Sun (Oracle) has done a lot of really brilliant work with the JVM over the years, and there's no reason that they can't do the same for a bytecode that is practically identical when it comes to instruction sets and development use behavior. There would need to be features added and some features re-tuned, but that doesn't mean that it can't be done or that there isn't any synergy between the two products. There's only going to be so many ways of JIT optimizing common operations, and if Oracle's already done it, they just have to write the heuristic detector for the Dalvik bytecode instead of the classfile formats.
I see that they're just adding a concurrent GC into Dalvik (which is great) that has been in the JVM for years (though admittedly less important for systems with single cores). What I'm trying to say is that Oracle now has the expertise to make great language VM's and they can actually sell their expertise to those who want to pay for it. If that means that we still get a well performing and improving JVM for years to come then great for us all.
PS: There was a Windows version of UNIX, it was called SFU (Services for UNIX) which is a kernel mode API interface layer that generally mimics some variant of UNIX. I never saw it as being very popular, but its in every new Windows server shipped... Cheers
I am not a member of the JCP, but in my eyes, it was established so that solutions vendors for things like Application servers or other tools use to enhance the Java language/platform were done with everyone's best interests at heart.
It was meant to: 1. Reduce the number of competing platform standards so that vendors and developers can focus on making the best products that fulfill a standard set of tools / features (J2EE, Struts, etc..) 2. Allow external parties besides Sun to make an impact on how the Java language/platform would evolve over time
This debate in the JCP has sadly been dragged out for so long that it makes many question if the JCP has any hope of functioning as an open steering body for an open language/platform when a single dominant vendor has effectively veto powers over anything that is or could be decided in the committee. The field of use restriction on the TCK just solidifies the business realities of Sun (now Oracle) trying to freeze competition away from areas that they were actually making money with Java.
My suggestion: 1. Release OpenJDK as it is, and pretty much leave it unfettered. Don't invest a ton into the really high end features like high performance clustering etc. Release the super JVM as a paid offering or bundled with Weblogic, etc.. whatever. They can even release a Dalvik version of the JVM since they have such a rich set of experience making Java fast on desktops. 2. Make the TCK available so that any implementation of the Java platform can legitimately pass the spec (Android would still fail, but then...) 3. Sign whatever kind of deal necessary to get Oracle and Google on the same page with one another on the -Java- platform again. My suggestion would be to add a new device profile for android and allow both companies support in steering the platform. Maybe just having the platform supported through a -functional- JCP or through open handset would be a good place to go with it.
That would be cool, but it should be in the IDE instead considering that the item is in caps to signify a reserved word. If it was only changed at check-in, it would miss the whole point of having the item capitalized to begin with which is to make sure the developer of the code isn't accidentally typing reserved words into parts of the SQL that it isn't valid, or simply that the reserved word in the wrong spot is valid but changes the behavior from what the coder intended.
If you publish, you're a journalist. You may not be a -popular- journalist, but a journalist none the less. What are the professional requirements for being a journalist? Nothing.
Inspection is an interesting question. Assuming nothing is stored, and no encryption is infiltrated, is your data being compromised in any way that a half intelligent switch/router isn't already doing today?
1. The Customer's PC is not the property or under control of the ISP 2. The ISP can trivially detect the presence of 'questionable activity' like egress email in the 1000's for a consumer broadband account 3. ISP's can through deep packet inspection (if employed) easily detect the presence of well known computer viruses / exploits both ingress and egress 4. If decided to do so, an ISP can cut off a customer's line or block an IP both automated (based on some pre-defined traffic analysis) or manually due to human inspection
So we've established that assuming that an ISP has a packet inspection system handing customers (not guaranteed) that they can handle dealing with offending PC's if in fact they chose to. The government has said that this is not the responsibility of ISP's to do this, and I fully agree. The responsibility for such an action should not be on a common carrier. The one difference of opinion I have with some is that if I was the government and I wanted to make a better internet for our citizenry, I'd want to knock virus/exploit based customers out of the internet until they've taken the steps necessary to remove the infection and make the PC suitable for healthy internet activity.
