09-14-2014 11:01 PM
41 12
tools
  1. Jas00555's Avatar
    People seem to forget that Microsoft makes 20% of their annual revenue from consumer versions of Windows. That would be like Apple giving away the iPad for free. It just won't happen. I still stand by my theory that to please the consumer (thanks to mobile, is use to a few bigger updates a year) and the enterprise (huge updates once a year), they'll probably have to have a consumer and enterprise Windows where the enterprise would be comparable to Ubuntu's LTE version.
    LexTaylor likes this.
    09-10-2014 04:06 AM
  2. LexTaylor's Avatar
    After reading everyone's post , I feel like Windows 9 just might be the big break that Windows need like Win 7 was a huge success ONLY! cause the failure of Win Vista , I feel like Windows make up for their issues and faults when they release a new OS. As for giving it away for free , maybe a beta version like they did with Windows 8 - however you cannot revert back to Win 7 after doing that install trust me I learned the hard way , I do see that with Windows 9 it might just be the big thing Windows needs and releasing it free well thats another story cause like Jas00555 stated its like Apple giving away free iPads , however Apple isn't likely to give away their products , Microsoft might just do it
    09-11-2014 07:06 PM
  3. Mike Gibson's Avatar
    I think it's important to separate two different things: upgrades purchased by consumers vs. OEM licenses for new machines. AFAIK, consumers rarely purchase upgrades so it wouldn't hurt much to offer free upgrades to them. However, as someone previously noted, if that free upgrade hurts new machine purchases then the OEMs will scream ... and the OEM-MSFT relationship has already been strained by Surface and the Win8/WinRT disaster. The consumer-MSFT relationship has also been strained by the Win8/WinRT disaster, possibly permanently. The Win8/WinRT disaster could not have happened at a worse time because there are several viable alternatives to Windows these days. If I was Nadella I'd do the following:

    1. Offer free upgrades to Win9 for Win8 users as an apology.

    2. Charge a nominal fee for Win7 upgrades, say $49 (it won't matter since few will upgrade anyway).

    3. Fire everyone involved in Win8, especially those responsible for the WinRT framework. Getting rid of Sinofsky and Ballmer was a good start but the actual designers of the WinRT framework should never be allowed near a computer again.

    4. Deprecate the WinRT API.

    5. Add new touch/gesture support to Win32.

    6. Add a native scalable UI API to Win32 (might as well stick with XAML).

    7. Add a Win32 Store but only take a 10% cut of sales.

    8. Backport items 5-7 to Win7 ... or else it too will be DOA.

    I think that Win8 was a fatal blow to MSFT so it might not matter what they try. Most likely they'll retreat to be an enterprise-only company and then slowly fade away.
    DavidinCT likes this.
    09-11-2014 07:48 PM
  4. tgp's Avatar
    Most likely they'll retreat to be an enterprise-only company and then slowly fade away.
    I don't think that this is the beginning of the end for Microsoft. I do feel, however, that in a few years down the road Microsoft might turn out to be primarily enterprise. It's currently their bread & butter, besides Windows. But I don't think that Windows will continue to be the cash cow it has been in the past.

    The longer this mobile lethargy goes on, the more I question whether it will go anywhere. I don't see WP going away anytime soon, besides the name of course. Microsoft has the resources to subsidize it forever. But if it doesn't gain significant market share in the next few years, I'd revisit that opinion.

    Google doesn't make anything from Android directly either, but it is the means to an end. I'm sure they feel it's paying off. If not they would do something different. Was Microsoft's intention for WP to be a tool to get users onto Microsoft services, or was it to make a profit from sales? It was probably some of both. Microsoft originally sold the licenses to Nokia and other OEMs, and then later bought Nokia's devices division. But even if WP was solely intended as a vehicle to get users onto Microsoft's services, are the numbers enough to justify it?

