Dalkvik doesn't interpret Java bytecode. It's register based which makes it more analogous to a virtual CPU than a JVM. Performance isn't too bad with it and I really don't have any objection to virtualized instructions. I think however that there are apps where it is unreasonable to expect someone to port them to Java source. For example C++ apps, or even Objective C for that matter.
That said, I think it is retrograde to compile to any native instruction set, even ARM. There are differences between ARM chipsets (e.g thumb instructions), and who's to say it wouldn't make more sense to use MIPS or x86 or some other chipset some other day? LLVM exists precisely so devs don't have to care too much about the hardware, compilation can be deferred until the app is installed on a device. Ship the app as bitcode and only compile to native when the thing is installed or first executed. Android has recently introduced Renderscript which does support LLVM support but its being used in a way more analogous to CUDA / OpenCL than regular C++. I would like to see the LLVM extended to support regular C++ apps.
It's painfully obvious if you're a farmer in a field. It's not so obvious to me sitting in my house where I can't even get a signal lock. What I meant to do with it anyway? Google services can fallback on wifi spots which are reasonably adequate for location and failing that it could always offer me a dialog to "stick a pin where you are" for geolocation.
Rear facing camera is for reality overlay and for just good old picture-taking.
Which would be great if I want to take pictures of my balls while browsing. Not so much use otherwise. Who is seriously going to hold up an enormous slab to take pictures?
Lots of flash is obvious as well.
Most Android phones have micro SD slots. Stick 8GB of flash into the tablet and let users augment it if they need to.
3G is obvious but let me get to that in a moment.
Again it's obvious if you have no other means of using the internet. Most people do. They're in a house or office which has wifi. Android tablet owners most likely have a phone which can tether. Failing that, Android tablets sport a USB port, so why not support popular 3G dongles?
Super high end dual core mobile processor is so I can replace my desktop with it, which will work fine for most users if it has mini-HDMI
Not everyone wants to replace their desktop. They want to browse and do casual stuff. Stuff which shouldn't tax a moderately specced tablet. I also doubt you could replace your desktop with any tablet at the moment even if you wanted to.
Barometers are cool and practically free. A compass is needed for reality overlay and also practically free.
The point is it's superfluous fluff. It takes space, power and bloats the price while having dubious application in a tablet.
Yes, that is ludicrous bullshit. But 3G is useful. I want a Kindle (but am unwilling to shell out for the small one, and unwilling to shell out so much for the big one) with 3G because it is a mobile Wikipedia terminal. Being able to look up zonation at the nursery or compatibility at the electronics store when shopping the sale counter is eminently useful.
I'm not saying 3G is useless. I've been commenting a lot recently saying how cool it would be if Sony's NGP worked like a Kindle. What I'm saying is it really adds to the price of the tablet.
Or in short, don't be a Luddite.
I'm not being a Luddite. I just don't check my brains in at the door when the latest shiny thing comes out at ludicrous prices. Tablets can and should be cheaper.
Are you implying that the Xoom costs so much because of the marketing? I wouldn't be surprised if were true though.
As for the Archos, I'm just using it to demonstrate what's going to happen this year. 2011 started with very usable tablets in the $250-300 range and it's likely that the field is going to swell considerably before the year is out.
No, not at all. The Archos 101 is a perfectly acceptable tablet for 3/5 price of the iPad. It's slim, has a long battery life, 10.1" capacitive touch screen, memory slot, 8GB on board, HDMI / USB outputs and plays pretty much every common video format. The Nook is another example of a low cost tablet designed for a job (in its case ereader) and doing it well. They're not "junk", they're perfectly fine devices.
Tablets don't need to pack in as many features as Motorola, Samsung et all seem compelled to do. Why does a tablet need GPS again? Or a rear facing camera? Or 32GB flash? Or 3G? Come to that why does it need a super high end dual core mobile processor? Or barometer / compass? The 3G alone apparently slaps $200 onto the price of the Xoom which will be a mystery to anyone who has seen 3G dongles on sale for a fraction of that.
