Recommended Posts

How should they start building Windows 8?

Should they build off of previous OS releases and their code, or should they start from ABSOLUTE scratch and build everything from the registery to the calculator from nothing?

I can see benifiets on both sides, where you can cut time and resoucres starting on top of other OS's, or start fresh and build the bare nessities. I can only imagine they have improved their code writing and how they organize stuff.

Link to comment
https://www.neowin.net/forum/topic/828104-starting-windows-8/
Share on other sites

Win8 is already in development, They don't wait a few months or half a year after the previus OS before starting the next.

They start the next before the last one is done. And you won't see a full rewrite for a very long time, there's no need either. And major parts of the OS was already rewritten with Vista. And it took 1-2 years for the drivers vendors to catch up. which caused the media and anti people to go on a crusade about how vista sucks, yet somehow 7 which is for all intents and purposes, as far as stability, performance and compatibility go(the point the complained about Vista) identical is awesome (and yes 7 is awesome, but so was vista anyway)

it is like

Windows 2000 > Windows XP > Windows 2003 > Windows Vista/WS 2008 > Windows 7/WS 2008R2 > Windows 8/Windows server 2012(future release)

so nope , they aren't starting from scratch anytime soon . it will cost them billions and not mentioning the huge risk of failure for such a big project and in the end they render the whole previous platform software wont work in the new one

it is like

Windows 2000 > Windows XP > Windows 2003 > Windows Vista/WS 2008 > Windows 7/WS 2008R2 > Windows 8/Windows server 2012(future release)

so nope , they aren't starting from scratch anytime soon . it will cost them billions and not mentioning the huge risk of failure for such a big project and in the end they render the whole previous platform software wont work in the new one

I think the risk is minimal. Microsoft has experience so they wouldn't be really starting from scratch with ideas of how to go about doing things. I'm sure a new experience would allow more Plug and Play approach to hardware. Like Apple for instance. I'm just shooting out ideas and wondering why. I don't have any background.

I just think it would be go to "trim the fat." Vista and 7's footprint on hard drive space is huge. I think they are taking that for granted.

Thread moved to: Windows Support ? Windows Beta Discussion>

I agree with The_Decryptor. I really doubt they will ever do a full rewrite, but instead, they will just continue to re-write different parts with each release. I mean, in Windows 7, the Calculator application was re-written and came with some very useful improvements; it'd be silly for them to re-write that again for Windows 8:pp

I think the risk is minimal. Microsoft has experience so they wouldn't be really starting from scratch with ideas of how to go about doing things. I'm sure a new experience would allow more Plug and Play approach to hardware. Like Apple for instance. I'm just shooting out ideas and wondering why. I don't have any background.

I just think it would be go to "trim the fat." Vista and 7's footprint on hard drive space is huge. I think they are taking that for granted.

Apple can do what they did with OS X because at the time they had a negligible marketshare, so the upset wouldn't be that great. Not to mention they shipped OS X with (essentially) an entire other OS, for backwards compatibility reasons. Microsoft, having the greatest marketshare, is not afforded such luxuries, nor the ability to bully their users into something so drastic as what Apple did.

Really? In the day and age of the 1tb hard disk. You can argue that the software itself is bloated. The foot print really doesn't matter. Not even if you a newer generation Netbook. As for plug and play, it has been there since Windows 95 (and has worked well since Windows 2000). I think a reboot is unnecessary because the core fundamentals that were laid with Vista work well since.

Windows 7 is not bad. Neither is the server line. I am not really seeing a problem. It is far from perfect, but than again, so is Apple. Neither is exactly "well" designed on a theoretical level. Nor is Linux. There are certain concessions that vendors make when releasing a product for the mass market. This, however, doesn't mean that Microsoft is not working on a number of different research projects to extend Windows well into the next two decades. Windows 8, however, will be "probably" very much based on a Vista/7 model.

Really? In the day and age of the 1tb hard disk. You can argue that the software itself is bloated. The foot print really doesn't matter. Not even if you a newer generation Netbook. As for plug and play, it has been there since Windows 95 (and has worked well since Windows 2000). I think a reboot is unnecessary because the core fundamentals that were laid with Vista work well since.