    There are several possible reasons that Microsoft keeps WP going:

    1. They still think it will take off.
    2. It pays for itself by bringing users to their ecosystem.
    3. To save face.
    4. Poor management.

    Anyway, let's hope it takes off. I do my part: in the last year and a half or so I've bought around 10 new or used Windows Phones! I've since sold most of them, but I still have 2.
    09-11-2014 09:04 PM
  5. RajeevT's Avatar
    4. Deprecate the WinRT API.
    Not gonna happen. Instead I expect it to be improved in the future.

    8. Backport items 5-7 to Win7 ... or else it too will be DOA.
    [strikeout]Or else Win7 will DOA? In 2014? What?[/strikeout] Ok, re-read it and it's badly written. By "it" you were referring to these 3 features.

    Why would they backport all these features to a 2 version old OS that is barely 4 months away from end of mainstream support? Just to increase their headaches, waste resources and so that it becomes even more certain that Win7 will be the new XP when it comes to prolonging support? Adding them to Win8 also along with Win9 I can understand, but adding to Win7 will just mean that much less incentive for people to ever upgrade.

    I think that Win8 was a fatal blow to MSFT so it might not matter what they try. Most likely they'll retreat to be an enterprise-only company and then slowly fade away.
    Disagree on both counts.
    DoctorSaline likes this.
    09-11-2014 11:29 PM
  6. Mike Gibson's Avatar
    I don't think that this is the beginning of the end for Microsoft. I do feel, however, that in a few years down the road Microsoft might turn out to be primarily enterprise. It's currently their bread & butter, besides Windows. But I don't think that Windows will continue to be the cash cow it has been in the past.
    I think that Windows is around 1/3 of MSFT's profit. If that goes away then MSFT becomes a much smaller company. It also means that they would have less cash to use in new areas ... i.e. the snowball effect.

    The longer this mobile lethargy goes on, the more I question whether it will go anywhere. I don't see WP going away anytime soon, besides the name of course. Microsoft has the resources to subsidize it forever. But if it doesn't gain significant market share in the next few years, I'd revisit that opinion.
    If they lose their Windows cash-cow then MSFT will not have the "resources" to fund unprofitable ventures. MSFT has a lot of cash in the bank but that won't last forever if they're forced to spend it defending their cash cow(s).

    Was Microsoft's intention for WP to be a tool to get users onto Microsoft services, or was it to make a profit from sales? It was probably some of both. Microsoft originally sold the licenses to Nokia and other OEMs, and then later bought Nokia's devices division. But even if WP was solely intended as a vehicle to get users onto Microsoft's services, are the numbers enough to justify it?
    MSFT's "services" like MSN, Bing, etc. have been a money pit from the beginning, kept alive by the profits from their Windows/Office platform. I think MSFT realized that iOS and Android's popularity was eventually going to pull people off of MSFT's platform (Windows, Office, etc.) once AAPL and GOOG evolve their phone userbases to more capable systems with larger screens (OSX and ChromeOS).

    What's funny is that MSFT did the exact same thing to minicomputer and workstation companies back in the 1980s and 1990s. PCs were considered a joke ("toys") back then but got bigger and better over time. The mini and workstation vendors (Wang, DEC, Sun, etc.) got eaten alive from the bottom up.

    To prevent being eaten alive, MSFT needed to get serious in the phone/tablet space. So, they whipped out their big gun, Windows, and added a Phone UI (Metro) to it. I imagine their thinking was to get the 1+ billion Windows users used to the new Phone UI on their computers and then those users would naturally start buying Windows phones and tablets as companion devices. It didn't work. Not only did users not migrate to Windows phones/tablets as hoped -- users stopped buying Windows computers because of its inappropriate UI! That's the user side of the disaster, I won't even go into the developer disaster that is WinRT.

