Stand' from Amazon should not be considered a fundamentally different product than 'The Stand' from Barnes and Noble. If you obey the letter of the law and do not undo the DRM schemes, those are two different products.
But look at all the success they have had with the Windows editions for MIPS, Alpha, Itanium, and Power! (No, I don't count the kernel on xBox 360 in the same realm here).
The argument that DRM is ok because it is futile against those in the know is horrible. The ostensible point is to combat piracy, but in this case the pirates aren't even slowed down and legitimate customers are forced to violate DMCA (in US) to get what should be protected as fair use.
Also, easily is a relative term. Last I checked it was non-trivial to do if all you had was a linux box and a Kindle, but trivial if you had a Windows box. Even then though, it required the right version of the Kindle app that was obfuscated in the way that your particular de-DRM script understood.
The business merit for charging/restricting a media should not be based on how comparatively crappy the pirates do at breaking the law. It should be based on value delivered to the consumer.
Relative to the cost of maintaining a sound studio and the number of people required to get the job done in the shorter amount of time, it is low. The key distinction is that movies and music require relatively expensive facilities and people that are reusable between projects, making 'publishers' more inherently valuable in distributing costs that would be unreasonable for a single effort to incur. In purely digital text media, the author doesn't get that much out of the publisher comparatively speaking, and can more realistically have a middle-man free relationship with vendors.
Sorry, but $3 for something that takes a year or more to create isn't much money.
That argument would make sense if incremental cost were the dominating factor or even a major contributer. Even if a paperback costs 50 cents to create as an incremental cost, that's still probably two orders of magnitude higher than a digital version. With negligible incremental cost, you can play all sorts of games with the demand curve and so talking about an individual purchase without the context of how that price increases/decreases volume of units sold is kind of pointless. If hypothetically charging $5 gets you 10,000 copies, but $1 got you one million copies, then $1 *is* a lot of money for a man-year of work, it's even more than $5. It's hard to play those games with even mass-market paperback, because you need to guess pretty well in advance what the required run will be or else get eaten alive by the incremental costs.
And if you want to bring up paperback pricing, nearly any book on Amazon in paperback form is available for basically the same price as an ebook
I consider this to be unreasonable. If I were going to buy an eBook and get no physical copy, I should benefit from the decreased cost from not having to create the book and carry stock that may or may not sell. If you insist on gouging me for the eBook, then I should be able to buy a physical book and be entitled to an eBook copy to go with it. If you insist on new-release pricing being high, then do hardcover only and save the e-book for paperback time. An 'early adopter' for a book getting a hardback at least gets a product with extra value that persists after the paperback comes out, but the eBook edition will be *identical* before and after a hypothetical price drop, leaving the original purchaser with nothing tangible to show after that arbitrary point.
Finally, I'm tired of people only looking at costs and using that to justify piracy.
I'm not doing that, my stance is abstain from the industry. I feel the need to make it known why I'm almost abstaining from the market (have taken advantage of some appropriately priced ebooks when they are on 'special'), and how I (and presumably others like me) could be persuaded to participate. Pricing is one issue and DRM is another. DRM makes it damn near impossible for law-abiding (since DMCA screwed over fair use) people to do things like move a book from their Nook to a new Kindle they got, but it's absolutely useless at deterring unauthorized copies for people willing to break laws (just run two commands and poof, all DRM gone from your eBook library).
There's more to any business than per-unit costs. And if you think you're entitled to everything at cost, just go into MacDonalds and try paying a dime (cost) for a Coke.
Well, for one, I won't pay the $2.50 many restaurants charge for tea or soft drink and instead go for water. But in the larger scheme, incremental cost is a key factor in contrasting digital distribution from physical distribution. The cost of producing a copy, of stocking surplus, of risking spending money on copies that will never sell, of shipping that stuff all over the place, that cost is significant and yet consumers don't see any significant savings at all by participating in a model that saves the vendor and publisher from all of that.
