iPhones are still really the king of the smartphone world but they are more the Microsoft Windows of the smartphone world. They have a vast array of applications and the user experience is pretty consistent across their devices which makes it easy for users.
Their flagship, "best we can do" phone is really just a copy of the Essential Phone's design and having a fingerprint reader, even if it was relegated to the back of the phone, would have made sense particularly for things like Apple Pay and for unlocking the phone when you don't want to bring it up in front of your face.
My next phone will most likely still be an iPhone simply because I see no reason to switch, everything works fine and there's no really compelling alternative - maybe I just have a relatively boring or average set of requirements for a smartphone - but if they are continuing this path it will be the iPhone 9 (or whatever they call the successor to the iPhone 8) rather than whatever the next iPhone X.
There's already too much Chrome-specific stuff and the point is: it's growing. Sure, Chrome started as fully standards-based, but then so did IE in the (very) early days when it was the best browser around. Then the years went on, and the IE-specific stuff grew until we had the world of IE6.
This is the problem with the standards organizations, companies do not want to be limited by the standard but adding to it takes far too long. So instead you get companies like Google and Microsoft creating technologies that enable new features beyond the standard. Microsoft had this with IE and ActiveX, now web standards have caught up in terms of functionality such that Microsoft has been able to abandon proprietary extensions because the standard is capable. This will likely come around for Google as well but it will take time.
You lose a significant amount of performance by connecting the GPU that way just on the bus speed alone, that's even before you get to the inability to leverage Infinity Fabric or NVLINK which are incredibly important to multi-GPU computing. The performance gap of communicating between GPUs over Thunderbolt VS PCIe is bad enough but compared to Infinity Fabric or NVLINK it's a complete non-starter.
If you've actually used the solution you suggest then you already know the performance is absolutely woeful.
I did a very quick comparable on the HP z840 to the base iMac pro and it was about $4665...it isn't exact, but is close to what Apple purports the base iMac pro to be. And also, the iMac pro includes the 5K 27" monitor which you would have to add to the HP or other comparable unit.
But you also get expandability, and of particular interest in your specific case Resolve leverages multiple GPUs so if things start to slow down you can simply add another GPU to the HP system and you're good to go, to improve the performance on the iMac you have to scrap the entire system, including the display, and buy a new one.
That's not going to be the case for everybody but certainly in your case the choice of something expandable is definitely the right one. If you really just want an Apple then holding out for the improved Mac Pro would be the right choice over the iMac.
If you have something very narrow and specific, you might be able to run.NET on Linux or OS X. If you expect to do a full-fledged client app, good luck with that.
Why? We're talking about the open source initiatives around.Net, if you use those open source bits there's no reason you can't run them on various different platforms and you can couple it with a platform-agnostic GUI toolkit (since the winforms bit of.Net is not open source) if you need a GUI. If you're going to use closed-source bits that are only available on one platform then yes you're not going to be able to run that on multiple platforms, but that is obvious already and isn't really related to this discussion. The open source bits are the things that are neutral and that is what solves the vendor (or rather platform) lock-in issue.
While many open source licenses provide this benefit (BSD, MPL), they also allow for the source code to be rolled into proprietary products (such as OsX).
I think perhaps it is you who misunderstands Free Software, just because a derivative work could be non-free does not make the original work non-free. Restrictive open source licenses are not a requirement for Free Software.
By guaranteeing freedom to downstream users, GPL maintains these four freedoms for all and forever.
Nowhere is that a requirement for Free Software, indeed the permissive open source licenses you mention are Free Software licenses precisely because they satisfy the 4 conditions you list. TensorFlow, for example, is Free Software.
Yes there is user/community benefit to having open source even on a proprietary platform - another example is Microsoft's various open source initiatives around.NET. But, what it does not solve is the matter of lock-in.
Of course it does, you don't have to run Google's open source TensorFlow on Google's cloud infrastructure, you can run it on Azure, EC2, your own cluster or even just your own local system. Likewise the open source.Net technologies don't require Microsoft's Windows platform, you can run them on Linux if you want to and the code itself is open and available. How exactly is it you think you are locked to a vendor?
I use a Macbook because MacOS is Unix that "just works".