    Anyway, let's hope it takes off. I do my part: in the last year and a half or so I've bought around 10 new or used Windows Phones! I've since sold most of them, but I still have 2.
    As an MSFT shareholder I thank you for the support! As a regular person I always say that people should buy what meets their needs. That's the best way to send a signal to a company. MSFT needs to see that they're on the wrong track and plunging sales is the best way to open their eyes.
    09-12-2014 09:53 AM
  7. Mike Gibson's Avatar
    Not gonna happen. Instead I expect it to be improved in the future.
    Lipstick on a pig.

    [strikeout]Or else Win7 will DOA? In 2014? What?[/strikeout] Ok, re-read it and it's badly written. By "it" you were referring to these 3 features.
    Yes, it wasn't clear that I was referring to the three items together as "it". Basically, any new stuff they add to Win32 in Win9 won't succeed with developers unless it is compatible and available on Win7.

    Why would they backport all these features to a 2 version old OS that is barely 4 months away from end of mainstream support? Just to increase their headaches, waste resources and so that it becomes even more certain that Win7 will be the new XP when it comes to prolonging support? Adding them to Win8 also along with Win9 I can understand, but adding to Win7 will just mean that much less incentive for people to ever upgrade.
    Win7 is the new XP. MSFT needs to deal with and take advantage of that fact. The current "Windows Store" is really a "WinRT Store" and it is failing badly because:

    1. Users don't like the Metro UI on Win8 ==> so no viable user base
    2. It requires ISVs to almost completely rewrite existing Win32 software ==> requires significant ISV investment
    3. MSFT takes a 20-30% cut of already dismal sales ==> less return on ISV investment

    The WinRT API/framework is all pain, no gain for ISVs. It was DOA.

    If MSFT had simply defined a cleaned up Win32 API (cut out old, obsolete stuff), added more security in the API (not go overboard like in WinRT), created a Store that only took a 10% cut of sales, and backported the system to Win7 then they would have had a winner. Instead they started with a new, crippled WinRT framework that had zero users and zero developers. What were they thinking???
    09-12-2014 11:07 AM
  8. RajeevT's Avatar
    Instead they started with a new, crippled WinRT framework that had zero users and zero developers.
    Everything has to start somewhere, even Win32 did.

    Lipstick on a pig.
    So what did Win32 look like back when it was launched I wonder? WinRT sucks right now according to you. Why can't it be improved?
    DoctorSaline likes this.
    09-12-2014 10:23 PM
  9. Mike Gibson's Avatar
    Everything has to start somewhere, even Win32 did.

    So what did Win32 look like back when it was launched I wonder? WinRT sucks right now according to you. Why can't it be improved?
    Win32 was a straightforward extension of Win16. You could (almost) just recompile a Win16 program to make it Win32 and gain all the advantages. If you did a lot of segment-specific stuff in your Win16 code it was far more involved.

    With WinRT, you basically have to start from scratch with any code that deals with the OS and UI. My particular Win32 program was lucky in that much of its work is performed via Direct3D, which was brought over to WinRT almost intact. Everything else was a disaster. One culprit was the stupid Async fetish of the WinRT designers. Win32 already had the best multithreading support of any OS. Instead of taking advantage of that fact, WinRT chops every operation into Async calls. Every source code example I've seen simply turns them into a set of sequential Async calls (in C# you see "await OpenAsync; await ReadAsync, etc.). It's stupid and a performance killer. You get far better performance, in the order-of-magnitude range, kicking off a thread to do standard synchronous Open+Read+Close. However, it wouldn't matter because the actual file operations go through the WinRT File Broker, which is (I'm not joking) two orders-of-magnitude slower than the Win32 file APIs.

    In WinPRT 8.1, the "universal" WinRT version for WP 8.1, you have the FilePicker abomination. Unbelievably, before calling the FilePicker API you have to prepare your WP 8.1 app to be shut down. It may not get shut down but you have to be able to handle it anyway. Who would ever design such an abomination and why are they still allowed near a computer anymore???

