The Windows 10 Anniversary update, due later this summer, represents a major landmark for Microsoft. As well as being a significant update for Windows 10 on the desktop and Windows 10 Mobile on phones, the release is also coming to the Xbox One. For the first time, the Xbox One will be running essentially the same operating system as desktop Windows. Critically, it will also be able to run many of the same applications as desktop Windows.
In a lot of ways, this represents the realization of a vision that Microsoft has been promoting for more than 20 years: Windows Everywhere. Always important to Microsoft's ambitions for Windows as a platform, the Windows Everywhere ideal has a renewed significance with Windows 10 and CEO Satya Nadella's promise that Windows 10 will have one billion users within the first three years of its availability. The purpose of that promise is to send a message to developers that Windows is a big platform, a platform that they should still think about and create software for.
yeah I have him (assuming we're thinking of the same long-winded poster) on ignore too. But it does indicate that posts are hidden and you can temporarily unhide them from just the thread.
XBox One is a PC!
Microsoft is now pushing only x86 Atom-based phones for a reason!
Total marketing, and you guys have been totally been duped. Any core Windows architect (no longer with Microsoft) will tell you so! Several of them are good, long-standing colleagues of mine.
Windows is 32-bit x86, byte-aligned. Anything non-x86 has to be byte-code emulated-translated, with a huge run-time blob and performance hit. That's why performance is crap on non-x86, and Microsoft has dropped non-x86 from their "Everywhere" program. This was announced just several months back, quietly, silently. The codebase is still not even remotely 64-bit clean ... something Linux has been since the '90s on Alpha (which made it almost day 1 for x86-64)!
The "way forward" for Microsoft is new codebase, driven by their separate Azure teams. I know, I worked with some of the same people there, and was even recruited by Microsoft (like others from Red Hat)! They are re-writing entire programs for Windows Server 10 Nano, and which is also portable to GNU/Linux, which will be compatible with their "after-Windows" API. These platforms, especially Nano itself, break 100% of existing Windows applications.
Seriously ... you guys have ate up the marketing. Microsoft completely reneg'd on their promises for non-x86. I love it. Many of us are proven completely right (about this being x86-only), and you guys say otherwise. Why? Marketing!
This is what drives me up-the-wall. The Azure team in Microsoft makes fun of these people!
XBox One is a PC!
Microsoft is now pushing only x86 Atom-based phones for a reason!
Total marketing, and you guys have been totally been duped. Any core Windows architect (no longer with Microsoft) will tell you so! Several of them are good, long-standing colleagues of mine.
Windows is 32-bit x86, byte-aligned. Anything non-x86 has to be byte-code emulated-translated, with a huge run-time blob and performance hit. That's why performance is crap on non-x86, and Microsoft has dropped non-x86 from their "Everywhere" program. This was announced just several months back, quietly, silently. The codebase is still not even remotely 64-bit clean ... something Linux has been since the '90s on Alpha (which made it almost day 1 for x86-64)!
The "way forward" for Microsoft is new codebase, driven by their separate Azure teams. I know, I worked with some of the same people there, and was even recruited by Microsoft (like others from Red Hat)! They are re-writing entire programs for Windows Server 10 Nano, and which is also portable to GNU/Linux, which will be compatible with their "after-Windows" API. These platforms, especially Nano itself, break 100% of existing Windows applications.
Seriously ... you guys have ate up the marketing. Microsoft completely reneg'd on their promises for non-x86. I love it. Many of us are proven completely right (about this being x86-only), and you guys say otherwise. Why? Marketing!
This is what drives me up-the-wall. The Azure team in Microsoft makes fun of these people!
There is not a single x86 Windows Mobile (nee Phone) on the market (yet). Rumors of a future intel-powered Surface Phone notwithstanding.
Raspberry Pi is ARM and runs Windows 10 just great.
I love how you expect me to throw out my computer science knowlege and everything I've seen, done, and read in favor of some inside secrets a Linux fanboi supposedly has from supposed former architects at Microsoft. You really failed English Comp 2.
Sounds to me like you've bought the Linux marketing hook, line, and sinker.
Have you tried out or read up on the Bash on Windows 10 stuff? It's pretty damn cool and comes in handy when one runs across open source stuff that doesn't work on Windows.
Huh? You're kidding, right? There are several out right now.
They killed the legacy RT with x86, ARM and others, and now they have just Windows 10. But what you're failing to realize -- again, marketing -- is that they've basically "re-invented" the "mobile" for PC-architectures only, plus they are trying to get some over to ARMv8, usually with emulation! That's the joke this is! It's Windows ... for PC architectures, just mobile, portable, set-top, etc... ones. Please explain how this isn't marketing? They are basically telling everyone to get to x86 if possible!
But ... here's the problem (e.g., Asus Chairman)! Developers do not trust Microsoft! They do not like Windows on even x86 phones!!! They have been f'd over on CE, Embedded NT, RT and now back to a x86-centric NT strategy!
The XBox One is a PC!
Literally ... you eat up the marketing, and fail to understand what they industry is saying! They are tired of Microsoft doing this!
Define "great"? Seriously dude. Do you understand how much existing code will not run on it? Are you literally making my point for me by using the Pi as an example? Seriously?
Microsoft's entire Windows -- and I stress Windows -- future is based on 1 preferred platform ... x86-64, porting as much as they can from their 32-bit only codebase, dropping what they can, emulating 32-bit structures as necessary -- and, a "second class citizen" in ARMv8 (64-bit) with uEFI, using x86 byte code emulation as needed. They are completely shifting to x86 by default, and have told all their partners to do such.
But because you only listen to marketing, they literally have you fooled. Everything I predicted about RT and ARM came true! But in true Microsoft fashion, everytime they abandon a strategy, leaving developers stranded, they push another strategy.
I love how you expect me to throw out my computer science knowlege and everything I've seen, done, and read in favor of some inside secrets a Linux fanboi supposedly has from supposed former architects at Microsoft. You really failed English Comp 2.
Dude, where do you get Linux "fanboi"? That's like saying a Microsoft employee is a Windows "fanboi"! WTF?!
I spend nearly a decade working for Red Hat, as a direct employee for 7 of them, several years (and still on and off) as a contractor (that presents myself as Red Hat). I literally left them because I got tired of 100+ hour/weeks, first when my father was ill and I couldn't get away (2012), and then again when I was totally over-utilized (2014), and will only work hourly for them now. You do know Microsoft has recruited me several times, correct? In the mid '00s, late '00s and again just 2 years ago (with the big Azure ramp-up, and being a leading SDN expert in Red Hat). I know some of the guys with the Azure team (I was almost one of them!), along with past, core NT/Win32 architects now at Google.
Do you know how much open source I've done on Windows? Mono projects? Everyone talks about Mono being a portability solution to non-Windows, but that's a joke. If you don't write for Mono, and stick with Visual Studio, you get 32-bit Windows-only code. You are absolutely listening to marketing, not reality. I've ported to/from both.
Here's an ultra simple test, and no Googling ...
Q: What are the primary differences between Cygwin and MinGW? Why would a product developer choose one or the other?
You realized I was one of Red Hat's leading Senior Consultants at major accounts, even done some custom embedded and a crapload of Windows integration.
I have cost Microsoft over 9 figures in revenue -- mid 8 figures at just one account in 3 months. Not being a Linux "fanboi," but literally eliminating the waste that the generic PC platform with Windows and "fat" solutions is in so many roles it should not be. They literally throw money at the problem when their customers refuse to do so any more. At that same customer, I was enemy #1 for 400 Windows developers for a reason.