There is nothing in the DRM encumbered market that makes this true. The stewards of the DRM are the likes of Amazon, B&N, Kobo, etc. Even if an author *did* consider DRM a must-have for him to be comfortable publishing, the vendors will gladly help that author self-publish with DRM in order to cut out the publisher middle man. The publisher doesn't implement any technical infrastructure required for DRM to function.
Even if it were the case that DRM is inaccessible, sure they don't get DRM, but they also don't have to let a publisher gouge them for money in the middle.
The reasons for that are technical. Books on a technical level are more comparable to film reels and vinyl records without tape or any sort of digital representation existing.
With the popularity of eBooks, that playing ground is now nearly level on distribution, leaving little more than production cost distinguishing the mediums from a business standpoint.
The seemingly most popular eReader can't 'legally' load copyrighted ebooks from Borders, B&N, or public libraries. Any user doing so violates the DMCA to get it there.
It's worse when you see people advocating buying dedicated eReaders per store as a reasonable thing to have to do with the reasoning 'why would you expect to use Gillette blades with a Bic handle?'.
Music has by and large ditched DRM efforts on purchased content (may still factor in subscription/streaming services, I'm not paying much attention there. I think music has found a comfortable low price point that renders the point mostly moot. Music may be DRMed on streaming, and the best protection their is that a lot of people who would deobfuscate their stuff have no motivation to since purchases aren't afflicted by DRM. It's almost reaching a point of sanity, that the per-unit cost can be brought low because the distribution overhead is minimal (even more minimal without DRM) and the production cost is sizable, but not horribly bad.
Books, on the other hand are still DRMed by the dominant vendors. They also charge outrageous amounts and want to compare the price to the hardcover editions, completely ignoring the fact that per-unit cost is next to nothing compared to even paperback. They don't even have a significant up-front cost to recover (Movies/TV have actors/sets/etc, music has engineers and sound studios that are really needed for respectable sound, books don't *need* much more than a diligent author with a computer, though editors and artists frequently help). The DRM on at least the epub stuff is laughably easy to remove (because without removing it, it's pretty damn hard to actually put it on many devices, so they get a large volume of people out to get it). I wonder if publishers are keeping prices high and the distribution overly complicated just to slow down the electronic market because they know full well they play a negligible role if distribution becomes trivial to do.
I company I used to work for was operated by someone too naive to understand it was the support and not the intrinsic equipment that mattered. I came in and they were using some Cheapo NAS box that used 4 drives on two IDE channels as the storage in RAID5. Of course one drive failed and took out a channel, so after I recovered the data (it was backed up, but I recovered data off the three working drives), it was time to look for a replacement. I was told a very very very puny budget, so I priced out basically the same thing except with 4 IDE channels and bought as a new, supported, warrantied server. After a few days, the President announced he found the perfect 'enterprise' Dell enclosure on eBay and that was going to be the solution.... Even off random ebay guy, as an enclosure only (excluding controller and server to drive it), at much reduced capacity it was about twice as much as the server I suggested (President was willing to stretch the budget he dictated because he found himself such a 'good deal').
Of course, after two weeks of operation, the midplane somehow glitched in a way to corrupt drives and lose all data. Dell charged quite a bit for the service call to come in and fix the stupid thing since there was no support or warranty remotely associated with the thing, and the enclosure still had no warranty associated with it. He never admitted that it was a mistake. My second-worst job didn't last too much longer after that (got fired because they decided they needed to open a position to afford to look for a person with MCSE at the time).
I think the point of that example is that people will do seemingly innocuous things like mention a grandfather with a differing last name from yours and not think twice about it, while at the same time way too many financial institutions use that like a secret to prove your identity. Yes, email could be forwarded/etc, but generally the recipient list is more pruned and targetted. Part of this whole social networking thing is that people can blast out data to many people and let most of them ignore it if they don't care. Explicitly thoughtless dissemination of info.