As much as the various hardware limitations, software limitations (max of OpenGL 4.1 in 2017? Come on) and closed, proprietary technology over open standards (Metal over Vulkan, Airplay, Airdrop, Facetime, etc) is annoying I do like that the kernel can be updated without it breaking the display driver and having to go into a terminal to recompile the kernel module just to get the GUI working again. It's all those niche little annoyances that still persist all across Linux that add up to it being a poor user experience. Macs aren't perfect but on a day-to-day basis I often prefer to deal with their flaws than Linux's.
If that's the case then fine, I don't really care whether it's DisplayPort or HDMI and I'm not suggesting one is better than the other just that advancing standards like this is important and has valid use cases. At the moment DisplayPort does not have a standard that supports 8k @ 120Hz, it's simply on the roadmap for a future version (which may or may not be the next version) but if they manage to specify it, add it to the standard and ship it faster than HDMI then great, if not then I don't really care either.
HDMI 2.1 can do 8k @ 120Hz with 3:1 DSC, DisplayPort 1.4 can only do 8k @ 60Hz with the same compression ratio. The next version of DisplayPort should match HDMI 2.1 though. 120Hz is VERY important in VR.
In any case I said 'standards like this', whether that is DisplayPort or HDMI doesn't really matter.
I think that _you_ missed the point. When the existing standards have bandwidth requirements that are beyond the ability of content providers to distribute and there is virtually no planned upgrade path, further upgrades to that standard are of little/no use.
The ability to have VR headsets with a pair of 4k screens that you can drive with stereoscopic visuals is certainly a reason to have standards like this.
Well with replaceable/expandable components there's no reason they shouldn't last a very long time indeed, add that there's huge amounts of existing stock. If you like it, invest in a couple of those now for the event of catastrophic failure and you should be good beyond the next decade or so.
The simple answer is that people don't need to upgrade their phones all the time, they're good enough now. As you say, the S5 already has all the features you want.
The company sort of had conflicting goals of its own... they'd rather have sold them a groupware solution to replace Exchange, and ultimately the Linux migration projects were seen as a way to get more customers to use their own back-end solutions and server hardware. Having a perfect integration into exchange wasn't seen as a desirable outcome.
Right so when the consulting company isn't willing to present the ideal solution you can hardly blame the company for not coming up with it themselves.
Current versions of windows still require a hotfix if you have users with more than 120 AD groups.
Yes, as I said neither is perfect but you're just trading one set of problems for a different set of problems which may or may not be worse depending on many many variables and going through a painful migration process to do so. You can trade problems with the various options all day, that doesn't get you anywhere though unless you can somehow enumerate all the problems of all the options and weigh them all up for a particular use case.
I may be taking my use case (and the use case of my current market segment) for a generality, but I stopped caring about what OS/software I am using as long as I have uniform ability to access my data. If I have an external requirement for an application-specific format at some point (external entity insisting on PDF or word), the data gets converted to that format when needed but is stored as application-agnostic as possible.
Which is fine for simple things like documents, but it is not suitable for the various types of data involved in the media & entertainment, construction and manufacturing industries for example. For office workers, by and large that strategy might work. But that isn't an argument for Linux, it's an argument for 'don't care' which ultimately means the incumbent because the incumbent covers not just the general case but the niche cases as well.
The problem is that, when a great F/OS application is written for Linux, it's usually profitable to port it to Windows.
Yes so the operating system itself needs to offer something, some innovation above and beyond what Windows can offer. If there's no innovation then of course nobody is going to switch. The whole notion of free software is that it's ok to give away your 'secret sauce' because you capitalized on getting to market first and people are using your product.
None of these are issues in a corporate environment because the user won't install the OS himself and won't have administrative rights on his machine. The machine will be installed with the corporate image that already has all the required components installed and configured.
Right but these are the sorts of issues that crop up all the time, I'm not saying this is going to apply to everybody but niche cases like this (and there are hundreds of them) are one of the things that keeps people away from Linux and increases the support cost of Linux.
The biggest issue I was facing at the time was that the Exchange connector of Evolution didn't cope well with large scale Exchange deployments where you had more than one exchange server... so, for example, the calendar free/busy was only working for users that were on the same exchange server as you... this would have been a trivial fix just requiring an extra lookup per user in the calendar, but nobody was interested at the time.
Being in pre-sales did the company you were working for not see value in investing in such a thing? Particularly if it really was so trivial?