At one point Microsoft and Red Hat were even talking about the political aspects of me becoming a MCM on Directory Server, because they are usually all Microsoft employees or Gold Partners. I literally got pulled into a lot of AD at many accounts because I do a lot of Kerberos and LDAP internals, migrations, etc... (to the point Red Hat often had to pull the SoW out). I mean, really ... I was one of the half-dozen guys in the company focused on large-scale Linux-Windows interoperability as a SME for major accounts.
Microsoft, of course, shut down the Masters program abruptly (less than 30 days notice), pissing a lot of us MCPs off, especially those who had sunk tens of thousands of dollars into it.
Have you tried out or read up on the Bash on Windows 10 stuff? It's pretty damn cool and comes in handy when one runs across open source stuff that doesn't work on Windows.
No crap. Because I do a lot of open source on Windows ... and more than just rebuilding Apache on Windows (although that alone is a sustainable nightmare with the countless ISVs who built abandoned Apache Win32 solutions). Geez, you're talking down to me like I haven't done this, when I've done a crapload more open source on Windows than you because open source was (and still is) my life!
Literally dude. It's called subject matter expertise, and you keep crossing into mine! Open source, Mono, portability. Microsoft ... is a joke when it comes to this. Their own Azure team understands this. You are totally a marketing spewing fool. And I'm the fanboi? Because I understand the non-sense?
I literally have been in the code, over and over! You're talking marketing and conference non-sense. How many times has Microsoft reneg'd on this? They shift the strategy! Even just from that last developer conference you even quoted! This is an x86 strategy! ARM will never materialize as a first-class citizen in comparison, and is only there are a carrot that will never happen en masse. Or at least it will take a lot of emulation and a lot of new code.
The Azure team is the only one working on actual portability, not the Windows team! And they are focused on cross-GNU/Linux and Windows codebases, unlike the Windows team who is still focused on x86 only or x86 first (when things actually port). That's the team that may have some things happen, but it's going to be post Windows 10.
Honestly, enjoy the marketing.
It's the same thing in securities and trading. People assume Windows dominates the backend, because they see .NET on Windows at Bloomberg and Reuters. The backend is where it is at, and where the money is. But perception = reality, marketing is king. Doesn't matter that Microsoft and Accenture got their @$$es handed to them for the debacle of the LSE, Accenture being heavily subsidized by Microsoft, their specified partner, not LSE's choice.
People who say such have never written portable code. All they know is Visual Studio and 100% assume it ports.
No offense guys, but I'm actually one of the most uniquely qualified and exposed professionals in this space. I done both sides, inside of major accounts integrating both, including user interfaces and end-customer solutions. Most people haven't even scratched anything close to what I've done in this space.
Literally, you guys read marketing ... I live it, especially after one of you fools in a key stakeholder position believe it and pay a lot of money, then find out the truth. Only Oracle is worse.
That's one of the reasons Microsoft tried to recruit me, among others I worked with, from Red Hat. Most of us were intrigued, but didn't like the IP agreement. Seriously, GFY. You'all are naive fools, but most of the mobile market isn't tolerating it any more. Even the Azure team inside of Microsoft is tired of it too. Why do you think Ballmer hates them?
I see the lunatic keeps on responding. No matter what he writes, team woot. Woot could be 110% wrong and I would still think that he won the argument. I cannot support mental midgets and people who should be institutionalized/euthanized.
I see the lunatic keeps on responding. No matter what he writes, team woot. Woot could be 110% wrong and I would still think that he won the argument. I cannot support mental midgets and people who should be institutionalized/euthanized.
Cliff Version: Microsoft's new strategy is exactly what I predicted, x86 first (including all their own devices -- my main prediction) with most things x86 only (so much doesn't run non-x86). They are in real trouble as few want to get on board with yet their latest, mobile strategy (among all the changes). Windows mobile is essentially limited to x86 tablets and set-tops at this point (set-tops may be their future mainstay).
Huh? You're kidding, right? There are several out right now.
They killed the legacy RT with x86, ARM and others, and now they have just Windows 10. But what you're failing to realize -- again, marketing -- is that they've basically "re-invented" the "mobile" for PC-architectures only, plus they are trying to get some over to ARMv8, usually with emulation! That's the joke this is! It's Windows ... for PC architectures, just mobile, portable, set-top, etc... ones. Please explain how this isn't marketing? They are basically telling everyone to get to x86 if possible!
But ... here's the problem (e.g., Asus Chairman)! Developers do not trust Microsoft! They do not like Windows on even x86 phones!!! They have been f'd over on CE, Embedded NT, RT and now back to a x86-centric NT strategy!
The XBox One is a PC!
Literally ... you eat up the marketing, and fail to understand what they industry is saying! They are tired of Microsoft doing this!
Define "great"? Seriously dude. Do you understand how much existing code will not run on it? Are you literally making my point for me by using the Pi as an example? Seriously?
Microsoft's entire Windows -- and I stress Windows -- future is based on 1 preferred platform ... x86-64, porting as much as they can from their 32-bit only codebase, dropping what they can, emulating 32-bit structures as necessary -- and, a "second class citizen" in ARMv8 (64-bit) with uEFI, using x86 byte code emulation as needed. They are completely shifting to x86 by default, and have told all their partners to do such.
But because you only listen to marketing, they literally have you fooled. Everything I predicted about RT and ARM came true! But in true Microsoft fashion, everytime they abandon a strategy, leaving developers stranded, they push another strategy.
This is an x86 strategy. And you ate it up!
Dude, where do you get Linux "fanboi"? That's like saying a Microsoft employee is a Windows "fanboi"! WTF?!
I spend nearly a decade working for Red Hat, as a direct employee for 7 of them, several years (and still on and off) as a contractor (that presents myself as Red Hat). I literally left them because I got tired of 100+ hour/weeks, first when my father was ill and I couldn't get away (2012), and then again when I was totally over-utilized (2014), and will only work hourly for them now. You do know Microsoft has recruited me several times, correct? In the mid '00s, late '00s and again just 2 years ago (with the big Azure ramp-up, and being a leading SDN expert in Red Hat). I know some of the guys with the Azure team (I was almost one of them!), along with past, core NT/Win32 architects now at Google.
Do you know how much open source I've done on Windows? Mono projects? Everyone talks about Mono being a portability solution to non-Windows, but that's a joke. If you don't write for Mono, and stick with Visual Studio, you get 32-bit Windows-only code. You are absolutely listening to marketing, not reality. I've ported to/from both.
Here's an ultra simple test, and no Googling ...
Q: What are the primary differences between Cygwin and MinGW? Why would a product developer choose one or the other?
You realized I was one of Red Hat's leading Senior Consultants at major accounts, even done some custom embedded and a crapload of Windows integration.
I have cost Microsoft over 9 figures in revenue -- mid 8 figures at just one account in 3 months. Not being a Linux "fanboi," but literally eliminating the waste that the generic PC platform with Windows and "fat" solutions is in so many roles it should not be. They literally throw money at the problem when their customers refuse to do so any more. At that same customer, I was enemy #1 for 400 Windows developers for a reason.
At one point Microsoft and Red Hat were even talking about the political aspects of me becoming a MCM on Directory Server, because they are usually all Microsoft employees or Gold Partners. I literally got pulled into a lot of AD at many accounts because I do a lot of Kerberos and LDAP internals, migrations, etc... (to the point Red Hat often had to pull the SoW out). I mean, really ... I was one of the half-dozen guys in the company focused on large-scale Linux-Windows interoperability as a SME for major accounts.
Microsoft, of course, shut down the Masters program abruptly (less than 30 days notice), pissing a lot of us MCPs off, especially those who had sunk tens of thousands of dollars into it.
No crap. Because I do a lot of open source on Windows ... and more than just rebuilding Apache on Windows (although that alone is a sustainable nightmare with the countless ISVs who built abandoned Apache Win32 solutions). Geez, you're talking down to me like I haven't done this, when I've done a crapload more open source on Windows than you because open source was (and still is) my life!