You point out that he posted examples that demonstrate how capable he is of being savvy about it, however that is when he sits down to explicitly think about it. Even for those that *would* be aware of the potential implications if they thought it through, if they get into it they are unlikely to thoroughly think through everything they post.
The notion of hundreds upon hundreds of 'friends' is ludicrous on the surface and dangerous when you consider the potential import of everyday thoughts and conversations.
Also, AOL had their own walled garden of value that evaporated with an internet that automatically looked the same regardless of whether you accessed it from home, work, or school. AOL on broadband would have been insufficient to keep them relevant (it was on broadband of course, and available anywhere you could install their software with any internet access, which was not most school and work cases).
In terms of predicting whether or not facebook or anyone else will die and when, all of the rational discussion in the world won't matter. It's a whim that is an emergent feature of a population that simply is unknowable.
The people whom you lost respect for because of stupid or melodramatic stuff don't care. A whole lot of people love stupid and even very destructive drama. Facebook is a medium where those exchanges are made more real by their typical association to real identities. This is not a problem or failing of social networking, it's a reflection of human nature and owes no small part of their success to that sort of asinine stuff.
From a technical perspective, sure, lots of companies could do this, it's not hard.
Facebook is *not* about a technical offering. A technical offering is a prerequisite, but the success is entirely predicated on networking effects. Therefore, popularity is the single most important thing. By some combination of 'getting it right', strategy, and sheer luck, they have ultimately built the most ubiquitous offering. I would say facebook is the first widespread provider to make aggregation the central feature, which was key to establish the networking effect as an intrinsic factor where it wasn't so much for geocities and myspace where you had to explicitly check in on your 'friends'.
I would like to agree with you, but I think the issue is the more bullet-proof, federated architectures of piece-parts that are readily available otherwise. If that were the case, neither MySpace nor Facebook would have come to prominence. AOL died because the internet became a superset of what AOL contained, could be accessed equally from school campuses, work, and home, and opened the door for cheaper and higher-bandwidth providers. Content providers couldn't restrict themselves to a service that most people couldn't get to from work or school, so they published at least on the internet and maybe AOL. That was the key nail in the coffin IMO.
On user provided content, the trend seems clear enough from geocities/angelfire, to myspace, to facebook. Each required less and less work from users (by sacrificing customization capability). Also, as they ditched customization, one navigating content provided by several friends is faced with a more and more consistent set of data. Of course, much of this could be hypothetically reproduced by a near monopoly releasing a distributed approach to this, but it would *have* to be on some embedded device that people wouldn't tend to sleep/turn off, eliminating the most likely candidate for that.
Another issue of this is the natural aggregation of information in facebook. The users look at one page and data provided by all of their 'friends' is aggregated into one apparent stream of data. This is the tricky part to do in a distributed fashion. Becoming a friend or fan of someone could hypothetically request an explicit push of data from one person to all friends/fans they have then and there even if that friend never reads it. However, this would scale worse than what a site like facebook deals with. The converse of scan on read would similarly be bad, with the additional problem that the experience would be apparently sluggish and miss/hiccup as peers are down. Approaches to make it appear responsive would cause 'invisible' gaps to be in the data where unavailable data is missing and magically appears when available. If not assembled by the home routers, it would have to be a federated set of content-providers. e-mail is probably the most similar model, and that carries the problem of inefficient storage. Each user gets a disparate copy of the data. Also, once out there, it cannot be edited in a single copy. Basically, people explicitly want to put all their data in a *single* authoritative place, which naturally leads to things like facebook.
Which brings us to another point, people are used to what datacenters provide in terms of reliable data delivery. Home routers would be subject to ISP outages (which are not infrequent in residential price points), power outages/blips, people just turning their stuff off, router hardware/firmware problems without elaborate failover or even someone paying particularly close attention. Rationally, a 'facebook' application isn't critical enough to consider an outage in a vacuum particularly important, but the emergent behavior of a few missing pieces would be perceived as garbage. Not to mention more work on the part of the participants to notice when their stuff isn't working and fix something.
