What was the purpose of your XP machine at the time and was this at home or office?

It sounds more like you are some sort of web developer or at least someone in the IT department making sure no breaches occur. If that's the case I understand why you don't want to use it.

I already know about Google Chrome dropping XP support but like I said they have their own goals and it was never for the consumer but for themselves.

What version of Opera are you referring to using Chrome Blink engine?

I'm using version 10.00 but again I only use it to visit certain sites and I keep multiple browsers open so I do switch back and forth between them.

I think we shall see what happens in the upcoming years and if it truly looks like you cannot use any internet browser on XP I have no problem going to Vista and continuing there at that time.

I'm not doing anything overly complex that someone would get boggled by just yet in order to keep an older OS working on the internet.

Windows XP Professional SP3 and just running Firefox.

Who knows in 2021 Microsoft might come out with XP Ultimate 128-bit edition and shock us with the same classic interface capable of running DirectX 15.0. I'd instantly switch if that was the case. It would probably support USB 4.0 so it'd be a worthy upgrade over USB 2.0.

I think the biggest obstacle for 64-bit Windows was backward compatibility with 32-bit Windows software was not 100% which meant buying a lot of 64-bit based equivalent version of the same software they already bought.

Again I'm going to stress I'm not a big fan of Google Chrome. It's bloated for one and uses more memory than Firefox. It keeps "GoogleUpdate.exe" in the background process if you check Windows Task Manager even when its not running. Possible data mining going on sending info to Google about your OS, websites you go to, cookie spying perhaps? Who knows. Look up "scroogle" and you will see there was once an alternative to google searching without being spied on. Google Chrome was made for those cheap Google Chromebooks to compete and kill off Sony Vaio's ultralite laptops. Vaio had to get out of that business. They made some of the nicest looking slimmest ultralite laptops weighing 2 pounds or less.

As for the comment about them won't letting go of using IE6-IE8 on XP. I think those people already moved on a long time ago when websites started to not load properly. They probably got infected with malware and the computer seemed sluggish or no longer booting up. They ended up taking it to a computer repair place like Geek Squad and got quoted $200 to look at it or buy a new laptop with Windows 7 or later for $300. Those guys are gone man. Only resourceful people who don't want to spend any money and have 32-bit applications they like are probably running Firefox because they want to use the internet until it no longer works.

I just happen to be the opposite putting XP on modern hardware because I liked the OS enough and wanted to see how much more performance I got out of it. Mind you I upgraded from a P4 3.06 GHz Unicore to a Ivy Bridge 3.4 GHz Quadcore so I was expecting a big performance boost after holding out from upgrading for so long.

Passmark puts the P4 at 348 which was top of the line for a Unicore back then.
Passmark puts the i5-3570K at 7151.

This put it at 20.55 or roughly 21 times better performance.

However I downclocked mine down to 2.2GHz and it runs on just 35 Watts of power out of the socket.
The P4 3.06GHz Unicore sucked 120 Watts.

So let's assume the appropriate passmark score is 4627 / 328 = 13.3 times faster.

So you can imagine what a tremendous potential CPU boost it should be from the P4 to Ivy Bridge for XP.

As for Firefox vs Internet Explorer 6.0-8.0 differences. It's actually quite similar. If you've used Internet Explorer then you can use Firefox quite easily. I recommend Firefox version 22.0. This is the LAST version that still allows you to enable or disable javascript in the options. A lot of times websites slow down the entire computer because of javascript and a simple OFF ON of Javascript fixes the CPU load lag. I actually use a Firefox Addon to Toggle Javascript ON OFF on the fly instead of going through the Menu options.

At one point I actually preferred Opera at first because it runs faster than Firefox and has more options like saving pictures despite websites forbidding it. But Firefox seems to have the most support on most websites including banks, Amazon, eBay, Paypal, et cetera so I use it for my email and whatever websites require a more up to date browser.

