Coffeehouse Thread

52 posts

Forum Read Only

This forum has been made read only by the site admins. No new threads or comments can be added.

Windows 8.1 RTM delayed for MSDN/Technet Subscribers?

Back to Forum: Coffeehouse
  • User profile image
    bondsbw

    @ScanIAm:  In computer architecture class, I remember learning a bit about the history of Intel's 8086 architecture and how it was in competition with MIPS and their RISC processor.  Intel knew that x86 was an inferior technology, but they were ahead of MIPS by a few months in terms of Moore's Law.  Intel had the chance to move to a RISC architecture, but by the time Intel implemented one that could compete with MIPS, they would be behind.  So Intel scrapped the idea and stayed the course with x86.

    x86 was inferior to RISC, but Intel has a canyon full of cash for their decision that first to market was better than going with a superior technology.

  • User profile image
    DeathBy​VisualStudio

    , ryanb wrote

    That's not necessarily a reflection of anything more than the software not being ready yet.  You can't release it on MSDN a couple months early if it isn't done a couple months early.  As product cycles get shorter, that two months becomes a more significant chunk of bug-fixing time.

    But that's still a problem if you are trying to have your product ready at launch.

    Nail meets hammer. I believe this may be the new reality due to Rapid Release. So then the question is what are OEMs actually getting with the RTM version? A placeholder that will need to receive a sizable day-one update?

    If we all believed in unicorns and fairies the world would be a better place.
    Last modified
  • User profile image
    evildictait​or

    , DeathBy​VisualStudio wrote

    Nail meets hammer. I believe this may be the new reality due to Rapid Release. So then the question is what are OEMs actually getting with the RTM version? A placeholder that will need to receive a sizable day-one update?

    Typically, OEMs put the OS on during manufacture anyway (it's called ZDP), and the date when the OEMs will be given the image will have been fixed long, long ago.

    In this case, as well, Win8.1 is being distributed as an online update anyway, so it's not even clear that for this release OEMs will have the image before other customers will anyway. OEMs will of course want to ship with Win8.1 soon after Win8.1 is available, but there's no downside to their customers of OEMs shipping with Win8 even after Win8.1 installs, because Win8.1 will then install during their first boot of the machine as part of the install process.

    So if OEMs are a week or two late to the Win8.1 party, their customers, the OEMs themselves and frankly Microsoft probably don't really care all that much; if they turn on Automatic Updates, they'll have it in a day or two anyway (and if they don't have automatic updates turned on, they have bigger problems than not getting Win8.1).

    Also: if you really want a Win8.1 machine to develop against, what's wrong with using the Win8.1 preview that Microsoft released some weeks back? The changes between the preview and the RTM are really bugfixes at this point.

    Sure, it's not byte-for-byte what users will be using. But all of the documented APIs in kernel and usermode are fixed at this point, and unless your code is particularly fragile I can't think of a reason why you wouldn't be able to develop your app against it and have it work out of the box when 8.1 launches later this year.

  • User profile image
    elmer

    , ZippyV wrote

    *snip*

    Not if MSDN delays it.

    Sorry, what I meant was, that I thought device-drivers are typically written by devs working for hardware OEMs, and I was under the impression that they will be receiving the RTM immediately.

  • User profile image
    DeathBy​VisualStudio

    , evildictait​or wrote

    *snip*

    Typically, OEMs put the OS on during manufacture anyway (it's called ZDP), and the date when the OEMs will be given the image will have been fixed long, long ago.

    In this case, as well, Win8.1 is being distributed as an online update anyway, so it's not even clear that for this release OEMs will have the image before other customers will anyway. OEMs will of course want to ship with Win8.1 soon after Win8.1 is available, but there's no downside to their customers of OEMs shipping with Win8 even after Win8.1 installs, because Win8.1 will then install during their first boot of the machine as part of the install process.

    So if OEMs are a week or two late to the Win8.1 party, their customers, the OEMs themselves and frankly Microsoft probably don't really care all that much; if they turn on Automatic Updates, they'll have it in a day or two anyway (and if they don't have automatic updates turned on, they have bigger problems than not getting Win8.1).

     

    I thought the idea behind W8.1 was in part to help spur the adoption of W8.x devices. If I were an OEM I'd want my product box, product literature, etc all to be stating that this is a W8.1 device. When my customer fires up their brand spanking new device I'd want them to be able to use it immediately and have a fantastic out-of-box experience. Having to wait for a massive W8.1 download and install does not provide that great experience IMO.

    Also does it really only take a week or two for OEMs to switch over their product lines to use a new OS and get that product boxed and to retail? That's pretty fast IMO. 

    Also: if you really want a Win8.1 machine to develop against, what's wrong with using the Win8.1 preview that Microsoft released some weeks back? The changes between the preview and the RTM are really bugfixes at this point.