Sure advanced features are nice, but most tablets are going to be used for playing videos, browsing, reading and don't need those high specs. Even 3G is superfluous for anyone with an Android handset which supports tethering, or owns a mifi device.
The Archos 101 demonstrates that a tablet can be affordable and reasonable specced. And I expect more examples will trickle out this year. There will be tablets to match all price points, many of which will be less than an iPad and certainly much less than a Xoom.
It would actually be in Nokia's interests to hang onto QT and spread it EVERYWHERE. If they were smart they even got MS to allow them to use it in WP7, possibly even licence it for app development. Wouldn't that be a turn up for the books?
QT isn't big enough to compete. The other juggernauts have the momentum and QT will fail. it is not because it is a bad technology - it just doesn't have the traction.
I think QT has a place on Android. Think of all those useful Linux apps you'd like to see on your tablet for example. Then there are apps migrating from Nokia's platform which could find it useful.
What I would like to see on Android is a proper alternative to the Dalvik framework. There is the NDK but it would be nice to see a proper LLVM like environment where you can write C++ code but it is turned to an intermediate bitcode and isn't tied to one platform or architecture.
HTML within a page is inherently single threaded. All JS loaded against that page runs from the same thread and manipulates the DOM, CSS etc. from that single thread. So if we had 4 HTML5 ads doing what Flash normally does with animation etc. they'd all be setting timers, e.g. every 1/30th second. Timer callbacks would be handled sequentially and time taken to do this would be taken away from the same thread. In addition all these animations would clobber the browser's layout manager with potentially 4 chunks of the page expecting to be repainted in rapid succession. I expect the browser would be dragged to its knees.
So can timers be done on different threads? No. If JS (e.g. on a timer) were permitted to run on a different thread, there would be all kinds of horrible race conditions, not just as it touched the DOM but also for any objects the timer interacted with. JS is probably too loose even to contemplate some locking mechanism where objects on one thread were locked for the duration of use to cause other threads to wait until they were released.
The only exception is the recently introduced web workers and if you read the spec you'll see they have to rely on messaging if they want to interact with the main thread at all, i.e. web workers live in their own little world, they receive messages and they can send them. If they want to do something to the DOM they have to send a message to do it and wait for the main thread to process their message. Potentially HTML5 ads could spawn a web worker which would do the heavy work in the thread and then message the main window when it needed an update. That might work assuming all browsers supported web workers which they don't currently.
I'm sure there are ways for browsers to improve performance but they are subject to these constraints. For example every browser tab could conceivably be a separate thread, especially if it has no window.opener. Cases where one window opens another and iframes is more problematic but could probably done with some marshaling.
Flash isn't evil, it's just abused. When you load a page and it has 3 or 4 flash ads and every tab in your browser is the same your computer is going to have a hernia. Some people pretend this is Flash's fault but the reality is that if pages were serving up the equivalent workload in HTML5 performance is bound to be even worse. At least the Flash plugin can spawn threads, do background rendering and so on. Everything in HTML5 on the same page will be competing on the same thread (web workers could potentially handle some load but nothing DOM related).
The remedy is to use an ad blocker so you can pick and choose what content to receive. In time I expect Ad Block will be used as much to curb the abuses of HTML5 as it is for Flash now. Assuming HTML5 ads aren't inlined and obfuscated which is a distinct possibility.
Enjoy the subpar performance and non-honeycomb interface...
You don't need to spend $800 to get performance which would be entirely adequate for watching videos, browsing, reading books. An Archos 101 satisfies the things most people need from a tablet for a price which is 3/5 of an iPad. The Nook is cheaper again. I would hope that there will be android 3.0 tablets with beef up specs hitting those price points before the year is out.
There is no avoiding that the Xoom is way too expensive. It makes the overpriced iPad look like a bargain. Perhaps it's all a cunning plan to make the price under contract look like a bargain. I wouldn't rule that out.