Literally dude. It's called subject matter expertise, and you keep crossing into mine! Open source, Mono, portability. Microsoft ... is a joke when it comes to this. Their own Azure team understands this. You are totally a marketing spewing fool. And I'm the fanboi? Because I understand the non-sense?
I literally have been in the code, over and over! You're talking marketing and conference non-sense. How many times has Microsoft reneg'd on this? They shift the strategy! Even just from that last developer conference you even quoted! This is an x86 strategy! ARM will never materialize as a first-class citizen in comparison, and is only there are a carrot that will never happen en masse. Or at least it will take a lot of emulation and a lot of new code.
The Azure team is the only one working on actual portability, not the Windows team! And they are focused on cross-GNU/Linux and Windows codebases, unlike the Windows team who is still focused on x86 only or x86 first (when things actually port). That's the team that may have some things happen, but it's going to be post Windows 10.
Honestly, enjoy the marketing.
It's the same thing in securities and trading. People assume Windows dominates the backend, because they see .NET on Windows at Bloomberg and Reuters. The backend is where it is at, and where the money is. But perception = reality, marketing is king. Doesn't matter that Microsoft and Accenture got their @$$es handed to them for the debacle of the LSE, Accenture being heavily subsidized by Microsoft, their specified partner, not LSE's choice.
People who say such have never written portable code. All they know is Visual Studio and 100% assume it ports.
Please link to all of these Windows 10 Mobile/Phone devices on x86 SoCs. I'm unaware of any and google seems to be unaware of any as well. Microsoft's own Lumia devices released late last year are on Qualcomm ARM chips.
I have no clue as to the differences between mingw and Cygwin and don't care. I'm legit surprised you're using Windows 10 Insider builds so props for that.
Let's just ignore your constant diatribes against windows. The fact that open source is your self-professed life shows where your loyalties lie.
There's definitely some marketing to the whole "one windows" thing. But it is rooted in computer science. What does it mean to be an operating system? Is each Linux distro a different operating system? I'd argue that Windows 10 Mobile is much more Windows than Android is Linux.
I don't know why you're so hung up on the whole "WINDOWS CANT RUN ON ARM" thing when it's demonstrably not true. If Microsoft really was emulating bytecode wouldn't the performance and battery life be absolutely dreadful on Windows Phones on ARM SoCs? I mean if Microsoft can implement high perf emulation like that that deserves even more props than compiling an os for multiple architectures.
Why do you keep bringing up your own experience working for a Linux distributor and integrator as some sort of expert testimony that we should automatically take as truth? Are you saying that Ars Technica is lying and shilling for Microsoft?
My work life is around identity and federation. We integrate various platforms but are primarily a Microsoft shop. That said, I'm primarily working in Java these days as the products we're using are written in Java so the integration must be as well. No big deal but I do miss Visual Studio. So much better than NetBeans and the other IDEs I've tried.
Please link to all of these Windows 10 Mobile/Phone devices on x86 SoCs. I'm unaware of any and google seems to be unaware of any as well. Microsoft's own Lumia devices released late last year are on Qualcomm ARM chips.
I have no clue as to the differences between mingw and Cygwin and don't care. I'm legit surprised you're using Windows 10 Insider builds so props for that.
Let's just ignore your constant diatribes against windows. The fact that open source is your self-professed life shows where your loyalties lie.
There's definitely some marketing to the whole "one windows" thing. But it is rooted in computer science. What does it mean to be an operating system? Is each Linux distro a different operating system? I'd argue that Windows 10 Mobile is much more Windows than Android is Linux.
I don't know why you're so hung up on the whole "WINDOWS CANT RUN ON ARM" thing when it's demonstrably not true. If Microsoft really was emulating bytecode wouldn't the performance and battery life be absolutely dreadful on Windows Phones on ARM SoCs? I mean if Microsoft can implement high perf emulation like that that deserves even more props than compiling an os for multiple architectures.
Why do you keep bringing up your own experience working for a Linux distributor and integrator as some sort of expert testimony that we should automatically take as truth? Are you saying that Ars Technica is lying and shilling for Microsoft?
My work life is around identity and federation. We integrate various platforms but are primarily a Microsoft shop. That said, I'm primarily working in Java these days as the products we're using are written in Java so the integration must be as well. No big deal but I do miss Visual Studio. So much better than NetBeans and the other IDEs I've tried.
Please link to all of these Windows 10 Mobile/Phone devices on x86 SoCs. I'm unaware of any and google seems to be unaware of any as well. Microsoft's own Lumia devices released late last year are on Qualcomm ARM chips.
So, this is what I love, the nice circular. So, let's summarize ...
The only Windows 10 x86 phones are R&D samples, with not much interest, and ...
Have you seen the software limitations of Windows 10 on non-x86?
In other words ...
as Microsoft tries to push everyone towards x86, and drops any ARM other than v8 w/uEFI (support for which will still be questionable), many vendors are getting out. Even Microsoft itself. The existing, pre-uEFI, pre-v8 ARM platforms are only getting sustainment. Many predicted this, including myself, and that's what you're in total denial over. This entire, new strategy is basically what we all said ... x86 first, possibly only, including for new Microsoft products.
Side Note: I've long conceded that Google sucks as sustaining Android in comparison, but that's another story. Trust me, I'm very objective on this, and have my complains about non-GNU, non-Java Android.
Exactly! You have never ported code outside of Windows, likely not even x86 (am I right? )
Because the second you do, you quickly learn you don't write on a Windows platform -- much less in Visual Studio -- first. And that's the thing ... I have extensive experience with porting from/to multiple platforms. I can instantly tell when people think what they do about Windows.
I was an original NT 3.1 Beta tester at the largest installed site of the first native NT app. I've build run-times for every CE and NT release, including having to deal with some of the early StrongARM builds with systems that utterly lacked anything that looked remotely like a PCI hub. I've had Alpha access on several releases. Heck, for years I was on Microsoft Gold Partner lists thanx to colleagues until a few years as an employee for Red Hat.
I'm not some ignorant puke, I've been into the internals of CE and NT for decades.
You mean like the Microsoft Azure team? They are huge on open source, even GPL in several cases!
Or the fact that 85% of Red Hat's Professional Services are not Linuxexerts (HINT: JBoss and application-level).
Seriously ... right there you told me everything!
Knowledge is not bias. Lack of doing something is bias, ergo yourself on porting.
I mean, do you know how much open source Microsoft itself uses in core NT libraries?
E.g., when the zlib security hole hit, there were 2 patches required for Linux ... one in the kernel, one in the shared object (.so). Windows? 17 core files related to the NT executive because they statically linked. And that's just one example. libssl is the second most statically linked library, and even Microsoft branded software (admittedly a lot of acquisitions) have used it instead of the NT security libraries.
Open source is not bigotry ... it's total reality of where most innovation has always come from. That's why Microsoft uses it ... always has! They're stupid not to, although they statically link way, way too much, as do their ISVs, leaving old versions around in various directories for various services.
Microsoft has even crossed from BSD/MIT to GPL licensed in the case of Azure. Geez, even core Microsoft people understand this! Even Gates since '09. Where have you been? You do realize that Microsoft, as of '09, stopped using Cisco and outsourcing to others and just started directly having their own IT and developers admit how much open source they use directly.
You are literally Ballmer. In denial ... and clueless. I cannot educate you because of your assumptions, including about me, and anyone doing open source.
I cannot stress that enough ... there is Ballmer, and the new, Azure-focused leadership, and they are utterly at odds. Ballmer thinks PCs are the future, and x86 the staple. The Azure team realizes that the world has changed, and Microsoft is the #5 player in the mobile space. Microsoft will become a PC vendor to sustain their revenue, all while they try to push people to Office 365.