99% of the issues with Open Office had been caused by improper Office template design and were very quickly fixed. Some other interesting issues were caused by the way Microsoft overloaded their Kerberos tickets by enumerating the group membership of the user in a comment field of the ticket... users that are members of more than 120 AD groups generate tickets above 64K, which is creating funny issues in all kind of places (web services, IPSEC, login,...).
But this, and your Exchange issue above and the issues I had and all the other niche cases and "easy fixes" are the things that keep people and companies away. Why use a tool that doesn't work properly and requires fixing when you can just pay for a tool that does work properly?
And I'm not saying Windows is anywhere near perfect but you just end up trading one set of problems for a different set of problems and arguing that it will be cheaper in the long run and maybe that's true but when you go to companies and tell them they may need to employ or contract software development teams to fix, improve or develop the applications they use that raises alarm bells.
What has been demonstrated with smartphones/tablets is that software compatibility and a stable UI are actually a red herring.
In terms of software compatibility that's simply not true, the primary reason people use Windows over Linux on the desktop is application compatibility, it's the same reason people use Android over Windows Phone on mobile. In terms of a stable UI I agree, in fact software compatibility is so important that Microsoft was able to completely flip over the apple cart in terms of UI and people still used it because their applications ran on it.
> As for the other 300 programs, what are they? Do they all have Linux alternatives?
Yes.:)
What are they? They've said there are hundreds of programs that they use that do not run on Linux and you, seemingly without knowing anything about what those programs are, insist that there are alternatives. Where did you get the list of those programs from to be able to be so confident of that?
I happened to have an iMac and wanted a Linux machine (dual boot) so I tried to install it. The whole point is the clumsiness of the process of running Linux on it compared to macOS and Windows.
Wait, what? Are fucking nuts? Most of the users are software developers.
So why is application support worse on Linux than it is on macOS and Windows? There's very little software that is Linux exclusive yet there is lots and lots of desktop software that runs on macOS and/or Windows but not on Linux, a lot of it is the sort of software that is important across big industries like entertainment, construction and manufacturing. There's no technical reason you couldn't develop those applications for Linux but most users are not software developers and it's usually several orders of magnitude cheaper to buy a license for a program that already exists that does what you want than to contract a team of developers to build an alternative just because you want to run Linux as your operating system.
"According to the Document Foundation (parent organization of LibreOffice), this step back to Microsoft will cost around 90M of taxpayer’s money."
Oh that couldn't possibly be biased. While they may have saved the license fees the workarounds, the non-working applications and the need to invest huge sums of money developing alternative applications is likely what is driving the switch back.
In theory many councils of many countries could switch to open source and collaborate on building applications together, everybody would share the cost and responsibility for development and maintenance and it would be all warm and fuzzy. It's a lovely sentiment, but reality is simply not like that.
They're trying to keep using Microsoft Exchange and 300 other Windows programs, on Linux.
You can use Microsoft Exchange on Linux, there's a bunch of different clients that work with it. As for the other 300 programs, what are they? Do they all have Linux alternatives? If not what you need to do is to develop those programs, it seems they've decided that is cost-prohibitive, which is probably true.
You have a corporation who's business model builds upon not playing nice with the competition.
You mean like how they have their applications on Windows, macOS, iOS, Android and web-based platforms that make products like even their flagship Office 365 available on desktop Linux? There's mail clients like DavMail and Hiri that work on Linux and operate with Microsoft's Exchange email server. Even.Net Core is open source and supported on Linux allowing even more cross platform applications.
Linux has a "the chicken or the egg" problem. Without a sizable desktop user pool in whatever industry you're trying to switch over, all these proprietary products won't be written for Linux. Without proprietary products, industry is handicapping themselves by moving to Linux. It's a tough problem to solve, and that's why it hasn't really been solved yet.
What you need is a reason for people to use Linux. The argument that it's cheap is a non-starter because you need to contribute resources to maintaining it and to developing the applications that you need that don't yet exist for it. The motivation for doing that would be some killer, disruptive feature like the way iOS and Android killed off Blackberry and Windows Mobile =6. Then of course Microsoft suffered the same problem that desktop Linux is having with Windows Phone, not actually a bad product but not innovative or compelling enough to attract users from the incumbents.
Slashdot Mirror
Microsoft isn't betting against it, they just aren't betting on it.
iPhones are still really the king of the smartphone world but they are more the Microsoft Windows of the smartphone world. They have a vast array of applications and the user experience is pretty consistent across their devices which makes it easy for users.