Every time the government changes and the Conservatives get in they start going on about this. The 'Tories' are an english party, and in England it gets dark at 10pm in the summer. Boo-hoo. It gets dark in Spain at 9pm in the summer, but they are out having beers till 3am. In England everyone is in their beds at 10pm, what do you want an extra hour of light for? Why do you want it to be light while you are asleep or in your house watching cricket and drinking warm beer or whatever it is you guys do? Especially if it's raining. I may be scottish but I've never seen so much rain as I have in Oxfordshire. No wonder you like your boats, you need them.
If it's nearly always light in Scotland, what difference does it make to shift time a bit in the summer? If it extends daylight in southern parts then that is good for a raft of reasons - tourism, lower energy consumption, and just the ambient good feeling of being sat outside drinking beer at 10pm and it still being light.
Personally I think the government needs to set a timezone which optimizes daylight during waking hours all year around. Screw farmers and people living on remote islands and consider what benefits the majority of the population.
GPL: You need to comply only if you get caught! (And even in this case, you can wait for a year or two)
Yeah, lets encourage this behavior further...
I think there would be less violations if the FSF offered some useful targeted advice to the vendors. It is the vendor who is distributing the boxes, so they're the ones on the hook to supply the source code but they're often not the ones doing the firmware. Usually they'll subcontract that out and there could be subcontractors underneath them. Everyone in the chain needs to know their obligations during requirements gathering, contract negotiation phase, deliverables and in support. The FSF should produce a document which can be handed out with the contract that says in simple terms what the obligations of a subcontractor are with regard to his customer and the deliverable. Given the proper level of education the source would find its way up the chain and there would be less violations all around.
The MeeGo SDK has cross compilers for x86 and ARM which is what I found disappointing. It implies if I were writing apps to MeeGo I'd have to compile at least once for every target architecture. If it used LLVM bitcode I wouldn't have to do this. I'd compile to an intermediate bitcode format and when it was deployed to the target device it would be compiled to the native equivalent. It largely wouldn't matter to me what the target architecture was at all. It wouldn't even have to be MeeGo - the runtimes could be ported to work over Windows, OS X, Linux which I'm sure you'll agree would have massive potential.
QT quick appears to be something analogous to Mozilla's XUL. An XML + CSS + JS framework which can doubtless do a lot of stuff but isn't suitable for compiling C++ apps.
I agree MeeGo was a non-runner. While it would be nice to see a Linux based tablet, the reality is there are already two Linux based tablet operating systems and another simply isn't going to make it. What surprises me most about MeeGo is it targets multiple architectures but doesn't think to use something like LLVM to provide a layer of abstraction.
It's a wasted opportunity since one could envisage a situation where apps were built once and ran on any MeeGo device. Or for that matter any desktop OS providing the same runtime. Without any recompilation at all. It would have been a massive deal to be able to write apps in this way.
Nokia is a brand that relies on hardware and software to differentiate its products from other manufacturers on the market. When they go to WP7 there will be precious little to differentiate their phones from anyone else. The hardware will be the facilitator for the software which will be virtually the same from one handset to the next. So why exactly would anyone prefer to buy a Nokia phone running WP7 over one with HTC, LG, Samsung or Dell? Chances are they won't care at all because all the phones will be doing pretty much the same thing. Also note that only Nokia has burned its ships, all the other WP7 providers make handsets with other operating systems on them. So if WP7 doesn't work out they can withdraw without any loss of face. Except Nokia.
It's bizarre they didn't choose Android. There would have been plenty of room for customizing the experience while sharing the core functionality with other phones. They could have decked it out with a Symbian like icons & front-end, rejigged the settings to be more familiar, tossed in Ovi store and anything else they cared to do. Very little of that will happen with WP7. They're just one of Microsoft's bitches now and must come to heel when they're called.