I have dovecot, postfix, roundcube and squirrelmail running on my home system providing mail, and have other web content I share up too. I must confess I'm moving more and more of my email to gmail. My home system is pretty sufficient and I like the control, however when I'm on vacation I worry that something will happen that would require me to fix that would not be possible remotely. I also like doing a lot of things like upgrades and experimentation and the system with email responsibility I can't really mess with in a way that induces a long downtime. I would move my home domain to google and be done with it, except I use -<suffix> addresses and google only supports +<suffix> addressing.
I don't think facebook will continue indefinitely specifically.. It will be replaced by something
Slashdot Mirror
Well, you can use unique local addressing for internal stuff. NAT is called a big no-no, but pragmatically, it should be doable if required. Besides, *everything* was architected carefully to make renumbering easier if your subnet changes in IPv6, which was not done in IPv4. In short, theoretically no worse than a company having a route from an ISP and changing ISPs today, with internal addresses that remain constant and NAT to remap on the way out. It's in practice significantly better because renumbering is baked in all levels. If you are talking about large companies that have their own allocations in v4 that do *not* vary on their 'ISP', I'll wager they will get /32s like an ISP and get to keep their address no matter who they partner with. In terms of security, it's no more work than maintaining the NAT rules even if not NATing. Replace the configuration required to do NAT with an equivalent amount of effort on firewalling and you are there.
I don't see why IPv6 makes it an urgent need for medium and small companies to get PI space. If they had no urgent need in IPv4, they can either accept the renumbering risk like they do today, or use ULA and NAT like they do today.
Now, they still have a few kinks to work out before DHCPv6 is even theoretically as workable as DHCPv4, and even with all the pending drafts going through, operators would have to rethink how they do things entirely, though they may be able to get to the same general goal.
More like they *want* the future to be vendor-controlled. They always hoped that, but never thought the consumers to be *that* self-destructive until Apple essentially did it. Now they hope they can ape Apple's success on that front.
Presumably hired to patch any apparent 'exploits' they would have otherwise caught.
Not a big fan of this, but it is more than a shade better than Sony trying to sue their problems out of existence.
Not just Mac, Linux and Windows too.
I've had that problem too. Sometimes I can't even click to position the text cursor beyond a certain threshold.
I had that problem in Linux, and just now checked in this very comment that happens to be chrome on a Windows box and it won't paste etiher.
It's peculiar, because it seemed spotty.
'The Stand' from Amazon should not be considered a fundamentally different product than 'The Stand' from Barnes and Noble. If you obey the letter of the law and do not undo the DRM schemes, those are two different products.
ARM windows which in my opinion will be DOA
But look at all the success they have had with the Windows editions for MIPS, Alpha, Itanium, and Power! (No, I don't count the kernel on xBox 360 in the same realm here).
But those are actual services to provide. None of that is inherently tied to the DRM aspect of the eBooks.
The argument that DRM is ok because it is futile against those in the know is horrible. The ostensible point is to combat piracy, but in this case the pirates aren't even slowed down and legitimate customers are forced to violate DMCA (in US) to get what should be protected as fair use.
Also, easily is a relative term. Last I checked it was non-trivial to do if all you had was a linux box and a Kindle, but trivial if you had a Windows box. Even then though, it required the right version of the Kindle app that was obfuscated in the way that your particular de-DRM script understood.
http://www.ebaumsworld.com/video/watch/80953120/
The business merit for charging/restricting a media should not be based on how comparatively crappy the pirates do at breaking the law. It should be based on value delivered to the consumer.
no up-front cost my ass
Relative to the cost of maintaining a sound studio and the number of people required to get the job done in the shorter amount of time, it is low. The key distinction is that movies and music require relatively expensive facilities and people that are reusable between projects, making 'publishers' more inherently valuable in distributing costs that would be unreasonable for a single effort to incur. In purely digital text media, the author doesn't get that much out of the publisher comparatively speaking, and can more realistically have a middle-man free relationship with vendors.