    These are just two examples of the WinRT API disaster. Everywhere you look you find similar inexplicable design decisions. Note that this isn't simply a developer issue, you see the WinRT API design and implementation defects surface to the UI and apps, too. For example, an app that displays a list of photos becomes slow as molasses if there are hundreds of image files to enumerate. This is the File Broker getting in the way. In Win32 is would be almost instantaneous while the File Broker causes WinRT apps to take 10-30 seconds to do the same.
    DavidinCT likes this.
    09-13-2014 01:05 AM
  10. RajeevT's Avatar
    (Why WinRT sucks)
    Ok, I read through your post and what I got from it was that WinRT sucks and a few major reasons why, but again no answer to my question about why it can't be improved. I'm optimistic and if it's bad I see no reason for it to remain that way forever, but I can see how some people might look at it pessimistically and say nothing will ever improve. Hopefully MS proves those people wrong.
    09-13-2014 08:54 PM
  11. Jas00555's Avatar
    Win32 was a straightforward extension of Win16. You could (almost) just recompile a Win16 program to make it Win32 and gain all the advantages. If you did a lot of segment-specific stuff in your Win16 code it was far more involved.

    With WinRT, you basically have to start from scratch with any code that deals with the OS and UI. My particular Win32 program was lucky in that much of its work is performed via Direct3D, which was brought over to WinRT almost intact. Everything else was a disaster. One culprit was the stupid Async fetish of the WinRT designers. Win32 already had the best multithreading support of any OS. Instead of taking advantage of that fact, WinRT chops every operation into Async calls. Every source code example I've seen simply turns them into a set of sequential Async calls (in C# you see "await OpenAsync; await ReadAsync, etc.). It's stupid and a performance killer. You get far better performance, in the order-of-magnitude range, kicking off a thread to do standard synchronous Open+Read+Close. However, it wouldn't matter because the actual file operations go through the WinRT File Broker, which is (I'm not joking) two orders-of-magnitude slower than the Win32 file APIs.

    In WinPRT 8.1, the "universal" WinRT version for WP 8.1, you have the FilePicker abomination. Unbelievably, before calling the FilePicker API you have to prepare your WP 8.1 app to be shut down. It may not get shut down but you have to be able to handle it anyway. Who would ever design such an abomination and why are they still allowed near a computer anymore???

    These are just two examples of the WinRT API disaster. Everywhere you look you find similar inexplicable design decisions. Note that this isn't simply a developer issue, you see the WinRT API design and implementation defects surface to the UI and apps, too. For example, an app that displays a list of photos becomes slow as molasses if there are hundreds of image files to enumerate. This is the File Broker getting in the way. In Win32 is would be almost instantaneous while the File Broker causes WinRT apps to take 10-30 seconds to do the same.
    I understand what you're saying and trust when I say that I have experienced the pain you're feeling, but as a developer, I think that getting rid of the WinRT API would be a very very stupid and short-sighted decision.

    Microsoft has started something and gone back on a lot of things and it constantly P/O developers. Microsoft needs to show that it has a consistent vision so that ways developers don't go "well, Microsoft could change this at any time, so it's not worth the potential risk". About a 3rd of the ISVs that I've talked to have said this. Many of the issues with WinRT can be fixed and I think the better option would be to make the transition from Win32 to WinRT more seamless while continuing to expand WinRT. I get what you're saying, I really do, but I personally think it's a little short-sighted and a little stuck in the past.
    Last edited by Jas00555; 09-13-2014 at 10:19 PM. Reason: clarification
    RajeevT likes this.
    09-13-2014 09:21 PM
  12. Mike Gibson's Avatar
    Ok, I read through your post and what I got from it was that WinRT sucks and a few major reasons why, but again no answer to my question about why it can't be improved. I'm optimistic and if it's bad I see no reason for it to remain that way forever, but I can see how some people might look at it pessimistically and say nothing will ever improve. Hopefully MS proves those people wrong.
    Frankly, I don't think the SDEs designing and implementing WinRT are capable of producing a quality system no matter how long they work on it. They make Programming 101 mistakes constantly. The horrible File Broker performance is a perfect example. If I, as an SDE implementing this thing, know that the IPC call to the FB is slow then I'm going to go to great lengths to hide it. The most obvious solution is to do a full FindFirst/FindNext enumeration in the FB then throw the entire block of results back over the IPC fence to the WinRT app. The enumeration takes 15-30 milliseconds and tossing the results over the fence would be, I don't know, 50 more milliseconds. This is an obvious optimization that anyone with any experience at all would do.