The iOS app store can have it's fair share of malware too. It's easy to hide snooping software behind a simple game for example. In fact, all apps can access the contacts list, recent youtube searches, email settings and even non-password field keystrokes [cnet.com]. When developers submit apps they only submit the binary and not the source code so Apple's app approval monkeys basically only cover what they can see. This "walled garden" argument is stupid for this reason.
I also expect the approval process is more geared to validating an app "works" and doesn't violate the ever-shifting, wibbly-wobbly, pernicious T&Cs. On the security side they probably just run the app through a scanner of some sort that looks for suspicious code which flags the app for human security review. I expect Android Marketplace works in a similar fashion when it comes to security reviews although they don't especially care about the app's quality or what it does (although it can't be a rival marketplace / distribution store).
As I said, it appears to be high enough. Anyone can put together malware for android and get it distributed with no investment other than the time and effort it takes. To put an app in Apple's store is not only not free, it's also not a sure thing your app will be approved. And finally, there is no money in Android, whereas quite a number of people do make a living developing iOS apps.
Well if we're going to split hairs, it costs $25 to get a licence key to use Google marketplace, although it's free to develop apps if you shove the apks on your own site. Also, you claim there is no money to be made in Android which is strange seeing as an ever increasing number of popular and well known apps are appearing there.
I certainly don't see $99 being ANY barrier to entry if a malware user wished to upload to appstore. If they make more than $99 then it's been a profitable exercise. I expect that malware could be hidden fairly easily too, e.g. within obfuscated / encrypted strings, or data files. Perhaps the malware would even start relatively innocuously doing what it's meant to do until some predetermined event (e.g. number of installs, date / time, remote command) kicks it off.
Of course, the fact that there is very little malware for iOS and tons of it for Android tends to confirm it as well. Personally, I love Linux and think it's a damned shame that Android becomes the first widely-distributed, mainstream version, because it's really a crappy OS security-wise.
I'm only aware of one high publicity case of malware on Marketplace and it was remotely killed. I also assume that Google have automated & manual security audits that they run over apps plus security teams doing analysis based on reports by users flagging apps as malicious. I assume apps are scanned as soon as they are uploaded to the store much as they probably are for the Apple Store. Depending on the threat they could retroactively kill an app, "upgrade" it into oblivion or otherwise neutralize it.
So people are extremely unlikely to get malware. And even if there is a small chance, is the risk outweighed by the benefits of the freedom of choice offered? Just the fact that I can install Opera or Firefox or Flash or a host of other apps that don't see the light of day on iOS makes it worthwhile to use Android to me.
As for security, Android has quite good security. Firstly Android uses Linux style security - processes run with their own user / group ids to stop them interfering with processes / files and there is a fine grained security model built into Dalvik too. I think it could be improved, e.g. UAC style controls of "untrusted" apps would be a huge benefit, and it would be nice if FAT32 SD cards could benefit from some kind of ACL extensions to enforce permissions. But claiming it has crappy security is not understanding what it has in the first place.
...why Apple's "Walled Garden" for the iPhone is such a bad thing?
Because it is constraining, anti open source, anti consumer, anti competitive, highly restrictive, doesn't even carry certain kinds of apps, imposes limits on other kinds of apps.
Yes there are bad things on the Internet but let's face it, unless you bought a shitty no name phone / tablet from a Chinese seller which was preloaded with some dubious alternative to Google Marketplace, you're *never* going to see this app. If by malice or misfortune it did turn up on Google Marketplace (or appslib or Amazon's tentative app store) it would probably get killed remotely.
So preferring to sit in a golden cage for fear of scary things you're never going to meet seems like an overreaction. Kind of reminds me of people stuck in AOL for so long because the internet was so scary and difficult.
Ah right, so if a computer or site's security is not perfect, the person who chooses to bust in bears no responsibility or blame for subsequent actions? Don't be stupid.
Your sarcasm is misplaced. What's the point of Sony going to the expense of providing services to legit users if a bunch of assholes are allowed to cheat rendering them useless? The answer is there is no point. Clearly Sony intend to shut modders out of PSN both as a deterrent and also to protect the service from vandalism, exploits, damage etc.