In order to accomplish said law, you could: 1. Add government honeypots and detect incoming exploit requests -- If the exploit is detected, then a letter is sent to the ISP requiring take-down until the problem has been cleaned up. Multiple offenses by the same subscriber results in fees? This would put a real financial onus on end users to make sure they're operating their PC's correctly 2. The same as the first idea, but instead of just honeypots, the ISP's can use deep inspection to detect exploits ahead of time. The ISP's aren't required by law to do this, so make a law that they are required to do this, but make the ISP's compensated for doing so so that it isn't directly levied from the customers. The fees charged to the offenders would then help to offset the ongoing costs of the system. Yes, it can be exploited as an copyright enforcement tool or the like, but I'm talking best measures here and assuming that it will only encompass the exploits, etc..
Do you think they care? Maybe a few do, but the people sticking their necks out to get cut off are venomously religious and truly believe that the US and the west in general have wronged them and their god and the only way to escape the tyranny of their rule is to fight back with holy jihads by killing or blowing things up, etc..
The religion is totally messed up in its modern belief systems, but that can be said for so many other faith based groups over the life of religion on Earth. The point I want to make is that the west can look at the problem pragmatically and coin terms like "a little terror goes a long way" but the people who are committing these attacks probably don't care about economics as much as the price paid in blood, or maybe the fame, but I'm somewhat skeptical of that motive.
1. Get hired as a sales rep at a major security systems vendor 2. Find a flimsy but potential hole in the current security process of a given country (hopefully a reactive country that only fights fires when they're on their doorstep) 3. Start developing a solution for said problem 4. Hire a shady business person loosely associated with a criminal or terrorist group to orchestrate an 'act of terror' using said exploit and offer $10mil for 'security consulting' or the like 5. Start knocking on doors about selling your newly developed product 6. Wait 7. Reap the billions the gov will throw at you to make their latest problem go away
How? How many clients will actually work their way up the chain to resolve against the hosted DNS server? That makes any initial engagement with raw (or cache expired) domains much slower. For a web site that is a looking for drive by service, this would be less appealing than say going to a Google derived alternative which is always well buried in cache. If you really want is a way of verifying that the upstream data source isn't tampered with, and I'm sorry but that's not going to happen, at least not on a root server level.
After reading the article, its still entirely unclear. There's a person referred only as Zmijewski who is never given context at all in the story. Their talk points are half the story and you don't even have the wit to say who the person is.
Going back to the original US document, it seems the Chinese root server was erroneously sending censored responses to non-chinese IP blocks and was for a while pulled of its authority until the problem was resolved. As bad as national censorship can be, I suppose its acceptable to be able to pull the cord on issues of the sort. After all the news of having the US seize domains, is it really worth noting a bug in the great firewall's DNS processing that was fixed months ago?
I completely agree. For the large segment of multiplayer Wii games (of which I own a few) they work well for local coop. I wish to amend through that these games are almost never split screen in nature(beside Mario kart which I've never played local multiplayer), and the games themselves lend to the ability to have either all players in screen at the same time, or they have a variation of turn based action.
What really makes me sad these days aren't so much the graphical sacrifices that PC gamers have to suffer from, its the control limitations. Whenever I play Mass Effect, I feel like punching the developers, because its perfectly clear that whomever designed it was only targeting console game play mechanics.
I think you'll find that any PC being sold the last few years has higher system specs than any console today as long as the PC was sold with any sort of discrete video card. I've got a PC that plays modern PC games and it cost me around $700. Because I can do so, much of that cost won't be renewed because I can reuse several of the pieces for my next generation or even the one after that.
"I fail to see what hardware has to do with creativity anyways" Hardware doesn't in itself encourage creativity, but it sure can stifle it. If I want to create an amazing live action CG rendering of something fantastic and beautiful, it would suck if my engine supported a max of 10 objects on the screen at once. Bad example, but the analogy still applies. Unlimited hardware allows artists and designers a wider canvas to express themselves and the potential restriction of said expression is frustrating. Imagine you were Square developer making final fantasy XIII and your boss tells you: "Oh, by the way Yoshi, this has to run on a Wii since we decided to do a 3 console release." The designer would be rolling over in frustration, or else you'd have 15 DVD's with all pre-rendered scenes for everything, and the game can only be played on rails, etc..