    Sure, it's not byte-for-byte what users will be using. But all of the documented APIs in kernel and usermode are fixed at this point, and unless your code is particularly fragile I can't think of a reason why you wouldn't be able to develop your app against it and have it work out of the box when 8.1 launches later this year.

    Is this in response to this post of mine or are you just adding to your original post on the same subject here? I'm a little confused on the context of these remarks.

    If we all believed in unicorns and fairies the world would be a better place.
    Last modified
  • User profile image
    evildictait​or

    , DeathBy​VisualStudio wrote

    I thought the idea behind W8.1 was in part to help spur the adoption of W8.x devices. If I were an OEM I'd want my product box, product literature, etc all to be stating that this is a W8.1 device. When my customer fires up their brand spanking new device I'd want them to be able to use it immediately and have a fantastic out-of-box experience. Having to wait for a massive W8.1 download and install does not provide that great experience IMO.

    Also does it really only take a week or two for OEMs to switch over their product lines to use a new OS and get that product boxed and to retail? That's pretty fast IMO. 

    No more so than an OEM would want to advertise "Windows 7 SP2" versus "Windows 7" on their box. There are a couple of UX changes, sure, but first boot of a Win8 device will download the Win8.1 update via Windows Update before the user gets to the desktop (or when they first have access to Internet with which to download the Win8.1 update), so most users accidentally purchasing a Win8 rather than Win8.1 device won't notice the difference, other than the longer than expected boot-to-first-desktop experience.

    Is this in response to this post of mine or are you just adding to your original post on the same subject here? I'm a little confused on the context of these remarks.

    I was adding to my previous post. 

    Admittedly I have no intention of writing Win8.1 specific products any time soon, but it looks to me like if you write software for Win8.1 Preview you'll have no problem getting it to work on Win8.1 RTM, with the only exception of some pathological cases like DRM or things like StarDock that have horrendous dependencies on undocumented or volatile features of the OS, in which case working on Win8.1 Preview is probably good enough for them to only have a few weeks of work left when RTM ships anyhow.

  • User profile image
    DeathBy​VisualStudio

    , evildictait​or wrote

    *snip*

    No more so than an OEM would want to advertise "Windows 7 SP2" versus "Windows 7" on their box. There are a couple of UX changes, sure, but first boot of a Win8 device will download the Win8.1 update via Windows Update before the user gets to the desktop (or when they first have access to Internet with which to download the Win8.1 update), so most users accidentally purchasing a Win8 rather than Win8.1 device won't notice the difference, other than the longer than expected boot-to-first-desktop experience.

    So W8.1 is just a service pack? Seems to me Microsoft packed in a whole lot more than that. I thought Microsoft was trying to get people to take another look at Windows 8. Kind of like the idea that "We fixed what you didn't like about Windows 8 and made it even better." IMO W8.1 is an opportunity for Microsoft to get people to take that second look. They won't get that with OEMs shipping W8 devices. With Ballmer admitting that they need to do better with W8 sales I'd think framing it as a service pack would be contrary to that effort.

    If we all believed in unicorns and fairies the world would be a better place.
    Last modified
  • User profile image
    evildictait​or

    , DeathBy​VisualStudio wrote

    *snip*

    So W8.1 is just a service pack? Seems to me Microsoft packed in a whole lot more than that. I thought Microsoft was trying to get people to take another look at Windows 8. Kind of like the idea that "We fixed what you didn't like about Windows 8 and made it even better." IMO W8.1 is an opportunity for Microsoft to get people to take that second look. They won't get that with OEMs shipping W8 devices. With Ballmer admitting that they need to do better with W8 sales I'd think framing it as a service pack would be contrary to that effort.

    It depends what you mean by service pack. It's about as substantial as WinXP SP3 was, which is more than all service packs since, and it's a genuinely new build of everything from the kernel down. But from a users perspective, only a few cosmetic changes occur (like the reemergence of the start menu, enhanced and unified search from the start screen and new apps) but the rest of the OS fundamentally looks the same

  • User profile image
    blowdart

    , evildictait​or wrote

    the reemergence of the start menu

    Button. Not menu.

  • User profile image
    ScanIAm

    , bondsbw wrote

    @ScanIAm:  In computer architecture class, I remember learning a bit about the history of Intel's 8086 architecture and how it was in competition with MIPS and their RISC processor.

    You know how I know that I'm old?  Cuz I remember this happening in real life.

    Intel knew that x86 was an inferior technology, but they were ahead of MIPS by a few months in terms of Moore's Law.  Intel had the chance to move to a RISC architecture, but by the time Intel implemented one that could compete with MIPS, they would be behind.  So Intel scrapped the idea and stayed the course with x86.

    x86 was inferior to RISC, but Intel has a canyon full of cash for their decision that first to market was better than going with a superior technology.

    That sounds a lot like revisionist history.  Intel was getting beaten senseless by AMD's competing and compatible x86 chip at the time, and while RISC was theoretically a better tech, it would also have thrown away all of the apps written for x86 (specifically, all apps in the MS-DOS and emerging windows ecosystem). 