Their flagship, "best we can do" phone is really just a copy of the Essential Phone's design and having a fingerprint reader, even if it was relegated to the back of the phone, would have made sense particularly for things like Apple Pay and for unlocking the phone when you don't want to bring it up in front of your face.
My next phone will most likely still be an iPhone simply because I see no reason to switch, everything works fine and there's no really compelling alternative - maybe I just have a relatively boring or average set of requirements for a smartphone - but if they are continuing this path it will be the iPhone 9 (or whatever they call the successor to the iPhone 8) rather than whatever the next iPhone X.
There's already too much Chrome-specific stuff and the point is: it's growing. Sure, Chrome started as fully standards-based, but then so did IE in the (very) early days when it was the best browser around. Then the years went on, and the IE-specific stuff grew until we had the world of IE6.
This is the problem with the standards organizations, companies do not want to be limited by the standard but adding to it takes far too long. So instead you get companies like Google and Microsoft creating technologies that enable new features beyond the standard. Microsoft had this with IE and ActiveX, now web standards have caught up in terms of functionality such that Microsoft has been able to abandon proprietary extensions because the standard is capable. This will likely come around for Google as well but it will take time.
You lose a significant amount of performance by connecting the GPU that way just on the bus speed alone, that's even before you get to the inability to leverage Infinity Fabric or NVLINK which are incredibly important to multi-GPU computing. The performance gap of communicating between GPUs over Thunderbolt VS PCIe is bad enough but compared to Infinity Fabric or NVLINK it's a complete non-starter.
If you've actually used the solution you suggest then you already know the performance is absolutely woeful.
I did a very quick comparable on the HP z840 to the base iMac pro and it was about $4665...it isn't exact, but is close to what Apple purports the base iMac pro to be. And also, the iMac pro includes the 5K 27" monitor which you would have to add to the HP or other comparable unit.
But you also get expandability, and of particular interest in your specific case Resolve leverages multiple GPUs so if things start to slow down you can simply add another GPU to the HP system and you're good to go, to improve the performance on the iMac you have to scrap the entire system, including the display, and buy a new one.
That's not going to be the case for everybody but certainly in your case the choice of something expandable is definitely the right one. If you really just want an Apple then holding out for the improved Mac Pro would be the right choice over the iMac.
If you have something very narrow and specific, you might be able to run .NET on Linux or OS X. If you expect to do a full-fledged client app, good luck with that.
Why? We're talking about the open source initiatives around .Net, if you use those open source bits there's no reason you can't run them on various different platforms and you can couple it with a platform-agnostic GUI toolkit (since the winforms bit of .Net is not open source) if you need a GUI. If you're going to use closed-source bits that are only available on one platform then yes you're not going to be able to run that on multiple platforms, but that is obvious already and isn't really related to this discussion. The open source bits are the things that are neutral and that is what solves the vendor (or rather platform) lock-in issue.
While many open source licenses provide this benefit (BSD, MPL), they also allow for the source code to be rolled into proprietary products (such as OsX).
I think perhaps it is you who misunderstands Free Software, just because a derivative work could be non-free does not make the original work non-free. Restrictive open source licenses are not a requirement for Free Software.
By guaranteeing freedom to downstream users, GPL maintains these four freedoms for all and forever.
Nowhere is that a requirement for Free Software, indeed the permissive open source licenses you mention are Free Software licenses precisely because they satisfy the 4 conditions you list. TensorFlow, for example, is Free Software.
Yes there is user/community benefit to having open source even on a proprietary platform - another example is Microsoft's various open source initiatives around .NET. But, what it does not solve is the matter of lock-in.
Of course it does, you don't have to run Google's open source TensorFlow on Google's cloud infrastructure, you can run it on Azure, EC2, your own cluster or even just your own local system. Likewise the open source .Net technologies don't require Microsoft's Windows platform, you can run them on Linux if you want to and the code itself is open and available. How exactly is it you think you are locked to a vendor?
I use a Macbook because MacOS is Unix that "just works".