It may just be me, but how many people play 4 player split screens for games these days? I thought it was bad during the time when tube TV's were standard, but how many games these days support more than 1/2 players simultaneously? This is of course ignoring platformer type games that can actually get away with having a few players on screen at the same time without degrading usability. Racing games? Split/Squinty screen. FPS? Split/Squinty screen. RTS? Not on consoles. Puzzle games? Sure, because they generally don't need spacial relevance.
I don't like software patents in general, but I can't really complain about the process of analyzing and granting patents beyond my notes below.
1. The software development ecosystem moves so fast that any normal developer would never file patents for work they do in developing their products. The cost and hassle of filing just isn't worth the lag when there's so much more money to be made by making products instead. 2. A well laid patent in any given area can have a severe impact on an industry that has a VERY hard time dealing with toll booths in software development. 3. OSS software is almost completely in direct opposition to patents, and OSS is working well and growing in the software development world. 4. 15 years or whatever the timeline is just seems to be excessively too long for software road-blocks, especially for patents that are so broad that they can't be written around. 5. The speed of innovation in computer science doesn't seem to be cooling down like many industries, and it means that USPO analyzers need to be very up to date with all facets of the software development industry in order to properly accept patents. The granting and eventual legal wrangling over patents seems to indicate that examiners just can't keep up with the industry.
I don't really have experience with the patent system in general (so this may already be the case) but it would make sense to me if there was a period of a patent's approval process where the public could issue their comments to the contents of the patent and if there's a valid problem with the patent, the examiner would send the request back to the submitter to fix up the submission. This would at least give interested parties early access to the patent before the submission is approved. Of course the system would need to be setup in a way to prevent competitors from (basically) permanently stalling the process. The one thing that is important about adding public interaction is that they'd also need a way of categorizing and notifying the public when there are new patents in given fields of interest / product sectors / etc.. so that the people who can have valid input into the system are given the opportunity to do so proactively. This last point could hypothetically be performed by an independent 3rd party, but you're spending so much for the patent system already, why not get some public good out of it?
Why in all the earth would the company give you money? Any money given to you can never affect the patent's validity. The only power the OP -could- wield would be to tell the defendants in a potential patent suit about it, but unless the lawyers are nitwits, they should be able to find that out through due diligence and discovery. Patent suits last years because they do a LOT of research for just this type of thing.
Secondly, lets assume that the patent is 100% covered by the software. Then what? The patent would effectively be dead and IBM would have no incentive to use the patent ever, so why would money threats against the company work exactly? It sounds more like extortion to me.
Looking through the comments up to now, it seems that the patent could still be valid, as the examiner themselves went out and found the OP's software and decided it wasn't exactly prior art. Looking through comments relating to dynamically loading the code during runtime without interfering with the original, it seems very likely that the OP's software doesn't invalidate the patent anyways.
Taking a look at diverging pieces of the FPS genre, there's totally a place to grow for people more engaged in the strategy aspect of the titles than the twitch action. I was in love with the play style of Natural Selection (and hoping the sequel can live up to the original). The fact that one can take control of a team RTS-style means that there's some room for an overview style control of a battle grid where some really interesting gameplay can occur.
I'm a little late to the discussion, but I'll throw in anyways.
The really important facet of what a Facebook alternative should look like is the ability to dis-intermediate the service from me and my use of the data that is collected about me. Facebook has barely supported an export feature, but removing my data from what is essentially a social connection tool to others is not a plan.
Example: I own my cell phone, but I can choose to move myself, my data, (and in most places my phone number) to a different carrier. That means that the separation of the carrier in itself doesn't break my ability to communicate with friends or family through a mobile device. As it stands with social networks, if you're all on the same network, you can talk to one another. If you decide A and my sister decides B then there's no communication flow, and the ability to interact comes to an end.