Heck, even Red Hat evaluated Office 365 and RT devices as a potential solution for exchange with customers who only use MS Office, and found the on-line solutions compatibility lacking because the Windows PC versions still don't use the ISO Office OpenXML 2008 spec. They are on a "transitional" format that is incompatible in some many implementations.
Stop saying "computer science" like I'm some ignorant f'.
I was deploying NT/Alpha -- yes Alpha, not x86 -- back in '96, using Digital FX!32 technologies which Microsoft now uses for its base binary translation in .NET. I'm utterly deep-dive technical on the internals here.
What does it mean to be an operating system? Is each Linux distro a different operating system? I'd argue that Windows 10 Mobile is much more Windows than Android is Linux.
Of course Android is not GNU/Linux or pure Java. I'm the first one to admit that!
Dude ... how many times do I have to say it, I do not like Google and Android. It's poorly sustained with security issues. Heck, Google pulled over Red Hat CTO Brian Stevens to get their containers under control! Because they have sustainment issues on their servers too! Not surprisingly, Google is increasingly tapping Red Hat for sustainment engineering now.
Google is not a good example of Linux! Not remotely!
Side Note (since you don't believe anything I say): I actually interviewed with Brian in April '01 before either of us were with Red Hat, a company later acquired by Red Hat. Unfortunately they went on a hiring freeze, like a lot of company in March-April '01. But I maintained some of the official, Upstream documentation (network, NFS), as well as had a book published, on the subject matter involved (CIFS/SMB, NFS, etc...). The later also featured the founder of Gentoo, who now works for Microsoft (Linux maintainers are not a bunch of Linux fanboys and bigots!). Not surprisingly, I would later work on, and even patch, some of the same software (Cluster Suite -- prime reason I was at the Census, the largest base with peta-bytes, and other places I cannot talk about) with Red Hat, and other, related acquisitions (Gluster, CEPH) as I was the lead, post-sales Enablement Architect for all of Storage ('12-14) and usually had to backfill Consulting engagements due to lack of expertise (especially early adopter stuff).
Most of Microsoft Windows is byte-alignment ignorant (pretty much unique to x86), 32-bit Windows x86 (Win32/x86) byte code C libraries and does not port to non-x86 platforms that don't offer Protected386 run-time memory model compatibility (something x86-64/Long Mode does, although not Virtual86, which is another story). That is the bane of Windows' legacy!
It's why RT died. It's too bad we cannot get the old server archive back, because that's exactly what I predicted, and how everything would be x86-first or even only. The entire pre-ARMv8/pre-uEFI Windows Mobile world is now in pure sustainment mode.
It's why most things from .NET never ported to Mono, although the Microsoft Azure is officially re-writing a lot, and no longer having to "play games" (kudos to the Azure team). It's why MS Office has always had issues in its 64-bit implementation, and the Winforms version (Windows version, but not non-Windows version) of even other, cross-platform suites have had issues as well.
If Microsoft really was emulating bytecode wouldn't the performance and battery life be absolutely dreadful on Windows Phones on ARM SoCs? I mean if Microsoft can implement high perf emulation like that that deserves even more props than compiling an os for multiple architectures.
You're like Intel with Atom v. ARM benchmarks and battery ... you claim one thing, then the reviews tell it totally different! You compare to old systems.
It's like my favorite argument ... "NT is as good as trading as Linux." Yes ... Linux 2.2 from 15 years ago! Same thing on Atom v. ARM for performance/power, Windows v. Linux for embedded, etc... Compare against an old version.
Why do you keep bringing up your own experience working for a Linux distributor and integrator as some sort of expert testimony that we should automatically take as truth?
I could name people you could talk to, and you wouldn't even know who they are. But if you've been around since the '90s in the open source world, you'd know exactly who they are. Sigh ...
Furthermore, the only reason I cannot name some of the companies is because I'm under NDA. But have a beer with me and I will tell you everything! This whole "conspiracy theory" stuff is not "conspiracy theory." It's just business. It happens. I've been in the middle of it, first-hand, often the guy on-site. Again, it's just the business realities. The stuff you hear in rumors, I was often the guy, or involved with the team.
The only project my team was ever fully recognize for, publicly, was at the Census ('09-11) on NPR, but we were not named by name, only by company (even though there was less than a handful of us from Red Hat). And I'm still under NDA for that.
The most I'll say is that over just a couple of nights (yes, I had other duties with ECON and other entities -- e.g., the ones that do the quarterly economic reports to Congress), I analyzed the prime contractor's deployment, and helped identify all the inconsistencies. We then took that analysis and redeployed the entire compute infrastructure in 3-days ... something they couldn't do in 9 months.
You're literally like the new, "we standardizing on Windows" CIO for the Commerce Department who came over to his own, primary DataCenter and thought we just had a few non-Windows systems. We showed him the few racks of Windows Servers in one module, and then all of our stuff in that single module, plus a half-dozen more modules of 0 Windows. He literally didn't understand, because all he sees is interfaces he thinks is Windows (but often is not). I deal with people like yourself every day, humble, smiling, account management is sometimes difficult, but I (pun)manage(pun).
Heck, we (Red Hat) got permission to bring GM in to look at how we run Datacenters, as they were trying to standardize post-bankruptcy.
Side Note: I turned down the call for the Tech Surge. I had my fill of federal politicking after the 2010 Census.
Ars Technica has its authors, including from various corporations. I don't blame various Microsoft Product Developers from writing articles there.
I'm just saying, in the grand scheme of things, Microsoft has a huge, legacy codebase that doesn't port! You keep missing the fact that I'm talking about existing apps. New apps are different. But a lot of .NET is using existing C/C++ code that is Win32/x86-only.
The Azure team is doing all they can to rewrite as much as they can. Kudos to them, SQL Server and other things are coming to Linux ... because they need them on Linux!
My work life is around identity and federation. We integrate various platforms but are primarily a Microsoft shop. That said, I'm primarily working in Java these days as the products we're using are written in Java so the integration must be as well.
That's good. Because -- assuming you're also doing IdM and federation in Java -- you're likely learning what AD LDS really is (generic LDAP server -- most specifically the same Michigan v3 code that even AD is based on too), and how the non-AD world works, and has always worked.
Yes, because it's not a "canned" solution designed for a specific platform with specific assumptions ... ones that are often only good for 1 Windows release too.
That's the negative side of open systems -- beyond just open standards -- no argument. You're having to learn a lot. Open systems aren't easy. You really have to know a lot! I'm the first one to admit this. It takes decades to get all of the standards and realities down.
Just like -- switching back to directory services -- AD v. LDAP (including AD LDS). If everything is "cannned" down, it's easy. But if it's open and flexible, it requires you to understand far more.
E.g., Everyone thinks the DoD CAC system is AD and its certificate services, because that's all they see. It's not. It will never be. You cannot put 10M (and possibly going to tens of millions in just the next few years) on AD! But you can use just one (1) Red Hat server, (2) for failover (load balancing not so much needed.
But even Red Hat has IPA for a reason, the "canned" version for POSIX attributes (instead of legacy NTLM/SAM). At the same time, IPA isn't going to solve every UNIX LDAP implementation need, far from it. It just offers some base compatibility.
But having SSSD like NT's LSA (which just had a huge, massive, long-standing security hole be exposed), and IPA like AD, is a start. IPA emulating some components that AD expects -- like AD thinking the external forest trust to IPA has a Global Catalog -- goes a long way.
Again, I'm bringing this up because you mentioned IdM and Federation. And that's just platform-level, not application level. I mean, even Oracle dropped Sun One and is focusing on just Oracle LDAP for application-level ... conceding all platform-level to Red Hat.