I think you are looking at it all wrong. I'm not expecting some company that decides to release hardware let's say a new USB Logitech webcam to purposely make a XP driver. However I expect it to be USB compliant at least and as you know even if hardware that comes out from a manufacturer has no driver support for it people are out there who hack the thing and reverse engineer the code to get it to work on their operating system. I've seen this done more on Linux.

The only two companies who are actively trying to kill XP would be Intel and Microsoft by dropping support of drivers which they know how to make and it would only constitute very little effort to release. I'm not saying all companies are targetting XP as why should they because it doesn't hurt their profits if they already have the experience and code for making previous XP products whereas if it was a completely new OS I understand that's going to strain resources even further and require extra manpower and funding. However most companies who have been making products run on XP wouldn't be reinventing the wheel just to make a product work on XP despite say if the amount of users is in the millions would seem pointless. But if the effort is only like 5% or 10% to create an XP driver and 80% to create a Windows 10 driver why would it really be a huge negative. If anything saying your product supports more operating systems is an advantage and I'm not saying this should support something unusable like Windows 98SE or DOS or Windows 3.1. XP is still useable today and anyone can find a cheap used laptop on eBay with XP for $30 bucks and hook up a USB webcam and they got a surveillance camera ready to go on the cheap.

As for why I brought up IE it was your statement that inferred using IE with XP.

Since you do use Firefox than you probably understand IE and Firefox are quite similar. At least when I used IE back in the day I think I never used anything above IE 6.0 because the higher the version number the more bloated it got and slower. I think at one point I liked IE 4.0 on Windows 95B the most because it was the smallest and most compact browser without slowing down my computer and it incorporated the famous Quick Launch user interface.
Thanks for the explanation.

It seems like a lot more work to quote within quotes. It would be easier if it was WYSIWYG.

But aside from that the other limitation is how much text you can put in the response. I've had to break my messages down in chunks and edit and add more chunk until I paste it all. For now this seems to work the fastest although I admire your use of in quote usage skill. :)

Take the original post, and put it in notepad. Notice the starting quote tag.

Once you find the part you want to answer to, isolate it, add the starting quote tag in front of it, and end the tag at the end of the fragment. Write your answer to it. Continue until your answers are done.

If you want the quote to contain a quote, add the relevant quote block inside the quote block

What were your system specs for the XP system before your retired the OS?
"
Reason being that I put security above all else so running obsolete OS is not something I'm personally willing to do. Linux is a winner in that situation in my eyes, although it may not be for everyone quite yet - but I do think it will get there and it is certainly worth getting familiar with if someone hasn't done so."

Okay this definitely explains your priorities although XP SP3 was released around 2009 and I have to assume the only reason you kept lugging around XP was the mini patches Microsoft sporadically released now and then kept you on board...

However how about running Linux WineHQ and XP inside it? Would this not be secure?

Obviously this is a joke. The driver support for those two OSs are just not there.

It would be more interesting if there was a GOG Client for pure DOS but unfortunately that would only work for DOS games only which probably wouldn't be worth the effort since DOSBOX exists.

Plus what Windows 3.11 and 3.51 games can you think of even worth playing?

Most of the games made for Windows 3.1X and NT just really were sad. I think DOS despite being older still was a better gaming platform until 98SE starting to catch steam.
I haven't used ReactOS but wouldn't WineHQ make a better choice to run Windows inside it and be protected by a Linux exterior?
What we really need is Apple ][, Commodore 64, and Atari 400 support.

I know these platforms seem dated but at least give them a GOG Galaxy Client chat support so they can at least talk to someone who can play games even if they can't join in on the fun.
I think he meant GOG Galaxy (Final) supported on XP.

I doubt he meant every new game released had to work with XP because unless the game isn't using DX 10 or higher than it probably won't happen.

But any GOG game that currently has XP support and has network play should be able to play with other GOG Galaxy players regardless if they are on XP, Vista, 7, 10, MAC OS X, or Linux as long as the game supports network play on that particular OS.