The ability to make an alternative Facebook is important in the ability to further control what I do with my own data, the ability to use my entered data outside of some company's pervue, and to have a service that I can easily add, interact with people and not feel like I'm tied to something I don't like. Facebook is a closed ecosystem. They consume content and lock it up from prying eyes. If Diaspora has or will have support for open inter-operating service offerings then great, otherwise they're just building another Facebook wanna be to take over the world. Who cares if Diaspora's code is Open Source if my interaction with the system and my data is shackled behind a single company's vision of how social networking should work?
I really wouldn't want to leave an iPad or anything like it in my car. If the unit looks flush and well protected by the dash, there's a much better chance that thieves won't want to dig into it. Does a stereo slot support 9-10"? I doubt it. Does it look like an iPad when a tief looks through your window? If so, expect a big broken window. Any extra cost for a manufacturer's integration is a lot better than having to constantly remove the unit from my car whenever I park my car in public.
At least with a self-built unit, you can make it look like its a manufacturer's unit, but you also need a way of syncing and updating the unit in a way that is simple and straight forward.
Archos, Zune, etc.. There are other little guys in the market as well, but I can't remember them off the top of my head. Archos has pegged on Android, so they'll probably get app store access and Zune is tied to the Microsoft ecosystem for better or worse.
"EC2? Yep. All of your financial reports and graphs will look just great coming from somebody else's data store. "
Really? Who cares. There are very few CDN's in the world and unless you're Google, who cares if your financial reports are being hosted through, Google, Akamai, etc..
I just attempted to replay Homeworld 1/2 which I had such fond memories for. Sadly, they feel barely playable and amazingly primitive in today's interface world. Its not an FPS, but definitely in need of a make over.
Slashdot Mirror
Sun (Oracle) has done a lot of really brilliant work with the JVM over the years, and there's no reason that they can't do the same for a bytecode that is practically identical when it comes to instruction sets and development use behavior. There would need to be features added and some features re-tuned, but that doesn't mean that it can't be done or that there isn't any synergy between the two products. There's only going to be so many ways of JIT optimizing common operations, and if Oracle's already done it, they just have to write the heuristic detector for the Dalvik bytecode instead of the classfile formats.
I see that they're just adding a concurrent GC into Dalvik (which is great) that has been in the JVM for years (though admittedly less important for systems with single cores). What I'm trying to say is that Oracle now has the expertise to make great language VM's and they can actually sell their expertise to those who want to pay for it. If that means that we still get a well performing and improving JVM for years to come then great for us all.
PS: There was a Windows version of UNIX, it was called SFU (Services for UNIX) which is a kernel mode API interface layer that generally mimics some variant of UNIX. I never saw it as being very popular, but its in every new Windows server shipped... Cheers
I am not a member of the JCP, but in my eyes, it was established so that solutions vendors for things like Application servers or other tools use to enhance the Java language/platform were done with everyone's best interests at heart.
It was meant to:
1. Reduce the number of competing platform standards so that vendors and developers can focus on making the best products that fulfill a standard set of tools / features (J2EE, Struts, etc..)
2. Allow external parties besides Sun to make an impact on how the Java language/platform would evolve over time
This debate in the JCP has sadly been dragged out for so long that it makes many question if the JCP has any hope of functioning as an open steering body for an open language/platform when a single dominant vendor has effectively veto powers over anything that is or could be decided in the committee. The field of use restriction on the TCK just solidifies the business realities of Sun (now Oracle) trying to freeze competition away from areas that they were actually making money with Java.
My suggestion:
1. Release OpenJDK as it is, and pretty much leave it unfettered. Don't invest a ton into the really high end features like high performance clustering etc. Release the super JVM as a paid offering or bundled with Weblogic, etc.. whatever. They can even release a Dalvik version of the JVM since they have such a rich set of experience making Java fast on desktops.
2. Make the TCK available so that any implementation of the Java platform can legitimately pass the spec (Android would still fail, but then...)