Heck, Microsoft dropped IdM for UNIX in Windows 10 Server because they believe Red Hat's IPA is the right way forward! External forest trusts, instead of mixed in the same domains/forests ... because schema will always be different!
When a single post reaches four screens in length and has a rabid use of the spoiler button for no apparent reason, there is no reason to read it by anyone, ever. That might need to replace the definition of insanity. I'm actually dumber for having read the small portion of the post that I did.
there is no reason to read it by anyone, ever. That might need to replace the definition of insanity. I'm actually dumber for having read the small portion of the post that I did.
Hey man, this is my career, my expertise, my life in a way. Feel free to make fun of it. But I have my experiences. There's a reason I got out of this, and took a pay cut. It's just not worth the hours.
I tried to help HP '14-'15, but their sales had me working 100 hour/weeks because they scoped things entirely wrong, and it took 2-3 weeks to wrestle control of the account management from them before we lost the account altogether. That and they knocked off the VP by '15 March that actually knew what he was doing -- not surprisingly he came from an acquisition that was heavily ex-Red Hat folk (hence why he did, although most scattered).
I'm literally done doing this stuff, which was always post-sales (although I did a lot of upsell and expansion of existing accounts, once on-site). Not worth the hours, with people -- especially those who did the pre-sales, or resources "acquired" that don't know anything -- dragging me. I'll never forget the newer manager (Red Hat grew from less than 1,000 and barely NYSE listed to nearly 10,000 and S&P500 in my time) who didn't want me for something, and preferred something from EMC, because it was "storage."
That made a lot of us older Fedoras just go facepalm. Because ... Open source experts study internals and know technologies ... not products.
So, this is what I love, the nice circular. So, let's summarize ...
The only Windows 10 x86 phones are R&D samples, with not much interest, and ...
Have you seen the software limitations of Windows 10 on non-x86?
In other words ...
as Microsoft tries to push everyone towards x86, and drops any ARM other than v8 w/uEFI (support for which will still be questionable), many vendors are getting out. Even Microsoft itself. The existing, pre-uEFI, pre-v8 ARM platforms are only getting sustainment. Many predicted this, including myself, and that's what you're in total denial over. This entire, new strategy is basically what we all said ... x86 first, possibly only, including for new Microsoft products.
Side Note: I've long conceded that Google sucks as sustaining Android in comparison, but that's another story. Trust me, I'm very objective on this, and have my complains about non-GNU, non-Java Android.
Exactly! You have never ported code outside of Windows, likely not even x86 (am I right? )
Because the second you do, you quickly learn you don't write on a Windows platform -- much less in Visual Studio -- first. And that's the thing ... I have extensive experience with porting from/to multiple platforms. I can instantly tell when people think what they do about Windows.
I was an original NT 3.1 Beta tester at the largest installed site of the first native NT app. I've build run-times for every CE and NT release, including having to deal with some of the early StrongARM builds with systems that utterly lacked anything that looked remotely like a PCI hub. I've had Alpha access on several releases. Heck, for years I was on Microsoft Gold Partner lists thanx to colleagues until a few years as an employee for Red Hat.
I'm not some ignorant puke, I've been into the internals of CE and NT for decades.
You mean like the Microsoft Azure team? They are huge on open source, even GPL in several cases!
Or the fact that 85% of Red Hat's Professional Services are not Linuxexerts (HINT: JBoss and application-level).
Seriously ... right there you told me everything!
Knowledge is not bias. Lack of doing something is bias, ergo yourself on porting.
I mean, do you know how much open source Microsoft itself uses in core NT libraries?
E.g., when the zlib security hole hit, there were 2 patches required for Linux ... one in the kernel, one in the shared object (.so). Windows? 17 core files related to the NT executive because they statically linked. And that's just one example. libssl is the second most statically linked library, and even Microsoft branded software (admittedly a lot of acquisitions) have used it instead of the NT security libraries.
Open source is not bigotry ... it's total reality of where most innovation has always come from. That's why Microsoft uses it ... always has! They're stupid not to, although they statically link way, way too much, as do their ISVs, leaving old versions around in various directories for various services.
Microsoft has even crossed from BSD/MIT to GPL licensed in the case of Azure. Geez, even core Microsoft people understand this! Even Gates since '09. Where have you been? You do realize that Microsoft, as of '09, stopped using Cisco and outsourcing to others and just started directly having their own IT and developers admit how much open source they use directly.
You are literally Ballmer. In denial ... and clueless. I cannot educate you because of your assumptions, including about me, and anyone doing open source.
I cannot stress that enough ... there is Ballmer, and the new, Azure-focused leadership, and they are utterly at odds. Ballmer thinks PCs are the future, and x86 the staple. The Azure team realizes that the world has changed, and Microsoft is the #5 player in the mobile space. Microsoft will become a PC vendor to sustain their revenue, all while they try to push people to Office 365.
Heck, even Red Hat evaluated Office 365 and RT devices as a potential solution for exchange with customers who only use MS Office, and found the on-line solutions compatibility lacking because the Windows PC versions still don't use the ISO Office OpenXML 2008 spec. They are on a "transitional" format that is incompatible in some many implementations.
Stop saying "computer science" like I'm some ignorant f'.
I was deploying NT/Alpha -- yes Alpha, not x86 -- back in '96, using Digital FX!32 technologies which Microsoft now uses for its base binary translation in .NET. I'm utterly deep-dive technical on the internals here.
Of course Android is not GNU/Linux or pure Java. I'm the first one to admit that!
Dude ... how many times do I have to say it, I do not like Google and Android. It's poorly sustained with security issues. Heck, Google pulled over Red Hat CTO Brian Stevens to get their containers under control! Because they have sustainment issues on their servers too! Not surprisingly, Google is increasingly tapping Red Hat for sustainment engineering now.
Google is not a good example of Linux! Not remotely!
Side Note (since you don't believe anything I say): I actually interviewed with Brian in April '01 before either of us were with Red Hat, a company later acquired by Red Hat. Unfortunately they went on a hiring freeze, like a lot of company in March-April '01. But I maintained some of the official, Upstream documentation (network, NFS), as well as had a book published, on the subject matter involved (CIFS/SMB, NFS, etc...). The later also featured the founder of Gentoo, who now works for Microsoft (Linux maintainers are not a bunch of Linux fanboys and bigots!). Not surprisingly, I would later work on, and even patch, some of the same software (Cluster Suite -- prime reason I was at the Census, the largest base with peta-bytes, and other places I cannot talk about) with Red Hat, and other, related acquisitions (Gluster, CEPH) as I was the lead, post-sales Enablement Architect for all of Storage ('12-14) and usually had to backfill Consulting engagements due to lack of expertise (especially early adopter stuff).
Listen dude ... for the last time ...
Most of Microsoft Windows is byte-alignment ignorant (pretty much unique to x86), 32-bit Windows x86 (Win32/x86) byte code C libraries and does not port to non-x86 platforms that don't offer Protected386 run-time memory model compatibility (something x86-64/Long Mode does, although not Virtual86, which is another story). That is the bane of Windows' legacy!
It's why RT died. It's too bad we cannot get the old server archive back, because that's exactly what I predicted, and how everything would be x86-first or even only. The entire pre-ARMv8/pre-uEFI Windows Mobile world is now in pure sustainment mode.
It's why most things from .NET never ported to Mono, although the Microsoft Azure is officially re-writing a lot, and no longer having to "play games" (kudos to the Azure team). It's why MS Office has always had issues in its 64-bit implementation, and the Winforms version (Windows version, but not non-Windows version) of even other, cross-platform suites have had issues as well.
Dude ... where have you been?
You're like Intel with Atom v. ARM benchmarks and battery ... you claim one thing, then the reviews tell it totally different! You compare to old systems.