Windows 7 is not bad. Neither is the server line. I am not really seeing a problem. It is far from perfect, but than again, so is Apple. Neither is exactly "well" designed on a theoretical level. Nor is Linux. There are certain concessions that vendors make when releasing a product for the mass market. This, however, doesn't mean that Microsoft is not working on a number of different research projects to extend Windows well into the next two decades. Windows 8, however, will be "probably" very much based on a Vista/7 model.

I don't want you to mis understand me. I don't think Windows 7 is bad. I feel it it the best OS since Vista. My point is, since they have a good OS out, they can afford to wait a little bit and build an OS from scratch. Obvioulsy nothing is perfect, but now is the time to build a secure OS with the things they have learned. It is time to put all the Linux boys and Unix boys to shame with a far superior product that is based on none of the above. I'm just dreaming i guess.

And yes, the footprint does matter. A streamlined OS will allow for better component talk and quicker/more stable fluency.

People say they doubt they'll start from scratch but there is a channel 9 video from a few months ago of their developers discussing the possibility of doing just that.

Do you have a link to that video, that sounds interesting.

Vista and Win 7 are already very secure, re-writing it would just make it less secure (chance of re-creating bugs you've already fixed)

Never mind that Windows has a huge library of programs built around their API's, a single change could break a lot of them (look how many apps stuffed up when UAC was brought in)

With the change in the way Microsoft develops Windows introduced in the recent OS' (Main code tree only contains complete features, modular branches for features and components) and the increased independence of core components and components at all (MinWin and the various changes Vista and 7 introduced), I'm not sure they even need a rewrite anymore.

You're pretty much at the place where you can just pick out problematic/outdated components and rewrite them alone.

What IS likely is simply throwing out old components that are not needed - 16bit went that way, And perhaps 32bit support systems (and the entire WOW64 system, hopefully) will get removed by the next OS' revision.

I would just like to point out that Windows has never been 're-written' as such.

7 was based on Vista, with more of a clean up than major changes.

Vista was based on XP, with major changes but no complete re-write.

XP was based on 2000, with big but not major changes and again no complete re-write.

2000 was based on NT4 with major changes, but no re-write.

NT4 was based on NT3.5 with major changes, but no re-write.

You can see where this is going. Even the 95/98/ME architecture was never re-written, it just got cleaned up and had major changes.

If Microsoft ever did want to re-write Windows from scratch it would probably take at least 10 years from original conception to a finished product. At which point, all of their 'original ideas' would have been tacked on to a competitors product because they had the idea on their own. Even Apple didn't do a re-write with OSX, even though they would like people to believe it. They lifted the core of Linux BSD, and built around it.

Absolutely NO-ONE rewrites anything anymore because it just takes too long. If they've got a product line already, they'll just improve and clean up the code and then add a few new features.

I don't think they will re write the OS anytime soon. But it will be done at some point. When the time comes I can see Microsoft including backwards compatibility in the form on a sandbox or VM. The Sandbox and VM will emulate the registry and let all the old apps function like they should. Not to mention the work Microsoft is doing with XP mode.

Never say never. I wouldn't be surprised if Windows is re-written, while adding compatibility for older software by means of virtualization.

It amazes me how well you can run certain Windows applications in a non-Windows environment like Crossover for Mac or Wine. I don't know how they do it but it's an incredible feat IMO. So, it may very well be possible to implement something similar in a new version of Windows.

With the change in the way Microsoft develops Windows introduced in the recent OS' (Main code tree only contains complete features, modular branches for features and components) and the increased independence of core components and components at all (MinWin and the various changes Vista and 7 introduced), I'm not sure they even need a rewrite anymore.

Well, "rewrite" doesn't make sense because it implies that you're still following the original specifications. What Windows needs is new features, not changes or rewrites of old ones. Half of the Windows APIs have gotten so complex over the years and have so many "gotchas" that it's borderline impossible for a developer to get things right. Other parts are just dated. The graphics, windowing, and user interface APIs are an example. These are not salvageable at this point. They have to be replaced with something brand new. Something modern and simple. All of these "core" components were designed in the 1980s when the requirements were very different.

You're pretty much at the place where you can just pick out problematic/outdated components and rewrite them alone.

What IS likely is simply throwing out old components that are not needed - 16bit went that way, And perhaps 32bit support systems (and the entire WOW64 system, hopefully) will get removed by the next OS' revision.