3. Sign whatever kind of deal necessary to get Oracle and Google on the same page with one another on the -Java- platform again. My suggestion would be to add a new device profile for android and allow both companies support in steering the platform. Maybe just having the platform supported through a -functional- JCP or through open handset would be a good place to go with it.
That would be cool, but it should be in the IDE instead considering that the item is in caps to signify a reserved word. If it was only changed at check-in, it would miss the whole point of having the item capitalized to begin with which is to make sure the developer of the code isn't accidentally typing reserved words into parts of the SQL that it isn't valid, or simply that the reserved word in the wrong spot is valid but changes the behavior from what the coder intended.
Apparently they still do, but you have to pay crazy amounts of money to keep them.
If you publish, you're a journalist. You may not be a -popular- journalist, but a journalist none the less. What are the professional requirements for being a journalist? Nothing.
Inspection is an interesting question. Assuming nothing is stored, and no encryption is infiltrated, is your data being compromised in any way that a half intelligent switch/router isn't already doing today?
1. The Customer's PC is not the property or under control of the ISP
2. The ISP can trivially detect the presence of 'questionable activity' like egress email in the 1000's for a consumer broadband account
3. ISP's can through deep packet inspection (if employed) easily detect the presence of well known computer viruses / exploits both ingress and egress
4. If decided to do so, an ISP can cut off a customer's line or block an IP both automated (based on some pre-defined traffic analysis) or manually due to human inspection
So we've established that assuming that an ISP has a packet inspection system handing customers (not guaranteed) that they can handle dealing with offending PC's if in fact they chose to. The government has said that this is not the responsibility of ISP's to do this, and I fully agree. The responsibility for such an action should not be on a common carrier. The one difference of opinion I have with some is that if I was the government and I wanted to make a better internet for our citizenry, I'd want to knock virus/exploit based customers out of the internet until they've taken the steps necessary to remove the infection and make the PC suitable for healthy internet activity.
In order to accomplish said law, you could:
1. Add government honeypots and detect incoming exploit requests -- If the exploit is detected, then a letter is sent to the ISP requiring take-down until the problem has been cleaned up. Multiple offenses by the same subscriber results in fees? This would put a real financial onus on end users to make sure they're operating their PC's correctly
2. The same as the first idea, but instead of just honeypots, the ISP's can use deep inspection to detect exploits ahead of time. The ISP's aren't required by law to do this, so make a law that they are required to do this, but make the ISP's compensated for doing so so that it isn't directly levied from the customers. The fees charged to the offenders would then help to offset the ongoing costs of the system. Yes, it can be exploited as an copyright enforcement tool or the like, but I'm talking best measures here and assuming that it will only encompass the exploits, etc..
*I'm going to get flamed for this....*
Do you think they care? Maybe a few do, but the people sticking their necks out to get cut off are venomously religious and truly believe that the US and the west in general have wronged them and their god and the only way to escape the tyranny of their rule is to fight back with holy jihads by killing or blowing things up, etc..
The religion is totally messed up in its modern belief systems, but that can be said for so many other faith based groups over the life of religion on Earth. The point I want to make is that the west can look at the problem pragmatically and coin terms like "a little terror goes a long way" but the people who are committing these attacks probably don't care about economics as much as the price paid in blood, or maybe the fame, but I'm somewhat skeptical of that motive.
1. Get hired as a sales rep at a major security systems vendor
2. Find a flimsy but potential hole in the current security process of a given country (hopefully a reactive country that only fights fires when they're on their doorstep)
3. Start developing a solution for said problem
4. Hire a shady business person loosely associated with a criminal or terrorist group to orchestrate an 'act of terror' using said exploit and offer $10mil for 'security consulting' or the like
5. Start knocking on doors about selling your newly developed product
6. Wait
7. Reap the billions the gov will throw at you to make their latest problem go away
How? How many clients will actually work their way up the chain to resolve against the hosted DNS server? That makes any initial engagement with raw (or cache expired) domains much slower. For a web site that is a looking for drive by service, this would be less appealing than say going to a Google derived alternative which is always well buried in cache. If you really want is a way of verifying that the upstream data source isn't tampered with, and I'm sorry but that's not going to happen, at least not on a root server level.