Sorry, but $3 for something that takes a year or more to create isn't much money.
That argument would make sense if incremental cost were the dominating factor or even a major contributer. Even if a paperback costs 50 cents to create as an incremental cost, that's still probably two orders of magnitude higher than a digital version. With negligible incremental cost, you can play all sorts of games with the demand curve and so talking about an individual purchase without the context of how that price increases/decreases volume of units sold is kind of pointless. If hypothetically charging $5 gets you 10,000 copies, but $1 got you one million copies, then $1 *is* a lot of money for a man-year of work, it's even more than $5. It's hard to play those games with even mass-market paperback, because you need to guess pretty well in advance what the required run will be or else get eaten alive by the incremental costs.
And if you want to bring up paperback pricing, nearly any book on Amazon in paperback form is available for basically the same price as an ebook
I consider this to be unreasonable. If I were going to buy an eBook and get no physical copy, I should benefit from the decreased cost from not having to create the book and carry stock that may or may not sell. If you insist on gouging me for the eBook, then I should be able to buy a physical book and be entitled to an eBook copy to go with it. If you insist on new-release pricing being high, then do hardcover only and save the e-book for paperback time. An 'early adopter' for a book getting a hardback at least gets a product with extra value that persists after the paperback comes out, but the eBook edition will be *identical* before and after a hypothetical price drop, leaving the original purchaser with nothing tangible to show after that arbitrary point.
Finally, I'm tired of people only looking at costs and using that to justify piracy.
I'm not doing that, my stance is abstain from the industry. I feel the need to make it known why I'm almost abstaining from the market (have taken advantage of some appropriately priced ebooks when they are on 'special'), and how I (and presumably others like me) could be persuaded to participate. Pricing is one issue and DRM is another. DRM makes it damn near impossible for law-abiding (since DMCA screwed over fair use) people to do things like move a book from their Nook to a new Kindle they got, but it's absolutely useless at deterring unauthorized copies for people willing to break laws (just run two commands and poof, all DRM gone from your eBook library).
There's more to any business than per-unit costs. And if you think you're entitled to everything at cost, just go into MacDonalds and try paying a dime (cost) for a Coke.
Well, for one, I won't pay the $2.50 many restaurants charge for tea or soft drink and instead go for water. But in the larger scheme, incremental cost is a key factor in contrasting digital distribution from physical distribution. The cost of producing a copy, of stocking surplus, of risking spending money on copies that will never sell, of shipping that stuff all over the place, that cost is significant and yet consumers don't see any significant savings at all by participating in a model that saves the vendor and publisher from all of that.
authors can't self-publish as easily
There is nothing in the DRM encumbered market that makes this true. The stewards of the DRM are the likes of Amazon, B&N, Kobo, etc. Even if an author *did* consider DRM a must-have for him to be comfortable publishing, the vendors will gladly help that author self-publish with DRM in order to cut out the publisher middle man. The publisher doesn't implement any technical infrastructure required for DRM to function.
Even if it were the case that DRM is inaccessible, sure they don't get DRM, but they also don't have to let a publisher gouge them for money in the middle.
The reasons for that are technical. Books on a technical level are more comparable to film reels and vinyl records without tape or any sort of digital representation existing.
With the popularity of eBooks, that playing ground is now nearly level on distribution, leaving little more than production cost distinguishing the mediums from a business standpoint.
The seemingly most popular eReader can't 'legally' load copyrighted ebooks from Borders, B&N, or public libraries. Any user doing so violates the DMCA to get it there.
It's worse when you see people advocating buying dedicated eReaders per store as a reasonable thing to have to do with the reasoning 'why would you expect to use Gillette blades with a Bic handle?'.