As much as the various hardware limitations, software limitations (max of OpenGL 4.1 in 2017? Come on) and closed, proprietary technology over open standards (Metal over Vulkan, Airplay, Airdrop, Facetime, etc) is annoying I do like that the kernel can be updated without it breaking the display driver and having to go into a terminal to recompile the kernel module just to get the GUI working again. It's all those niche little annoyances that still persist all across Linux that add up to it being a poor user experience. Macs aren't perfect but on a day-to-day basis I often prefer to deal with their flaws than Linux's.
If that's the case then fine, I don't really care whether it's DisplayPort or HDMI and I'm not suggesting one is better than the other just that advancing standards like this is important and has valid use cases. At the moment DisplayPort does not have a standard that supports 8k @ 120Hz, it's simply on the roadmap for a future version (which may or may not be the next version) but if they manage to specify it, add it to the standard and ship it faster than HDMI then great, if not then I don't really care either.
HDMI 2.1 can do 8k @ 120Hz with 3:1 DSC, DisplayPort 1.4 can only do 8k @ 60Hz with the same compression ratio. The next version of DisplayPort should match HDMI 2.1 though. 120Hz is VERY important in VR.
In any case I said 'standards like this', whether that is DisplayPort or HDMI doesn't really matter.
I think that _you_ missed the point. When the existing standards have bandwidth requirements that are beyond the ability of content providers to distribute and there is virtually no planned upgrade path, further upgrades to that standard are of little/no use.
The ability to have VR headsets with a pair of 4k screens that you can drive with stereoscopic visuals is certainly a reason to have standards like this.
Well with replaceable/expandable components there's no reason they shouldn't last a very long time indeed, add that there's huge amounts of existing stock. If you like it, invest in a couple of those now for the event of catastrophic failure and you should be good beyond the next decade or so.
The simple answer is that people don't need to upgrade their phones all the time, they're good enough now. As you say, the S5 already has all the features you want.
The company sort of had conflicting goals of its own... they'd rather have sold them a groupware solution to replace Exchange, and ultimately the Linux migration projects were seen as a way to get more customers to use their own back-end solutions and server hardware. Having a perfect integration into exchange wasn't seen as a desirable outcome.
Right so when the consulting company isn't willing to present the ideal solution you can hardly blame the company for not coming up with it themselves.
Current versions of windows still require a hotfix if you have users with more than 120 AD groups.
Yes, as I said neither is perfect but you're just trading one set of problems for a different set of problems which may or may not be worse depending on many many variables and going through a painful migration process to do so. You can trade problems with the various options all day, that doesn't get you anywhere though unless you can somehow enumerate all the problems of all the options and weigh them all up for a particular use case.
I may be taking my use case (and the use case of my current market segment) for a generality, but I stopped caring about what OS/software I am using as long as I have uniform ability to access my data. If I have an external requirement for an application-specific format at some point (external entity insisting on PDF or word), the data gets converted to that format when needed but is stored as application-agnostic as possible.
Which is fine for simple things like documents, but it is not suitable for the various types of data involved in the media & entertainment, construction and manufacturing industries for example. For office workers, by and large that strategy might work. But that isn't an argument for Linux, it's an argument for 'don't care' which ultimately means the incumbent because the incumbent covers not just the general case but the niche cases as well.
The problem is that, when a great F/OS application is written for Linux, it's usually profitable to port it to Windows.
Yes so the operating system itself needs to offer something, some innovation above and beyond what Windows can offer. If there's no innovation then of course nobody is going to switch. The whole notion of free software is that it's ok to give away your 'secret sauce' because you capitalized on getting to market first and people are using your product.
None of these are issues in a corporate environment because the user won't install the OS himself and won't have administrative rights on his machine. The machine will be installed with the corporate image that already has all the required components installed and configured.
Right but these are the sorts of issues that crop up all the time, I'm not saying this is going to apply to everybody but niche cases like this (and there are hundreds of them) are one of the things that keeps people away from Linux and increases the support cost of Linux.
The biggest issue I was facing at the time was that the Exchange connector of Evolution didn't cope well with large scale Exchange deployments where you had more than one exchange server... so, for example, the calendar free/busy was only working for users that were on the same exchange server as you... this would have been a trivial fix just requiring an extra lookup per user in the calendar, but nobody was interested at the time.
Being in pre-sales did the company you were working for not see value in investing in such a thing? Particularly if it really was so trivial?