After reading the article, its still entirely unclear. There's a person referred only as Zmijewski who is never given context at all in the story. Their talk points are half the story and you don't even have the wit to say who the person is.
Going back to the original US document, it seems the Chinese root server was erroneously sending censored responses to non-chinese IP blocks and was for a while pulled of its authority until the problem was resolved. As bad as national censorship can be, I suppose its acceptable to be able to pull the cord on issues of the sort. After all the news of having the US seize domains, is it really worth noting a bug in the great firewall's DNS processing that was fixed months ago?
"That and a lot of rich people spent a large amount of money to defeat it"
Your democracy in action? Where was the public outrage over special interest groups effectively buying votes?
I completely agree. For the large segment of multiplayer Wii games (of which I own a few) they work well for local coop. I wish to amend through that these games are almost never split screen in nature(beside Mario kart which I've never played local multiplayer), and the games themselves lend to the ability to have either all players in screen at the same time, or they have a variation of turn based action.
What really makes me sad these days aren't so much the graphical sacrifices that PC gamers have to suffer from, its the control limitations. Whenever I play Mass Effect, I feel like punching the developers, because its perfectly clear that whomever designed it was only targeting console game play mechanics.
I think you'll find that any PC being sold the last few years has higher system specs than any console today as long as the PC was sold with any sort of discrete video card. I've got a PC that plays modern PC games and it cost me around $700. Because I can do so, much of that cost won't be renewed because I can reuse several of the pieces for my next generation or even the one after that.
"I fail to see what hardware has to do with creativity anyways"
Hardware doesn't in itself encourage creativity, but it sure can stifle it. If I want to create an amazing live action CG rendering of something fantastic and beautiful, it would suck if my engine supported a max of 10 objects on the screen at once. Bad example, but the analogy still applies. Unlimited hardware allows artists and designers a wider canvas to express themselves and the potential restriction of said expression is frustrating. Imagine you were Square developer making final fantasy XIII and your boss tells you: "Oh, by the way Yoshi, this has to run on a Wii since we decided to do a 3 console release." The designer would be rolling over in frustration, or else you'd have 15 DVD's with all pre-rendered scenes for everything, and the game can only be played on rails, etc..
It may just be me, but how many people play 4 player split screens for games these days? I thought it was bad during the time when tube TV's were standard, but how many games these days support more than 1/2 players simultaneously? This is of course ignoring platformer type games that can actually get away with having a few players on screen at the same time without degrading usability. Racing games? Split /Squinty screen. FPS? Split /Squinty screen. RTS? Not on consoles. Puzzle games? Sure, because they generally don't need spacial relevance.
I don't like software patents in general, but I can't really complain about the process of analyzing and granting patents beyond my notes below.
1. The software development ecosystem moves so fast that any normal developer would never file patents for work they do in developing their products. The cost and hassle of filing just isn't worth the lag when there's so much more money to be made by making products instead.
2. A well laid patent in any given area can have a severe impact on an industry that has a VERY hard time dealing with toll booths in software development.
3. OSS software is almost completely in direct opposition to patents, and OSS is working well and growing in the software development world.
4. 15 years or whatever the timeline is just seems to be excessively too long for software road-blocks, especially for patents that are so broad that they can't be written around.
5. The speed of innovation in computer science doesn't seem to be cooling down like many industries, and it means that USPO analyzers need to be very up to date with all facets of the software development industry in order to properly accept patents. The granting and eventual legal wrangling over patents seems to indicate that examiners just can't keep up with the industry.