Music has by and large ditched DRM efforts on purchased content (may still factor in subscription/streaming services, I'm not paying much attention there. I think music has found a comfortable low price point that renders the point mostly moot. Music may be DRMed on streaming, and the best protection their is that a lot of people who would deobfuscate their stuff have no motivation to since purchases aren't afflicted by DRM. It's almost reaching a point of sanity, that the per-unit cost can be brought low because the distribution overhead is minimal (even more minimal without DRM) and the production cost is sizable, but not horribly bad.
Books, on the other hand are still DRMed by the dominant vendors. They also charge outrageous amounts and want to compare the price to the hardcover editions, completely ignoring the fact that per-unit cost is next to nothing compared to even paperback. They don't even have a significant up-front cost to recover (Movies/TV have actors/sets/etc, music has engineers and sound studios that are really needed for respectable sound, books don't *need* much more than a diligent author with a computer, though editors and artists frequently help). The DRM on at least the epub stuff is laughably easy to remove (because without removing it, it's pretty damn hard to actually put it on many devices, so they get a large volume of people out to get it). I wonder if publishers are keeping prices high and the distribution overly complicated just to slow down the electronic market because they know full well they play a negligible role if distribution becomes trivial to do.
I company I used to work for was operated by someone too naive to understand it was the support and not the intrinsic equipment that mattered. I came in and they were using some Cheapo NAS box that used 4 drives on two IDE channels as the storage in RAID5. Of course one drive failed and took out a channel, so after I recovered the data (it was backed up, but I recovered data off the three working drives), it was time to look for a replacement. I was told a very very very puny budget, so I priced out basically the same thing except with 4 IDE channels and bought as a new, supported, warrantied server. After a few days, the President announced he found the perfect 'enterprise' Dell enclosure on eBay and that was going to be the solution.... Even off random ebay guy, as an enclosure only (excluding controller and server to drive it), at much reduced capacity it was about twice as much as the server I suggested (President was willing to stretch the budget he dictated because he found himself such a 'good deal').
Of course, after two weeks of operation, the midplane somehow glitched in a way to corrupt drives and lose all data. Dell charged quite a bit for the service call to come in and fix the stupid thing since there was no support or warranty remotely associated with the thing, and the enclosure still had no warranty associated with it. He never admitted that it was a mistake. My second-worst job didn't last too much longer after that (got fired because they decided they needed to open a position to afford to look for a person with MCSE at the time).
That's the same mother's maiden name I use on my briefcase!
I think the point of that example is that people will do seemingly innocuous things like mention a grandfather with a differing last name from yours and not think twice about it, while at the same time way too many financial institutions use that like a secret to prove your identity. Yes, email could be forwarded/etc, but generally the recipient list is more pruned and targetted. Part of this whole social networking thing is that people can blast out data to many people and let most of them ignore it if they don't care. Explicitly thoughtless dissemination of info.
You point out that he posted examples that demonstrate how capable he is of being savvy about it, however that is when he sits down to explicitly think about it. Even for those that *would* be aware of the potential implications if they thought it through, if they get into it they are unlikely to thoroughly think through everything they post.
The notion of hundreds upon hundreds of 'friends' is ludicrous on the surface and dangerous when you consider the potential import of everyday thoughts and conversations.
Also, AOL had their own walled garden of value that evaporated with an internet that automatically looked the same regardless of whether you accessed it from home, work, or school. AOL on broadband would have been insufficient to keep them relevant (it was on broadband of course, and available anywhere you could install their software with any internet access, which was not most school and work cases).
In terms of predicting whether or not facebook or anyone else will die and when, all of the rational discussion in the world won't matter. It's a whim that is an emergent feature of a population that simply is unknowable.