You can't really. The APIs are set in stone and can't really be changed, and in many cases you have to be extremely careful about changing the implementation as well, as someone will inevitably have a dependency on it.

Even Apple didn't do a re-write with OSX, even though they would like people to believe it. They lifted the core of Linux BSD, and built around it.

Linux? What is OS X today predates Linux. I don't think people realize just how old it is. What you call "the core of Linux BSD" is actually the Mach kernel (which dates back to the mid-80s) and various components from FreeBSD. This hybrid kernel is called "XNU" and was created back in the 80s by the company Steve Jobs founded after leaving Apple. They also created an OS to sit on top of it, known as NEXTSTEP. When Jobs eventually rejoined Apple in the late 90s, he brought this OS with him, and it was heavily modified to become "OS X."

And NT3.5 was based on?

The NT architecture was written by Microsoft for the business market as computers took off more in the work place. It was designed to be more stable and robust than anything else they had and to also provide a server platform. Yes, the NT architecture was new, but that was back in the day when a full install only took a few hundred megabytes at the absolute most, and it was feasible to create something from scratch.

My point I was making before is no-one re-writes their software in todays market because it would take too much time. I believe when Vista was first planned out as Longhorn, that was a planned re-write and it took them years before they eventually abandoned most of the ground breaking ideas they had because it was simply taking too long to implement. I do think we will see all of the ideas they had planned for Longhorn but bringing it all out in one release was too much, and Microsoft should have realised it before they even started writing the first line of code.

So basically we won't see a re-write of Windows until we advance beyond our current hardware platform. And that will only be out of necessity.

It was the first version of Windows NT

The GUI was based on the model set by Windows 3 though

3.5 was based on 3.1, which was the first version of NT. NT itself took some inspiration from VMS, but was a new OS. It was just the core though, not a complete OS. They then ported the various Windows APIs from the 16-bit version of Windows, so in reality much of what people think of as "Windows" wasn't actually new in NT -- and it shows.

The GUI was based on the model set by Windows 3 though

So the claim it's never been re-written is no really true.

But, if MS ever did decide to re-write I think the only way they could do it is under another brand (not "Windows"), and offer a choice during a transitional phase.