    If you look at the biggie companies in the tech industry, I can guarantee that you will find very little tech that was unique and novel to them.  They did it better, not first. 

    And I'm talking about Google, Apple, Microsoft, and Intel.

    Edit: adding missing word.

  • User profile image
    DeathBy​VisualStudio

    , evildictait​or wrote

    *snip*

    It depends what you mean by service pack. It's about as substantial as WinXP SP3 was, which is more than all service packs since, and it's a genuinely new build of everything from the kernel down. But from a users perspective, only a few cosmetic changes occur (like the reemergence of the start menu, enhanced and unified search from the start screen and new apps) but the rest of the OS fundamentally looks the same

    I dunno about a "only a few cosmetic changes". This list (bottom half of the article) seems to be pretty substantial. To a geek I would agree that the changes to the OS fundamentally looks the same.

    If we all believed in unicorns and fairies the world would be a better place.
    Last modified
  • User profile image
    bondsbw

    , ScanIAm wrote

    That sounds a lot like revisionist history.  Intel was getting beaten senseless by AMD's competing and compatible x86 chip at the time, and while RISC was theoretically a better tech, it would also have thrown away all of the apps written for x86 (specifically, all apps in the MS-DOS and emerging windows ecosystem).

    I think I'm talking about earlier than AMD's involvement.  I'm too young to know from experience, so if I need to get off your yard just get out the water hose... but I'm talking more about the 8088 and 8086 chips introduced in the late 70s, and AMD wasn't really involved until around the 386 in the mid-80s.  And since these were the first x86 chips, there weren't apps written for them yet.

  • User profile image
    Jim Young

    , bondsbw wrote

    @ScanIAm:

    x86 was inferior to RISC, but Intel has a canyon full of cash for their decision that first to market was better than going with a superior technology.

     

    Ah, the old RISC vs CISC debate. Fanboyism from days of old.

    Each architecture has its advantages and disadvantages. One of the major reasons that CISC won out is that it is much easier to write compilers for it. RISC had their chance with Apple and the PowerPC systems. In the end the Intel won out because of its superior power management and near complete penetration in the desktop and laptop markets.

  • User profile image
    bondsbw

    @Jim Young: Hey, at least that fanboyism surrounded philosophy and not companies.

  • User profile image
    Jim Young

    , bondsbw wrote

    @Jim Young: Hey, at least that fanboyism surrounded philosophy and not companies.

    Eventually it did become company related. I remember back when Macs were running on the PowerPC CPUs and there were huge flame wars about how PowerPC was soooo superior to the unwashed masses using the x86. When rumors started circulating that Apple was going to abandon PowerPC there was a major freak out in the Mac community.

  • User profile image
    evildictait​or

    , Jim Young wrote

    One of the major reasons that CISC won out is that it is much easier to write compilers for it..

    What?! No! Quite the opposite! Compilers compile your code to an internal representation that looks a lot like RISC anyway, because it is much easier to pattern match on infinite register RISC-like instruction sequences than deal with the complexities of CISC. Consequently writing a RISC outputting compiler is a fairly trivial conversion of internal IL to RISC output, whereas outputting good CISC is itself a hard challenge.

    The big thing that CISC always had going for it was the fact that the output binary can be made a lot smaller than equivalent RISC, and also you have more information density going through the instruction pipeline.

    The big thing RISC traditionally had was that decoding a RISC instruction is cheap, and so RISC processors could be made to be very low power, and the compilers where cheaper to make.

  • User profile image
    Jim Young

    , evildictait​or wrote

    *snip*

    What?! No! Quite the opposite! Compilers compile your code to an internal representation that looks a lot like RISC anyway, because it is much easier to pattern match on infinite register RISC-like instruction sequences than deal with the complexities of CISC. Consequently writing a RISC outputting compiler is a fairly trivial conversion of internal IL to RISC output, whereas outputting good CISC is itself a hard challenge.

    I always assumed that compilers for CISC were easier to write and have better performance because the instruction set was much more high level than RISC. For example multiplication is a single instruction in x86 while the RISC approach takes four.

  • User profile image
    bondsbw

    Compared with lexical analysis, parsing, semantic checking, and of course doing all of this inside tooling, and then optimization, I've considered the emit phase of a compiler to be fairly straightforward.  Unless there are significant advantages in the area of optimization, I wouldn't expect building a compiler to be affected largely by the choice of RISC or CISC.

    And I can't consider Intel vs. PowerPC to show the advantages or disadvantages of either platform.  I'm sure that if Apple chose Intel and Microsoft/IBM chose PowerPC, we would all be using PowerPC computers right now.  We might even be using PowerPC phones and tablets instead of ARM.

Conversation locked

This conversation has been locked by the site admins. No new comments can be made.