It's like my favorite argument ... "NT is as good as trading as Linux." Yes ... Linux 2.2 from 15 years ago! Same thing on Atom v. ARM for performance/power, Windows v. Linux for embedded, etc... Compare against an old version.
I could name people you could talk to, and you wouldn't even know who they are. But if you've been around since the '90s in the open source world, you'd know exactly who they are. Sigh ...
Furthermore, the only reason I cannot name some of the companies is because I'm under NDA. But have a beer with me and I will tell you everything! This whole "conspiracy theory" stuff is not "conspiracy theory." It's just business. It happens. I've been in the middle of it, first-hand, often the guy on-site. Again, it's just the business realities. The stuff you hear in rumors, I was often the guy, or involved with the team.
The only project my team was ever fully recognize for, publicly, was at the Census ('09-11) on NPR, but we were not named by name, only by company (even though there was less than a handful of us from Red Hat). And I'm still under NDA for that.
The most I'll say is that over just a couple of nights (yes, I had other duties with ECON and other entities -- e.g., the ones that do the quarterly economic reports to Congress), I analyzed the prime contractor's deployment, and helped identify all the inconsistencies. We then took that analysis and redeployed the entire compute infrastructure in 3-days ... something they couldn't do in 9 months.
You're literally like the new, "we standardizing on Windows" CIO for the Commerce Department who came over to his own, primary DataCenter and thought we just had a few non-Windows systems. We showed him the few racks of Windows Servers in one module, and then all of our stuff in that single module, plus a half-dozen more modules of 0 Windows. He literally didn't understand, because all he sees is interfaces he thinks is Windows (but often is not). I deal with people like yourself every day, humble, smiling, account management is sometimes difficult, but I (pun)manage(pun).
Heck, we (Red Hat) got permission to bring GM in to look at how we run Datacenters, as they were trying to standardize post-bankruptcy.
Side Note: I turned down the call for the Tech Surge. I had my fill of federal politicking after the 2010 Census.
Ars Technica has its authors, including from various corporations. I don't blame various Microsoft Product Developers from writing articles there.
I'm just saying, in the grand scheme of things, Microsoft has a huge, legacy codebase that doesn't port! You keep missing the fact that I'm talking about existing apps. New apps are different. But a lot of .NET is using existing C/C++ code that is Win32/x86-only.
The Azure team is doing all they can to rewrite as much as they can. Kudos to them, SQL Server and other things are coming to Linux ... because they need them on Linux!
That's good. Because -- assuming you're also doing IdM and federation in Java -- you're likely learning what AD LDS really is (generic LDAP server -- most specifically the same Michigan v3 code that even AD is based on too), and how the non-AD world works, and has always worked.
Yes, because it's not a "canned" solution designed for a specific platform with specific assumptions ... ones that are often only good for 1 Windows release too.
That's the negative side of open systems -- beyond just open standards -- no argument. You're having to learn a lot. Open systems aren't easy. You really have to know a lot! I'm the first one to admit this. It takes decades to get all of the standards and realities down.
Just like -- switching back to directory services -- AD v. LDAP (including AD LDS). If everything is "cannned" down, it's easy. But if it's open and flexible, it requires you to understand far more.
E.g., Everyone thinks the DoD CAC system is AD and its certificate services, because that's all they see. It's not. It will never be. You cannot put 10M (and possibly going to tens of millions in just the next few years) on AD! But you can use just one (1) Red Hat server, (2) for failover (load balancing not so much needed.
But even Red Hat has IPA for a reason, the "canned" version for POSIX attributes (instead of legacy NTLM/SAM). At the same time, IPA isn't going to solve every UNIX LDAP implementation need, far from it. It just offers some base compatibility.
But having SSSD like NT's LSA (which just had a huge, massive, long-standing security hole be exposed), and IPA like AD, is a start. IPA emulating some components that AD expects -- like AD thinking the external forest trust to IPA has a Global Catalog -- goes a long way.
Again, I'm bringing this up because you mentioned IdM and Federation. And that's just platform-level, not application level. I mean, even Oracle dropped Sun One and is focusing on just Oracle LDAP for application-level ... conceding all platform-level to Red Hat.
Heck, Microsoft dropped IdM for UNIX in Windows 10 Server because they believe Red Hat's IPA is the right way forward! External forest trusts, instead of mixed in the same domains/forests ... because schema will always be different!
He actually cares to read and understand my points, while he wants me to understand his.
I've just long been very disappointed -- as a MCP/MCSD myself -- with Microsoft's utter lack of attempting any aspect of portability, many times purposefully, sometimes just ignorance (a MS Office for Mac developer was the most enlightening ... literally hated his x86/Win32 counterparts for lack of understanding anything about data alignment). Now they have this mess they are trying to clean up, along with fighting outsourcing to developers who honestly don't care.
And a lot of Microsoft's own Azure team shares my views too.
When a single post reaches four screens in length and has a rabid use of the spoiler button for no apparent reason, there is no reason to read it by anyone, ever. That might need to replace the definition of insanity. I'm actually dumber for having read the small portion of the post that I did.
And how many different architectures and platforms have you run Windows on? I.e., I seriously doubt it's even half as many as I have supported and developed for -- not just run -- but supported and developed for.
And that's the thing ... nothing I've said is untrue. Everything I've predicted has come true. Everything in this article is marketing with sleight of hand, "Ignore what we promised just 2 years ago!"
It's still amazing that one company can have so many divisions that hate one another and build incompatible core APIs, while an entire, disconnected, huge community will see over 99% agree on one set of core APIs. And you know who are some of people who say that most? The Azure team at Microsoft.
Experience and exposure is where this come from. But go ahead, marvel at the achievement of single codebase that is still heavily Win32/x86-only. E.g., even "Core" Windows still requires a Graphical Display Interface (GDI) as every Windows application is rooted on it, even services.
Again, Windows 10 Nano is the only edition that changes that, and it breaks 100% of existing Windows applications. But I'm glad to see that happen. There are many things that are finally nice to see at Microsoft. But it's sad it took this long ... and how much they've produced is Win32/x86-only.
The XBox One is a PC, entirely, fully. That's why it was able to be ported. And it's wholly unlike the Phones, which are now legacy and have a whole other set of issues. Many companies don't want to work with Microsoft because of them dropping entire developments and solutions.
So, this is what I love, the nice circular. So, let's summarize ...
The only Windows 10 x86 phones are R&D samples, with not much interest, and ...
Have you seen the software limitations of Windows 10 on non-x86?
In other words ...
as Microsoft tries to push everyone towards x86, and drops any ARM other than v8 w/uEFI (support for which will still be questionable), many vendors are getting out. Even Microsoft itself. The existing, pre-uEFI, pre-v8 ARM platforms are only getting sustainment. Many predicted this, including myself, and that's what you're in total denial over. This entire, new strategy is basically what we all said ... x86 first, possibly only, including for new Microsoft products.
Side Note: I've long conceded that Google sucks as sustaining Android in comparison, but that's another story. Trust me, I'm very objective on this, and have my complains about non-GNU, non-Java Android.
Exactly! You have never ported code outside of Windows, likely not even x86 (am I right? )
Because the second you do, you quickly learn you don't write on a Windows platform -- much less in Visual Studio -- first. And that's the thing ... I have extensive experience with porting from/to multiple platforms. I can instantly tell when people think what they do about Windows.
I was an original NT 3.1 Beta tester at the largest installed site of the first native NT app. I've build run-times for every CE and NT release, including having to deal with some of the early StrongARM builds with systems that utterly lacked anything that looked remotely like a PCI hub. I've had Alpha access on several releases. Heck, for years I was on Microsoft Gold Partner lists thanx to colleagues until a few years as an employee for Red Hat.
I'm not some ignorant puke, I've been into the internals of CE and NT for decades.