Not illegal and as for ethics, it depends who you are thinking of. From Sony's perspective it would be unethical to leave it in there and screw over their business interests & partners for a small number of people using a niche feature.
Sure. He only had to open the box, solder a bunch of connection points, insert a switch to cause deliberate hardware faults...
Yes a viable crack. He might initially have ripped a PS3 open to test the exploit but the subsequent code that he released enabled ring 0 access to Other OS which could used to read/write in memory, firmware or hdd with impunity. Sony removed Other OS to prevent people from running custom firmware installers which would have inevitably followed this exploit.
So yes you can blame Geohot for the current situation. Sony acted in an understandable and entirely predictable way. It's a shame that Other OS went but to leave it there would have jeopardized the entire platform. It's also laughable that people claim that hacking only started with its removal, it was removed BECAUSE of hacking. Maybe other exploits have subsequently surfaced but the Other OS hack was viable and if Sony had done nothing they'd probably in a worse place than they are now.
Sony could do much better working WITH the community on things like OtherOS. Instead, they've gone down the paranoid conspiracy-theorist route and look what it's gotten them - now the "pirate" scene and the homebrew scene (not always 100% separated anyways) are working on similar projects to open up the box for non-Sony code.
I doubt Sony gives a flying fuck if someone produces their own firmware to run XBMC for example. What they care about is people using modded firmware to play pirate software, cheat in multiplayer games, hack save games, hack trophies, scare off premium developers and generally screw around with something designed as a closed system.
Slashdot Mirror
I'd say give them a break. It's not hard to find firmware fuckups on other smart phones and devices.
That said, I think it is retrograde to compile to any native instruction set, even ARM. There are differences between ARM chipsets (e.g thumb instructions), and who's to say it wouldn't make more sense to use MIPS or x86 or some other chipset some other day? LLVM exists precisely so devs don't have to care too much about the hardware, compilation can be deferred until the app is installed on a device. Ship the app as bitcode and only compile to native when the thing is installed or first executed. Android has recently introduced Renderscript which does support LLVM support but its being used in a way more analogous to CUDA / OpenCL than regular C++. I would like to see the LLVM extended to support regular C++ apps.
GPS should be painfully obvious.
It's painfully obvious if you're a farmer in a field. It's not so obvious to me sitting in my house where I can't even get a signal lock. What I meant to do with it anyway? Google services can fallback on wifi spots which are reasonably adequate for location and failing that it could always offer me a dialog to "stick a pin where you are" for geolocation.
Rear facing camera is for reality overlay and for just good old picture-taking.
Which would be great if I want to take pictures of my balls while browsing. Not so much use otherwise. Who is seriously going to hold up an enormous slab to take pictures?
Lots of flash is obvious as well.
Most Android phones have micro SD slots. Stick 8GB of flash into the tablet and let users augment it if they need to.
3G is obvious but let me get to that in a moment.
Again it's obvious if you have no other means of using the internet. Most people do. They're in a house or office which has wifi. Android tablet owners most likely have a phone which can tether. Failing that, Android tablets sport a USB port, so why not support popular 3G dongles?
Super high end dual core mobile processor is so I can replace my desktop with it, which will work fine for most users if it has mini-HDMI
Not everyone wants to replace their desktop. They want to browse and do casual stuff. Stuff which shouldn't tax a moderately specced tablet. I also doubt you could replace your desktop with any tablet at the moment even if you wanted to.
Barometers are cool and practically free. A compass is needed for reality overlay and also practically free.
The point is it's superfluous fluff. It takes space, power and bloats the price while having dubious application in a tablet.
Yes, that is ludicrous bullshit. But 3G is useful. I want a Kindle (but am unwilling to shell out for the small one, and unwilling to shell out so much for the big one) with 3G because it is a mobile Wikipedia terminal. Being able to look up zonation at the nursery or compatibility at the electronics store when shopping the sale counter is eminently useful.