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • Zero tolerance for antisemitic social media posts. Thousands of arrests and fines.
    • It's not about the kids, it's about de-anonymizing the entire internet to punish people for wrongthink. The only way to ban kids, is to demand ID from everyone, a digital ID if you will.
    • QuickView 6.8.1 by Razvan Serea QuickView is the fastest image viewer for Windows, designed to open all your photos instantly. It supports popular formats like WebP, AVIF, JPEG XL, PNG, JPEG, TIFF, RAW, and PSD, making it perfect for photographers, designers, and everyday users. With lightning-fast load times and zero-lag previews, QuickView handles everything from small icons to massive 8K images effortlessly. Drag and drop files to view them instantly, zoom and pan smoothly, and enjoy a clutter-free interface built for speed and simplicity. QuickView also makes managing and analyzing images easy. You can preview thousands of photos instantly, view real-time color and brightness, and check image details without slowing down your computer. It automatically fixes common file issues and works perfectly offline, so your images stay private. QuickView supports multiple languages, is portable, and requires no installation. QuickView key features: Blazing Fast Loading – Open images instantly with zero lag. Modern Format Support – View WebP, AVIF, JPEG XL, and more. RAW File Ready – Handle photos from all major cameras effortlessly. Classic Format Friendly – Supports PSD, PNG, JPEG, TIFF, and BMP. Drag-and-Drop Convenience – Open files instantly without menus. Multilingual Interface – Works in English, Chinese, Japanese, German, Spanish, and Russian. Portable & Lightweight – Single executable, no installation required. Mass Image Preview – Instantly view thousands of images with HUD Photo Wall. Real-Time Color Tools – RGB histograms and color analysis overlays. Accurate Metadata – View EXIF and file information instantly. Smart File Fixes – Automatically repair incorrectly saved files. Fast or Full-Quality Toggle – One-click RAW preview adjustment. Smooth Navigation – Zoom, pan, and scroll without slowdowns. Privacy Focused – Fully offline operation keeps your images secure. QuickView 6.8.1 release notes: Dynamic Island, Filmstrip Gallery, Custom Hotkeys & Size Optimization QuickView v6.8.1 introduces redesigned window controls, an interactive filmstrip gallery, dual-mode slideshows, fully customizable hotkeys, and substantial binary size optimizations. Changelog: Floating 'Dynamic Island' Window Controls The window controls have been redesigned for a cleaner interface. Floating Capsule Pill (#199): Replaces traditional window controls with a floating pill-shaped widget in the top-right corner. Includes hover glow transitions. Compact Size: Reduced the size of caption buttons to maximize screen space for images. Interactive Filmstrip Gallery The filmstrip gallery has been redesigned with improved controls and animations. Top-Hover Trigger: Hover near the top edge to expand the filmstrip gallery. The trigger mode can be set to Hover, Pinned, or Disabled via the Settings menu. Auto-Centering Scroll: Selecting a thumbnail triggers a smooth scrolling animation that aligns the item to the center of the bar. Visual Refinements: Corrected visual gaps when pinned, fixed zoom anchor offsets, and restored smooth horizontal auto-scrolling. Dual-Mode Slideshow Spotlight Mode: Added a slideshow mode inspired by Picasa Spotlight, which dims the background and focuses on the active image. Normal Mode: Standard fullscreen slideshow functionality. Fully Customizable Hotkeys Custom Keyboard Mapping: Added support for completely customizing and rebinding all core keyboard shortcuts and navigation hotkeys directly within the Settings menu. UI/UX Adjustments & Window Snapping Magnetic Snapping (#90): Window borders now snap to screen edges (100% magnetic snap) when resized. Responsive Toolbar (#194): Toolbar buttons automatically hide based on the window width and active mode. Timeline Scrubbing: Implemented debounced asynchronous seeking for animated formats, providing smooth scrubbing without timeline lag. Fixed frame count and distortion issues on large GIF seeking (#197). Extended Mouse Mapping (#191): Added support for mapping multi-function mouse side buttons in settings. Archive Sorting (#193): Added an option to always sort archives by name ascending. Footprint Compression & Size Optimizations Reduced the binary size of the standalone executable by removing redundant templates and dependencies. C++ Stream Elimination: Removed dependencies, saving approximately 18.5 KB. Localization Deduplication: Consolidated localization string tables, saving 10.5 KB by preventing template duplication. Vector Icon Compression: Compressed static vector icon coordinates to 16-bit integers, saving 54 KB. Code Devirtualization: Replaced std::function callbacks with C-style function pointers and devirtualized core controllers to reduce overhead. LTO Debug Fix: Removed obsolete compiler flags (/MERGE:.rdata=.text) from Link-Time Optimization (LTO) builds to fix minidump crash debugging. Decoding & Memory Fixes Hybrid Allocation: Implemented a hybrid memory allocator to balance preloading and tile rendering. Access Violations: Fixed crashes when rapidly switching images. HDR in Archives: Resolved an issue where HDR images decoded from ZIP/RAR archives lost their peak luminance metadata or failed to render in float format. WebP, AVIF & JXL (#195): Fixed shadow transparency glitches in WebP/AVIF and image distortion in JPEG XL. Download: QuickView 64-bit | Portable 64-bit | ~5.0 MB (Open Source) Download: QuickView ARM64 | Portable ARM64 View: QuickView Website | Github | Screenshot Get alerted to all of our Software updates on Twitter at @NeowinSoftware
    • There was a Spider-Man game back in the day that had these as well. When other people played, they got gamer ads. When I played, I got ads for Visual Studio on the billboards in the city.
  • Recent Achievements

    • Conversation Starter
      NovaEdgeX earned a badge
      Conversation Starter
    • One Year In
      Console General earned a badge
      One Year In
    • One Year In
      Twozo Technologies earned a badge
      One Year In
    • One Month Later
      Twozo Technologies earned a badge
      One Month Later
    • Week One Done
      Twozo Technologies earned a badge
      Week One Done
  • Popular Contributors

    1. 1
      +primortal
      521
    2. 2
      +Edouard
      186
    3. 3
      PsYcHoKiLLa
      107
    4. 4
      Steven P.
      88
    5. 5
      ATLien_0
      67
  • Tell a friend

    Love Neowin? Tell a friend!