You mean like the Microsoft Azure team? They are huge on open source, even GPL in several cases!
Or the fact that 85% of Red Hat's Professional Services are not Linuxexerts (HINT: JBoss and application-level).
Seriously ... right there you told me everything!
Knowledge is not bias. Lack of doing something is bias, ergo yourself on porting.
I mean, do you know how much open source Microsoft itself uses in core NT libraries?
E.g., when the zlib security hole hit, there were 2 patches required for Linux ... one in the kernel, one in the shared object (.so). Windows? 17 core files related to the NT executive because they statically linked. And that's just one example. libssl is the second most statically linked library, and even Microsoft branded software (admittedly a lot of acquisitions) have used it instead of the NT security libraries.
Open source is not bigotry ... it's total reality of where most innovation has always come from. That's why Microsoft uses it ... always has! They're stupid not to, although they statically link way, way too much, as do their ISVs, leaving old versions around in various directories for various services.
Microsoft has even crossed from BSD/MIT to GPL licensed in the case of Azure. Geez, even core Microsoft people understand this! Even Gates since '09. Where have you been? You do realize that Microsoft, as of '09, stopped using Cisco and outsourcing to others and just started directly having their own IT and developers admit how much open source they use directly.
You are literally Ballmer. In denial ... and clueless. I cannot educate you because of your assumptions, including about me, and anyone doing open source.
I cannot stress that enough ... there is Ballmer, and the new, Azure-focused leadership, and they are utterly at odds. Ballmer thinks PCs are the future, and x86 the staple. The Azure team realizes that the world has changed, and Microsoft is the #5 player in the mobile space. Microsoft will become a PC vendor to sustain their revenue, all while they try to push people to Office 365.
Heck, even Red Hat evaluated Office 365 and RT devices as a potential solution for exchange with customers who only use MS Office, and found the on-line solutions compatibility lacking because the Windows PC versions still don't use the ISO Office OpenXML 2008 spec. They are on a "transitional" format that is incompatible in some many implementations.
Stop saying "computer science" like I'm some ignorant f'.
I was deploying NT/Alpha -- yes Alpha, not x86 -- back in '96, using Digital FX!32 technologies which Microsoft now uses for its base binary translation in .NET. I'm utterly deep-dive technical on the internals here.
Of course Android is not GNU/Linux or pure Java. I'm the first one to admit that!
Dude ... how many times do I have to say it, I do not like Google and Android. It's poorly sustained with security issues. Heck, Google pulled over Red Hat CTO Brian Stevens to get their containers under control! Because they have sustainment issues on their servers too! Not surprisingly, Google is increasingly tapping Red Hat for sustainment engineering now.
Google is not a good example of Linux! Not remotely!
Side Note (since you don't believe anything I say): I actually interviewed with Brian in April '01 before either of us were with Red Hat, a company later acquired by Red Hat. Unfortunately they went on a hiring freeze, like a lot of company in March-April '01. But I maintained some of the official, Upstream documentation (network, NFS), as well as had a book published, on the subject matter involved (CIFS/SMB, NFS, etc...). The later also featured the founder of Gentoo, who now works for Microsoft (Linux maintainers are not a bunch of Linux fanboys and bigots!). Not surprisingly, I would later work on, and even patch, some of the same software (Cluster Suite -- prime reason I was at the Census, the largest base with peta-bytes, and other places I cannot talk about) with Red Hat, and other, related acquisitions (Gluster, CEPH) as I was the lead, post-sales Enablement Architect for all of Storage ('12-14) and usually had to backfill Consulting engagements due to lack of expertise (especially early adopter stuff).
Listen dude ... for the last time ...
Most of Microsoft Windows is byte-alignment ignorant (pretty much unique to x86), 32-bit Windows x86 (Win32/x86) byte code C libraries and does not port to non-x86 platforms that don't offer Protected386 run-time memory model compatibility (something x86-64/Long Mode does, although not Virtual86, which is another story). That is the bane of Windows' legacy!
It's why RT died. It's too bad we cannot get the old server archive back, because that's exactly what I predicted, and how everything would be x86-first or even only. The entire pre-ARMv8/pre-uEFI Windows Mobile world is now in pure sustainment mode.
It's why most things from .NET never ported to Mono, although the Microsoft Azure is officially re-writing a lot, and no longer having to "play games" (kudos to the Azure team). It's why MS Office has always had issues in its 64-bit implementation, and the Winforms version (Windows version, but not non-Windows version) of even other, cross-platform suites have had issues as well.
Dude ... where have you been?
You're like Intel with Atom v. ARM benchmarks and battery ... you claim one thing, then the reviews tell it totally different! You compare to old systems.
It's like my favorite argument ... "NT is as good as trading as Linux." Yes ... Linux 2.2 from 15 years ago! Same thing on Atom v. ARM for performance/power, Windows v. Linux for embedded, etc... Compare against an old version.
I could name people you could talk to, and you wouldn't even know who they are. But if you've been around since the '90s in the open source world, you'd know exactly who they are. Sigh ...
Furthermore, the only reason I cannot name some of the companies is because I'm under NDA. But have a beer with me and I will tell you everything! This whole "conspiracy theory" stuff is not "conspiracy theory." It's just business. It happens. I've been in the middle of it, first-hand, often the guy on-site. Again, it's just the business realities. The stuff you hear in rumors, I was often the guy, or involved with the team.
The only project my team was ever fully recognize for, publicly, was at the Census ('09-11) on NPR, but we were not named by name, only by company (even though there was less than a handful of us from Red Hat). And I'm still under NDA for that.
The most I'll say is that over just a couple of nights (yes, I had other duties with ECON and other entities -- e.g., the ones that do the quarterly economic reports to Congress), I analyzed the prime contractor's deployment, and helped identify all the inconsistencies. We then took that analysis and redeployed the entire compute infrastructure in 3-days ... something they couldn't do in 9 months.
You're literally like the new, "we standardizing on Windows" CIO for the Commerce Department who came over to his own, primary DataCenter and thought we just had a few non-Windows systems. We showed him the few racks of Windows Servers in one module, and then all of our stuff in that single module, plus a half-dozen more modules of 0 Windows. He literally didn't understand, because all he sees is interfaces he thinks is Windows (but often is not). I deal with people like yourself every day, humble, smiling, account management is sometimes difficult, but I (pun)manage(pun).
Heck, we (Red Hat) got permission to bring GM in to look at how we run Datacenters, as they were trying to standardize post-bankruptcy.
Side Note: I turned down the call for the Tech Surge. I had my fill of federal politicking after the 2010 Census.
Ars Technica has its authors, including from various corporations. I don't blame various Microsoft Product Developers from writing articles there.
I'm just saying, in the grand scheme of things, Microsoft has a huge, legacy codebase that doesn't port! You keep missing the fact that I'm talking about existing apps. New apps are different. But a lot of .NET is using existing C/C++ code that is Win32/x86-only.
The Azure team is doing all they can to rewrite as much as they can. Kudos to them, SQL Server and other things are coming to Linux ... because they need them on Linux!
That's good. Because -- assuming you're also doing IdM and federation in Java -- you're likely learning what AD LDS really is (generic LDAP server -- most specifically the same Michigan v3 code that even AD is based on too), and how the non-AD world works, and has always worked.
Yes, because it's not a "canned" solution designed for a specific platform with specific assumptions ... ones that are often only good for 1 Windows release too.
That's the negative side of open systems -- beyond just open standards -- no argument. You're having to learn a lot. Open systems aren't easy. You really have to know a lot! I'm the first one to admit this. It takes decades to get all of the standards and realities down.
Just like -- switching back to directory services -- AD v. LDAP (including AD LDS). If everything is "cannned" down, it's easy. But if it's open and flexible, it requires you to understand far more.