I'm not saying 3G is useless. I've been commenting a lot recently saying how cool it would be if Sony's NGP worked like a Kindle. What I'm saying is it really adds to the price of the tablet.
Or in short, don't be a Luddite.
I'm not being a Luddite. I just don't check my brains in at the door when the latest shiny thing comes out at ludicrous prices. Tablets can and should be cheaper.
As for the Archos, I'm just using it to demonstrate what's going to happen this year. 2011 started with very usable tablets in the $250-300 range and it's likely that the field is going to swell considerably before the year is out.
I have no idea. What's that matter?
Tablets don't need to pack in as many features as Motorola, Samsung et all seem compelled to do. Why does a tablet need GPS again? Or a rear facing camera? Or 32GB flash? Or 3G? Come to that why does it need a super high end dual core mobile processor? Or barometer / compass? The 3G alone apparently slaps $200 onto the price of the Xoom which will be a mystery to anyone who has seen 3G dongles on sale for a fraction of that.
Sure advanced features are nice, but most tablets are going to be used for playing videos, browsing, reading and don't need those high specs. Even 3G is superfluous for anyone with an Android handset which supports tethering, or owns a mifi device.
The Archos 101 demonstrates that a tablet can be affordable and reasonable specced. And I expect more examples will trickle out this year. There will be tablets to match all price points, many of which will be less than an iPad and certainly much less than a Xoom.
There will be tablets running honeycomb for $300 before the year is out. Motorola is just being greedy.
It would actually be in Nokia's interests to hang onto QT and spread it EVERYWHERE. If they were smart they even got MS to allow them to use it in WP7, possibly even licence it for app development. Wouldn't that be a turn up for the books?
QT isn't big enough to compete. The other juggernauts have the momentum and QT will fail. it is not because it is a bad technology - it just doesn't have the traction.
I think QT has a place on Android. Think of all those useful Linux apps you'd like to see on your tablet for example. Then there are apps migrating from Nokia's platform which could find it useful.
What I would like to see on Android is a proper alternative to the Dalvik framework. There is the NDK but it would be nice to see a proper LLVM like environment where you can write C++ code but it is turned to an intermediate bitcode and isn't tied to one platform or architecture.
So can timers be done on different threads? No. If JS (e.g. on a timer) were permitted to run on a different thread, there would be all kinds of horrible race conditions, not just as it touched the DOM but also for any objects the timer interacted with. JS is probably too loose even to contemplate some locking mechanism where objects on one thread were locked for the duration of use to cause other threads to wait until they were released.
The only exception is the recently introduced web workers and if you read the spec you'll see they have to rely on messaging if they want to interact with the main thread at all, i.e. web workers live in their own little world, they receive messages and they can send them. If they want to do something to the DOM they have to send a message to do it and wait for the main thread to process their message. Potentially HTML5 ads could spawn a web worker which would do the heavy work in the thread and then message the main window when it needed an update. That might work assuming all browsers supported web workers which they don't currently.
I'm sure there are ways for browsers to improve performance but they are subject to these constraints. For example every browser tab could conceivably be a separate thread, especially if it has no window.opener. Cases where one window opens another and iframes is more problematic but could probably done with some marshaling.
The remedy is to use an ad blocker so you can pick and choose what content to receive. In time I expect Ad Block will be used as much to curb the abuses of HTML5 as it is for Flash now. Assuming HTML5 ads aren't inlined and obfuscated which is a distinct possibility.
Enjoy the subpar performance and non-honeycomb interface...
You don't need to spend $800 to get performance which would be entirely adequate for watching videos, browsing, reading books. An Archos 101 satisfies the things most people need from a tablet for a price which is 3/5 of an iPad. The Nook is cheaper again. I would hope that there will be android 3.0 tablets with beef up specs hitting those price points before the year is out.
There is no avoiding that the Xoom is way too expensive. It makes the overpriced iPad look like a bargain. Perhaps it's all a cunning plan to make the price under contract look like a bargain. I wouldn't rule that out.