I don't really have experience with the patent system in general (so this may already be the case) but it would make sense to me if there was a period of a patent's approval process where the public could issue their comments to the contents of the patent and if there's a valid problem with the patent, the examiner would send the request back to the submitter to fix up the submission. This would at least give interested parties early access to the patent before the submission is approved. Of course the system would need to be setup in a way to prevent competitors from (basically) permanently stalling the process. The one thing that is important about adding public interaction is that they'd also need a way of categorizing and notifying the public when there are new patents in given fields of interest / product sectors / etc.. so that the people who can have valid input into the system are given the opportunity to do so proactively. This last point could hypothetically be performed by an independent 3rd party, but you're spending so much for the patent system already, why not get some public good out of it?
Why in all the earth would the company give you money? Any money given to you can never affect the patent's validity. The only power the OP -could- wield would be to tell the defendants in a potential patent suit about it, but unless the lawyers are nitwits, they should be able to find that out through due diligence and discovery. Patent suits last years because they do a LOT of research for just this type of thing.
Secondly, lets assume that the patent is 100% covered by the software. Then what? The patent would effectively be dead and IBM would have no incentive to use the patent ever, so why would money threats against the company work exactly? It sounds more like extortion to me.
Looking through the comments up to now, it seems that the patent could still be valid, as the examiner themselves went out and found the OP's software and decided it wasn't exactly prior art. Looking through comments relating to dynamically loading the code during runtime without interfering with the original, it seems very likely that the OP's software doesn't invalidate the patent anyways.
"minus the "twitch" genre such as FPSs"
Taking a look at diverging pieces of the FPS genre, there's totally a place to grow for people more engaged in the strategy aspect of the titles than the twitch action. I was in love with the play style of Natural Selection (and hoping the sequel can live up to the original). The fact that one can take control of a team RTS-style means that there's some room for an overview style control of a battle grid where some really interesting gameplay can occur.
Windows for example. Why is this 'news' again?
I'm a little late to the discussion, but I'll throw in anyways.
The really important facet of what a Facebook alternative should look like is the ability to dis-intermediate the service from me and my use of the data that is collected about me. Facebook has barely supported an export feature, but removing my data from what is essentially a social connection tool to others is not a plan.
Example:
I own my cell phone, but I can choose to move myself, my data, (and in most places my phone number) to a different carrier. That means that the separation of the carrier in itself doesn't break my ability to communicate with friends or family through a mobile device. As it stands with social networks, if you're all on the same network, you can talk to one another. If you decide A and my sister decides B then there's no communication flow, and the ability to interact comes to an end.
The ability to make an alternative Facebook is important in the ability to further control what I do with my own data, the ability to use my entered data outside of some company's pervue, and to have a service that I can easily add, interact with people and not feel like I'm tied to something I don't like. Facebook is a closed ecosystem. They consume content and lock it up from prying eyes. If Diaspora has or will have support for open inter-operating service offerings then great, otherwise they're just building another Facebook wanna be to take over the world. Who cares if Diaspora's code is Open Source if my interaction with the system and my data is shackled behind a single company's vision of how social networking should work?
I really wouldn't want to leave an iPad or anything like it in my car. If the unit looks flush and well protected by the dash, there's a much better chance that thieves won't want to dig into it. Does a stereo slot support 9-10"? I doubt it. Does it look like an iPad when a tief looks through your window? If so, expect a big broken window. Any extra cost for a manufacturer's integration is a lot better than having to constantly remove the unit from my car whenever I park my car in public.
At least with a self-built unit, you can make it look like its a manufacturer's unit, but you also need a way of syncing and updating the unit in a way that is simple and straight forward.
Archos, Zune, etc.. There are other little guys in the market as well, but I can't remember them off the top of my head. Archos has pegged on Android, so they'll probably get app store access and Zune is tied to the Microsoft ecosystem for better or worse.
"EC2? Yep. All of your financial reports and graphs will look just great coming from somebody else's data store. "
Really? Who cares. There are very few CDN's in the world and unless you're Google, who cares if your financial reports are being hosted through, Google, Akamai, etc..
I just attempted to replay Homeworld 1/2 which I had such fond memories for. Sadly, they feel barely playable and amazingly primitive in today's interface world. Its not an FPS, but definitely in need of a make over.