E.g., Everyone thinks the DoD CAC system is AD and its certificate services, because that's all they see. It's not. It will never be. You cannot put 10M (and possibly going to tens of millions in just the next few years) on AD! But you can use just one (1) Red Hat server, (2) for failover (load balancing not so much needed.
But even Red Hat has IPA for a reason, the "canned" version for POSIX attributes (instead of legacy NTLM/SAM). At the same time, IPA isn't going to solve every UNIX LDAP implementation need, far from it. It just offers some base compatibility.
But having SSSD like NT's LSA (which just had a huge, massive, long-standing security hole be exposed), and IPA like AD, is a start. IPA emulating some components that AD expects -- like AD thinking the external forest trust to IPA has a Global Catalog -- goes a long way.
Again, I'm bringing this up because you mentioned IdM and Federation. And that's just platform-level, not application level. I mean, even Oracle dropped Sun One and is focusing on just Oracle LDAP for application-level ... conceding all platform-level to Red Hat.
Heck, Microsoft dropped IdM for UNIX in Windows 10 Server because they believe Red Hat's IPA is the right way forward! External forest trusts, instead of mixed in the same domains/forests ... because schema will always be different!
You keep saying false things. I show them to be false. Then you ignore this, move the goal post and say everything you've said ever has been correct, even your predictions. Again, please show me a shipping Windows 10 Mobile device running on x86. Here's the ones running on Arm.
Alcatel OneTouch Fierce XL
BLU Win HD LTE x150e
BLU Win HD W510U
BLU Win HD LTE X150Q
BLU Win JR x130e
20 different Lumia devices
Mouse Computer Madosma Q501
Xiaomi Mi4
Please show me a modern platform running some form of cohesive app model across phone, tablet, desktop, and beyond outside of Windows. I'm fully aware that legacy code does not work on all devices. When did I say otherwise? I don't know why you're freaking out about this.
You keep going on and on and on and on about your credentials... I DON'T CARE. Make a cohesive argument and support it with facts, not your credentials laced with things you can't talk about that you did in the 90s. You really need to go back to Comp 2 and learn how to write persuasively. This thread started with an article written by an independent journalist.
Why do you keep talking about Ballmer as if he's the CEO of Microsoft? Spoiler alert: he's not.
We've been using AD LDS (formerly ADAM) for probably a decade. To me it doesn't have much use. It's not agile enough to be a virtual directory server but it's not full featured enough as it doesn't support Kerberos.
The DoD CAC system is complete and utter garbage. If you had anything to do with that go ef yourself. They issue multiple new CAs each year presumably because their systems can't handle the amount of certs. They also insist on revoking all their certs when they expiring causing their CRLs to blow up to 20mb in some cases. I'm not sure what you're referring to with their CA issuing 10M certs. Why do they have 44 issuers (in their main line) if that's the case? Just so their CRLs don't get even larger?
Now my responses are getting stupidly long just because your bullshit is 100times longer. UUUggghhhh
You keep saying false things. I show them to be false. Then you ignore this, move the goal post and say everything you've said ever has been correct, even your predictions. Again, please show me a shipping Windows 10 Mobile device running on x86. Here's the ones running on Arm.
Alcatel OneTouch Fierce XL
BLU Win HD LTE x150e
BLU Win HD W510U
BLU Win HD LTE X150Q
BLU Win JR x130e
20 different Lumia devices
Mouse Computer Madosma Q501
Xiaomi Mi4
And all those are running Windows 10 with all applications across all platforms as if they were no different? Dude, seriously? I mean, what "GUI" is on it!
Does the Windows 10 Raspberry Pi run even the base Word+Excel+Powerpoint version of MS Office? Can it fully edit and exchange documents with PCs?
Now keep in mind that full-blown LibreOffice, every single GUI component, was running on Raspberry Pi almost 4 years ago ... basically when there was a version (just barely) powerful enough to do a GUI! Heck, we had Linux/ARM running full blown office suites from Corel and Sun in the early '00s!
Seriously ... what you sayis a joke to anyone who has done anything outside of Windows. I cannot believe you assert what you did about this. Forget me ... you really need to go meet some other GNU/Linux maintainer in this space!
It's mass assumptionbased on 0 -- zero -- experience outside Windows. And I'm the fanboi?
P.S. As far as Comp2, you know this is bottom-posting, correct? It's general rambling in response, not an organized, introduction, point by paragraph, and summary ... because it's led by someone else. I'd mention that I have over 400 pages in print books and magazines, and never had any difficulty with paragraph structure (my various editors at CMP Media, MacMilian and others can attest to that), but you'd just accuse me of stating more credentials. Did you ever think to realize that I only state my credential after you first accuse me of not having experience and exposure?
Does the lunatic still use that god damn annoying emoji after every effing sentence? The worse he gets his ass beat in an argument the more he uses those. I imagine he's using them a lot seeing that woot continues to whip his ass like the lunatic stole something from him.
Again I wouldn't know but we only have to assume this because horrible people make horrible arguments.
Does the lunatic still use that god damn annoying emoji after every effing sentence? The worse he gets his ass beat in an argument the more he uses those. I imagine he's using them a lot seeing that woot continues to whip his ass like the lunatic stole something from him.
Again I wouldn't know but we only have to assume this because horrible people make horrible arguments.
Does the lunatic still use that god damn annoying emoji after every effing sentence? The worse he gets his ass beat in an argument the more he uses those. I imagine he's using them a lot seeing that woot continues to whip his ass like the lunatic stole something from him.
Again I wouldn't know but we only have to assume this because horrible people make horrible arguments.
Yes, I'm a horrible person. I even shot JFK and influenced W. to invade Iraq. Seriously people, how many times have I cut down people here, compared to you guys? I'm screwed up because I try to stay logical and present a viewpoint? I'm pathetic because I won't just cut someone down in a short response? And that's the thing, you guys are beyond cruel. And you know it.
I focus on the discussion, the specifics and only push back when someone questions my experience, especially after someone else clearly doesn't (and he's already backtracking on that). Seriously, in this area, I'm an uniquely qualified expert, and come by 1st hand reference, nationally. But even if I wasn't ...
My whole, original point years ago predicted everything that has happened ... but for anyone who has done many platforms and ports in this space, it was inevitable, and easily predicted! Heck, the switch back (like the original XBox) to x86 by Microsoft in the XBox One was strategic in the first place. Sony regrets going with the AMD Jaguar, but Microsoft had to.
Disclaimer: Several of my fellow, Red Fedoras are now at Microsoft, and I still keep in contact with many.
As Microsoft keeps integrating, now openly, more Open Source for core infrastructure, it's inevitable. With a subset of Powershell coming over to Azure and Linux, they are running into the same things I have. Things just don't port. And the WINE (Windows Emulator) run-times, among others, are required to emulate portions of the Win32/x86-only Windows Executive, Task Manager and other functions in Visual Studio/WinForms standard objects.
I.e., way, way too much of the Windows ecosystem is just Win32/x86-only, and the overwhelming majority of .NET classes have dependencies on them, preventing native porting to any other architecture (let alone platform). This has now been proven and beaten like a dead horse as Microsoft tries to port more and more to Linux, as Fortune 500 companies demand such. The future is really an Intel-only Windows.
Even the rumors continue to swirl than any ARM-based Microsoft Phone, should they decide to release one, will be running Android. So far, it just seems like some Intel products will make it to market, but even Intel is having severe trouble keeping Atom competitive with ARM -- so much so they've just gone i-series in higher-performance cases.
So "One Windows" may be very well re-defined to include ... native iOS and Android apps. In fact, considering Windows/.NET is so unportable, like with many aspects of Azure's own infrastructure and services, it wouldn't be the first time Microsoft went Open Source and Linux for its own needs ... because it's more portable and supportable.