    The FilePicker/shutdown-your-wp81-app abomination is so bad that I question the competence of everyone involved in it. How did they not decide to go with reduced FilePicker functionality in low memory situations (e.g. show a simple text file list instead of a full-blown file sample)? In the worst case scenario, how about simply returning "out of memory"? No, they simply shifted the burden from one SDE in the WinRT group to every single freaking developer in the world for the life of WinPRT.

    Look, I've been there and done that. I know that writing OS code has constraints that aren't obvious to outsiders ... but the extremely obvious solutions to the problems I'm discussing are the first things to try when faced with those constraints. It's like the WinRT devs don't give a **** about anyone else and are just phoning-in their designs and implementations. MSFT needs to wipe the slate clean and bring in people who have a freaking clue.
    09-13-2014 11:06 PM
  13. Mike Gibson's Avatar
    Microsoft has started something and gone back on a lot of things and it constantly P/O developers. Microsoft needs to show that it has a consistent vision so that ways developers don't go "well, Microsoft could change this at any time, so it's not worth the potential risk". About a 3rd of the ISVs that I've talked to have said this.
    I would rather MSFT **** off the few WinRT ISVs than continue on wasting time and effort on a fundamentally flawed system. They pushed the reset button to a far greater extent when they nuked OS/2 in favor of Win16 and that led to the greatest expansion in MSFT's history.

    Many of the issues with WinRT can be fixed and I think the better option would be to make the transition from Win32 to WinRT more seamless while continuing to expand WinRT. I get what you're saying, I really do, but I personally think it's a little short-sighted and a little stuck in the past.
    There's no way to make the code transition from Win32 to WinRT more seamless. They changed everything, from internet access all the way down to file access.

    If MSFT wants to be relevant at all in the future then they need to regroup around Win7 compatibility. Improve Win32 for Win9 and backport those improvements to Win7. That's the only way to get the few remaining Windows ISVs to use them. If they're Win9-only then they will be DOA.
    09-14-2014 12:01 AM
  14. smoledman's Avatar
    So the huge premium they charged for the hardware counted for nothing at all? Also I bet it costs Apple far less to test their OS because of very limited hardware configurations, whereas MS has to invest huge amounts in setting up large test labs with a vast cross-section of hardware including peripherals.
    Exactly the cost of 4 years of OS upgrades is factored into the price of any Mac. By 4 years it's too underpowered to run the 5th update to OS X and it's time to buy another Mac.
    RajeevT likes this.
    09-14-2014 12:17 AM
  15. Jas00555's Avatar
    I would rather MSFT **** off the few WinRT ISVs than continue on wasting time and effort on a fundamentally flawed system. They pushed the reset button to a far greater extent when they nuked OS/2 in favor of Win16 and that led to the greatest expansion in MSFT's history.


    There's no way to make the code transition from Win32 to WinRT more seamless. They changed everything, from internet access all the way down to file access.