Every time the government changes and the Conservatives get in they start going on about this. The 'Tories' are an english party, and in England it gets dark at 10pm in the summer. Boo-hoo. It gets dark in Spain at 9pm in the summer, but they are out having beers till 3am. In England everyone is in their beds at 10pm, what do you want an extra hour of light for? Why do you want it to be light while you are asleep or in your house watching cricket and drinking warm beer or whatever it is you guys do? Especially if it's raining. I may be scottish but I've never seen so much rain as I have in Oxfordshire. No wonder you like your boats, you need them.
If it's nearly always light in Scotland, what difference does it make to shift time a bit in the summer? If it extends daylight in southern parts then that is good for a raft of reasons - tourism, lower energy consumption, and just the ambient good feeling of being sat outside drinking beer at 10pm and it still being light.
Personally I think the government needs to set a timezone which optimizes daylight during waking hours all year around. Screw farmers and people living on remote islands and consider what benefits the majority of the population.
GPL: You need to comply only if you get caught! (And even in this case, you can wait for a year or two)
Yeah, lets encourage this behavior further...
I think there would be less violations if the FSF offered some useful targeted advice to the vendors. It is the vendor who is distributing the boxes, so they're the ones on the hook to supply the source code but they're often not the ones doing the firmware. Usually they'll subcontract that out and there could be subcontractors underneath them. Everyone in the chain needs to know their obligations during requirements gathering, contract negotiation phase, deliverables and in support. The FSF should produce a document which can be handed out with the contract that says in simple terms what the obligations of a subcontractor are with regard to his customer and the deliverable. Given the proper level of education the source would find its way up the chain and there would be less violations all around.
QT quick appears to be something analogous to Mozilla's XUL. An XML + CSS + JS framework which can doubtless do a lot of stuff but isn't suitable for compiling C++ apps.
It's a wasted opportunity since one could envisage a situation where apps were built once and ran on any MeeGo device. Or for that matter any desktop OS providing the same runtime. Without any recompilation at all. It would have been a massive deal to be able to write apps in this way.
It's bizarre they didn't choose Android. There would have been plenty of room for customizing the experience while sharing the core functionality with other phones. They could have decked it out with a Symbian like icons & front-end, rejigged the settings to be more familiar, tossed in Ovi store and anything else they cared to do. Very little of that will happen with WP7. They're just one of Microsoft's bitches now and must come to heel when they're called.
The iOS app store can have it's fair share of malware too. It's easy to hide snooping software behind a simple game for example. In fact, all apps can access the contacts list, recent youtube searches, email settings and even non-password field keystrokes [cnet.com]. When developers submit apps they only submit the binary and not the source code so Apple's app approval monkeys basically only cover what they can see. This "walled garden" argument is stupid for this reason.
I also expect the approval process is more geared to validating an app "works" and doesn't violate the ever-shifting, wibbly-wobbly, pernicious T&Cs. On the security side they probably just run the app through a scanner of some sort that looks for suspicious code which flags the app for human security review. I expect Android Marketplace works in a similar fashion when it comes to security reviews although they don't especially care about the app's quality or what it does (although it can't be a rival marketplace / distribution store).
As I said, it appears to be high enough. Anyone can put together malware for android and get it distributed with no investment other than the time and effort it takes. To put an app in Apple's store is not only not free, it's also not a sure thing your app will be approved. And finally, there is no money in Android, whereas quite a number of people do make a living developing iOS apps.
Well if we're going to split hairs, it costs $25 to get a licence key to use Google marketplace, although it's free to develop apps if you shove the apks on your own site. Also, you claim there is no money to be made in Android which is strange seeing as an ever increasing number of popular and well known apps are appearing there.
I certainly don't see $99 being ANY barrier to entry if a malware user wished to upload to appstore. If they make more than $99 then it's been a profitable exercise. I expect that malware could be hidden fairly easily too, e.g. within obfuscated / encrypted strings, or data files. Perhaps the malware would even start relatively innocuously doing what it's meant to do until some predetermined event (e.g. number of installs, date / time, remote command) kicks it off.