The people whom you lost respect for because of stupid or melodramatic stuff don't care. A whole lot of people love stupid and even very destructive drama. Facebook is a medium where those exchanges are made more real by their typical association to real identities. This is not a problem or failing of social networking, it's a reflection of human nature and owes no small part of their success to that sort of asinine stuff.
From a technical perspective, sure, lots of companies could do this, it's not hard.
Facebook is *not* about a technical offering. A technical offering is a prerequisite, but the success is entirely predicated on networking effects. Therefore, popularity is the single most important thing. By some combination of 'getting it right', strategy, and sheer luck, they have ultimately built the most ubiquitous offering. I would say facebook is the first widespread provider to make aggregation the central feature, which was key to establish the networking effect as an intrinsic factor where it wasn't so much for geocities and myspace where you had to explicitly check in on your 'friends'.
I would like to agree with you, but I think the issue is the more bullet-proof, federated architectures of piece-parts that are readily available otherwise. If that were the case, neither MySpace nor Facebook would have come to prominence. AOL died because the internet became a superset of what AOL contained, could be accessed equally from school campuses, work, and home, and opened the door for cheaper and higher-bandwidth providers. Content providers couldn't restrict themselves to a service that most people couldn't get to from work or school, so they published at least on the internet and maybe AOL. That was the key nail in the coffin IMO.
On user provided content, the trend seems clear enough from geocities/angelfire, to myspace, to facebook. Each required less and less work from users (by sacrificing customization capability). Also, as they ditched customization, one navigating content provided by several friends is faced with a more and more consistent set of data. Of course, much of this could be hypothetically reproduced by a near monopoly releasing a distributed approach to this, but it would *have* to be on some embedded device that people wouldn't tend to sleep/turn off, eliminating the most likely candidate for that.
Another issue of this is the natural aggregation of information in facebook. The users look at one page and data provided by all of their 'friends' is aggregated into one apparent stream of data. This is the tricky part to do in a distributed fashion. Becoming a friend or fan of someone could hypothetically request an explicit push of data from one person to all friends/fans they have then and there even if that friend never reads it. However, this would scale worse than what a site like facebook deals with. The converse of scan on read would similarly be bad, with the additional problem that the experience would be apparently sluggish and miss/hiccup as peers are down. Approaches to make it appear responsive would cause 'invisible' gaps to be in the data where unavailable data is missing and magically appears when available. If not assembled by the home routers, it would have to be a federated set of content-providers. e-mail is probably the most similar model, and that carries the problem of inefficient storage. Each user gets a disparate copy of the data. Also, once out there, it cannot be edited in a single copy. Basically, people explicitly want to put all their data in a *single* authoritative place, which naturally leads to things like facebook.
Which brings us to another point, people are used to what datacenters provide in terms of reliable data delivery. Home routers would be subject to ISP outages (which are not infrequent in residential price points), power outages/blips, people just turning their stuff off, router hardware/firmware problems without elaborate failover or even someone paying particularly close attention. Rationally, a 'facebook' application isn't critical enough to consider an outage in a vacuum particularly important, but the emergent behavior of a few missing pieces would be perceived as garbage. Not to mention more work on the part of the participants to notice when their stuff isn't working and fix something.
I have dovecot, postfix, roundcube and squirrelmail running on my home system providing mail, and have other web content I share up too. I must confess I'm moving more and more of my email to gmail. My home system is pretty sufficient and I like the control, however when I'm on vacation I worry that something will happen that would require me to fix that would not be possible remotely. I also like doing a lot of things like upgrades and experimentation and the system with email responsibility I can't really mess with in a way that induces a long downtime. I would move my home domain to google and be done with it, except I use -<suffix> addresses and google only supports +<suffix> addressing.
I don't think facebook will continue indefinitely specifically.. It will be replaced by something
For the jerk and flake, there is simply more than 'intelligence' that matters, and sometimes that can offset any benefit.
The same can apply for the Heretic, but frequently the so-called heretic is right, which is a bigger problem than his apparent lack of morale.