    If MSFT wants to be relevant at all in the future then they need to regroup around Win7 compatibility. Improve Win32 for Win9 and backport those improvements to Win7. That's the only way to get the few remaining Windows ISVs to use them. If they're Win9-only then they will be DOA.
    And while from our perspective that would be amazing on the desktop, what would I do as a Windows Phone developer if I wanted to make a Universal app? Unless you're also suggesting that they port Win32 to WP, then WP will lose a big advantage that has thus far got some big name apps onto the platform that otherwise might not have got written (for either platform considering that W8's weak audience + WP's weak audience = just big enough to develop for) and could further leave WP behind in apps. That's why I think it's better to improve WinRT (and I honestly think that I can see some ways for them to help the transition from Win32 to WinRT, but that's a whole new thread) than to hit the reset button and scrap it.
    RajeevT likes this.
    09-14-2014 06:32 PM
  16. Mike Gibson's Avatar
    And while from our perspective that would be amazing on the desktop, what would I do as a Windows Phone developer if I wanted to make a Universal app? Unless you're also suggesting that they port Win32 to WP, then WP will lose a big advantage that has thus far got some big name apps onto the platform that otherwise might not have got written (for either platform considering that W8's weak audience + WP's weak audience = just big enough to develop for) and could further leave WP behind in apps. That's why I think it's better to improve WinRT (and I honestly think that I can see some ways for them to help the transition from Win32 to WinRT, but that's a whole new thread) than to hit the reset button and scrap it.
    All Win32 needed to cover multiple device types was a native scalable+touch UI API. Basically a replacement for USER. Win32 was already portable across CPU architectures. The Surface RT uses Win32 recompiled for ARM (along with Desktop Office). However, I wouldn't have ported *all* of Win32 but rather defined a more secure and modern subset with a native scalable+touch UI API to it. That way it could have been backported to Win7. It is an all gain, minimal pain solution to the problem. Think about it: what is the main difference between a phone, a tablet, a laptop, etc.? It's the screen size. MSFT should have tackled that problem, not wasted their time reinventing internet and file access.

    MSFT is almost there with their Universal WinRT API. It incorporates huge chunks of the modern Win32 API already: Direct2D, Direct3D, DirectWrite, etc. That's available on all devices, all the way down to a lowly 530. But then the ugly and unnecessary Async crap comes in and ruins the party in everything else (file access, internet access, etc.). Then they went overboard on security which essentially limits apps to simple trinkets. Then the app model itself where you're always being suspended or tossed out of memory -- which kills the potential for complex programs. The WinRT API and framework are simply not capable of supporting anything more than applets.

    Anyway, if you start a thread on how to ease the transition from Win32 to WinRT, send me a note or post a link here. I'd like to hear your thoughts.

    Edit: here's a post on MSDN just today from an ISV fighting with the absolutely horrible performance in WinRT/WinPRT's file functions:

    Terrible performance of StorageFolder.GetBasicPropertiesAsync. What are the alternatives?

    Sorry pal, there are no alternatives. Win32's FindFirst/FindNext would have returned the entire file list in the time taken for the Async bull**** to return a single entry --- but MSFT doesn't let you use FF/FN except for app local storage.
    Last edited by Mike Gibson; 09-15-2014 at 09:34 PM.
    09-14-2014 11:01 PM
41 12

Similar Threads

  1. Gameloft brings medieval battles to Windows 8 with Rival Knights
    By WindowsCentral.com in forum Windows Central News Discussion
    Replies: 0
    Last Post: 07-02-2014, 04:00 AM
  2. Microsoft wants you to beta test upcoming Office products and services
    By WindowsCentral.com in forum Windows Central News Discussion
    Replies: 0
    Last Post: 07-02-2014, 03:11 AM
  3. Microsoft joins Qualcomm-backed AllSeen Alliance to standardize connected home tech
    By WindowsCentral.com in forum Windows Central News Discussion
    Replies: 0
    Last Post: 07-02-2014, 02:50 AM
  4. Paytm gets updated, but still doesn't bring mobile shopping to Windows Phone
    By WindowsCentral.com in forum Windows Central News Discussion
    Replies: 0
    Last Post: 07-02-2014, 02:30 AM
LINK TO POST COPIED TO CLIPBOARD