Of course, the fact that there is very little malware for iOS and tons of it for Android tends to confirm it as well. Personally, I love Linux and think it's a damned shame that Android becomes the first widely-distributed, mainstream version, because it's really a crappy OS security-wise.
I'm only aware of one high publicity case of malware on Marketplace and it was remotely killed. I also assume that Google have automated & manual security audits that they run over apps plus security teams doing analysis based on reports by users flagging apps as malicious. I assume apps are scanned as soon as they are uploaded to the store much as they probably are for the Apple Store. Depending on the threat they could retroactively kill an app, "upgrade" it into oblivion or otherwise neutralize it.
So people are extremely unlikely to get malware. And even if there is a small chance, is the risk outweighed by the benefits of the freedom of choice offered? Just the fact that I can install Opera or Firefox or Flash or a host of other apps that don't see the light of day on iOS makes it worthwhile to use Android to me.
As for security, Android has quite good security. Firstly Android uses Linux style security - processes run with their own user / group ids to stop them interfering with processes / files and there is a fine grained security model built into Dalvik too. I think it could be improved, e.g. UAC style controls of "untrusted" apps would be a huge benefit, and it would be nice if FAT32 SD cards could benefit from some kind of ACL extensions to enforce permissions. But claiming it has crappy security is not understanding what it has in the first place.
...why Apple's "Walled Garden" for the iPhone is such a bad thing?
Because it is constraining, anti open source, anti consumer, anti competitive, highly restrictive, doesn't even carry certain kinds of apps, imposes limits on other kinds of apps.
Yes there are bad things on the Internet but let's face it, unless you bought a shitty no name phone / tablet from a Chinese seller which was preloaded with some dubious alternative to Google Marketplace, you're *never* going to see this app. If by malice or misfortune it did turn up on Google Marketplace (or appslib or Amazon's tentative app store) it would probably get killed remotely.
So preferring to sit in a golden cage for fear of scary things you're never going to meet seems like an overreaction. Kind of reminds me of people stuck in AOL for so long because the internet was so scary and difficult.
Ah right, so if a computer or site's security is not perfect, the person who chooses to bust in bears no responsibility or blame for subsequent actions? Don't be stupid.
Your sarcasm is misplaced. What's the point of Sony going to the expense of providing services to legit users if a bunch of assholes are allowed to cheat rendering them useless? The answer is there is no point. Clearly Sony intend to shut modders out of PSN both as a deterrent and also to protect the service from vandalism, exploits, damage etc.
Not illegal and as for ethics, it depends who you are thinking of. From Sony's perspective it would be unethical to leave it in there and screw over their business interests & partners for a small number of people using a niche feature.
Yes a viable crack. He might initially have ripped a PS3 open to test the exploit but the subsequent code that he released enabled ring 0 access to Other OS which could used to read/write in memory, firmware or hdd with impunity. Sony removed Other OS to prevent people from running custom firmware installers which would have inevitably followed this exploit.
So yes you can blame Geohot for the current situation. Sony acted in an understandable and entirely predictable way. It's a shame that Other OS went but to leave it there would have jeopardized the entire platform. It's also laughable that people claim that hacking only started with its removal, it was removed BECAUSE of hacking. Maybe other exploits have subsequently surfaced but the Other OS hack was viable and if Sony had done nothing they'd probably in a worse place than they are now.
Sony could do much better working WITH the community on things like OtherOS. Instead, they've gone down the paranoid conspiracy-theorist route and look what it's gotten them - now the "pirate" scene and the homebrew scene (not always 100% separated anyways) are working on similar projects to open up the box for non-Sony code.
I doubt Sony gives a flying fuck if someone produces their own firmware to run XBMC for example. What they care about is people using modded firmware to play pirate software, cheat in multiplayer games, hack save games, hack trophies, scare off premium developers and generally screw around with something designed as a closed system.