99% of the issues with Open Office had been caused by improper Office template design and were very quickly fixed. Some other interesting issues were caused by the way Microsoft overloaded their Kerberos tickets by enumerating the group membership of the user in a comment field of the ticket... users that are members of more than 120 AD groups generate tickets above 64K, which is creating funny issues in all kind of places (web services, IPSEC, login, ...).
But this, and your Exchange issue above and the issues I had and all the other niche cases and "easy fixes" are the things that keep people and companies away. Why use a tool that doesn't work properly and requires fixing when you can just pay for a tool that does work properly?
And I'm not saying Windows is anywhere near perfect but you just end up trading one set of problems for a different set of problems and arguing that it will be cheaper in the long run and maybe that's true but when you go to companies and tell them they may need to employ or contract software development teams to fix, improve or develop the applications they use that raises alarm bells.
What has been demonstrated with smartphones/tablets is that software compatibility and a stable UI are actually a red herring.
In terms of software compatibility that's simply not true, the primary reason people use Windows over Linux on the desktop is application compatibility, it's the same reason people use Android over Windows Phone on mobile. In terms of a stable UI I agree, in fact software compatibility is so important that Microsoft was able to completely flip over the apple cart in terms of UI and people still used it because their applications ran on it.
Yeah because switching back to Windows isn't due to any bias /s
Im sure it is, you have 2 biased camps fighting it out. Let's not pretend this is a case of one unbiased, objective party vs one biased party.
> As for the other 300 programs, what are they? Do they all have Linux alternatives?
Yes. :)
What are they? They've said there are hundreds of programs that they use that do not run on Linux and you, seemingly without knowing anything about what those programs are, insist that there are alternatives. Where did you get the list of those programs from to be able to be so confident of that?
I happened to have an iMac and wanted a Linux machine (dual boot) so I tried to install it. The whole point is the clumsiness of the process of running Linux on it compared to macOS and Windows.
A) 3rd party development support
Wait, what? Are fucking nuts? Most of the users are software developers.
So why is application support worse on Linux than it is on macOS and Windows? There's very little software that is Linux exclusive yet there is lots and lots of desktop software that runs on macOS and/or Windows but not on Linux, a lot of it is the sort of software that is important across big industries like entertainment, construction and manufacturing. There's no technical reason you couldn't develop those applications for Linux but most users are not software developers and it's usually several orders of magnitude cheaper to buy a license for a program that already exists that does what you want than to contract a team of developers to build an alternative just because you want to run Linux as your operating system.
"According to the Document Foundation (parent organization of LibreOffice), this step back to Microsoft will cost around 90M of taxpayer’s money."
Oh that couldn't possibly be biased. While they may have saved the license fees the workarounds, the non-working applications and the need to invest huge sums of money developing alternative applications is likely what is driving the switch back.
In theory many councils of many countries could switch to open source and collaborate on building applications together, everybody would share the cost and responsibility for development and maintenance and it would be all warm and fuzzy. It's a lovely sentiment, but reality is simply not like that.
They're trying to keep using Microsoft Exchange and 300 other Windows programs, on Linux.
You can use Microsoft Exchange on Linux, there's a bunch of different clients that work with it. As for the other 300 programs, what are they? Do they all have Linux alternatives? If not what you need to do is to develop those programs, it seems they've decided that is cost-prohibitive, which is probably true.
You have a corporation who's business model builds upon not playing nice with the competition.
You mean like how they have their applications on Windows, macOS, iOS, Android and web-based platforms that make products like even their flagship Office 365 available on desktop Linux? There's mail clients like DavMail and Hiri that work on Linux and operate with Microsoft's Exchange email server. Even .Net Core is open source and supported on Linux allowing even more cross platform applications.
Linux has a "the chicken or the egg" problem. Without a sizable desktop user pool in whatever industry you're trying to switch over, all these proprietary products won't be written for Linux. Without proprietary products, industry is handicapping themselves by moving to Linux. It's a tough problem to solve, and that's why it hasn't really been solved yet.
What you need is a reason for people to use Linux. The argument that it's cheap is a non-starter because you need to contribute resources to maintaining it and to developing the applications that you need that don't yet exist for it. The motivation for doing that would be some killer, disruptive feature like the way iOS and Android killed off Blackberry and Windows Mobile =6. Then of course Microsoft suffered the same problem that desktop Linux is having with Windows Phone, not actually a bad product but not innovative or compelling enough to attract users from the incumbents.