Please refrain from quoting posts from several months ago. I don't want to get notifications for some post I made months ago that a thread necro person wants to restart a discussion on.

There's no discussion it was a minor interpretation of what you thought. I didn't know there was a deadline. There you can go to sleep thread necro hater. Couldn't you just disable reply notifications from this forum on your own so you won't get any more notifications from this forum? I don't want to accidentally reply to something that has you in the thread.

This is something I'll probably have to play around with but it would be more interesting to see instead of quoting they use different colors to separate the user's text.

If I were to try to do multiple embedded quotes it would be a nightmare since I can only paste so much text after each edit and note if it was correctly updated before going back to edit and pasting more into the body.

However for very short one line comments I think it would be worth testing out.

I'm not sure where the exact limit is but if you hit "Post my message" and it doesn't post right away then the limit is reached. I remember encountering this when I first used this messaging system and waiting quite some time wondering why hasn't the post been sent. So far my technique has been the best way around it.

I tried reading one of your posts, but I nearly fell into a coma. Which is probably why my post wasn't in reply to anything you wrote.

Cheeky Blinkered Codswallop!

Return to your alcove Syndicate Governor!

Pip pip Cheerio.

There is no single answer, as it would vary greatly from one piece of hardware to another, and one driver to another. But in general it is a lot more than one might think it is. This is particularly true with drivers where a single change for one piece of hardware in a path that handles many hardware devices can cause problems on all the other hardware supported by the same driver. As a former video driver engineer I know this all too well. It becomes increasingly time consuming over time to both add support for new hardware and new hardware features across a number of operating systems while maintaining support for legacy hardware/OSs that an ever diminishing number of users are using. In some cases it puts a real road block on what changes you can make and eventually you have to make decisions about whether to get rid of support for older hardware, fork the codebase or other choices for example. Every single company out there has to make these type of decisions, and almost universally at some point old hardware and old operating systems support will get lopped off because it is no longer viable to pour the resources into maintaining.

Remember, no company has infinite resources/manpower in the end, and companies (or even open source projects/non-profits) will generally lean heavily towards trying to make the best use of their resources and manpower available. In many cases if just leaving some code alone that already works fine as-is on a piece of hardware, an OS, etc. then it may be left alone for example, but if/when it demands more engineering resources than the company/developer deems is reasonable for the perceived number (or percentage) of people using it out there then they may make a reasoned judgment call to no longer continue supporting the old OS or hardware as the case may be.

I myself have been in a position where I had to make this kind of judgment call. We had drivers for a multitude of video hardware, and as the years went on the video hardware marketplace solidified around 3 primary video hardware vendors. In terms of volume of hardware units available on the market, Intel was and is the largest vendor of video hardware out there due to the huge market for onboard/embedded GPUs they dominate, then there are ATI/AMD and nVidia which dominate performance graphics for more than a decade and a half now. Any other vendors that produced new hardware still at the time made up an extremely small percentage combined, and a great many of the other video hardware vendors simply left the market more or less so the drivers for their hardware were simply kept around for continued legacy compatibility. Several times a year a number of major architectural changes occurred in the video subsystem which required changes to be made to individual video drivers (and input drivers as well) which someone would have to do in order to keep those drivers functioning on the new system. Our hardware database showed that only a fraction of a percent of customers were using video hardware that was not from Intel/AMD/nVidia, and in most cases this legacy video hardware was 5-10 years old and well outside of our target marketplace. In some cases where it was a simple change to make, drivers would get updated and any customers that had indicated compatibility problems were asked to test the new drivers and we may have decided to ship them as-is just as a courtesy convenience, but we did not actively support them in terms of "we will dedicate engineer manpower to fixing problems with this but we hope it works for you anyway."

If a driver became so broken or unwieldy that it would require dedicating real manpower to fix, then we would generally decide to disable the non-functional driver in the build. We simply did not have infinite manpower to throw into maintaining ancient hardware support like that at the expense of product development geared towards the overwhelming majority of our actual customers and users. It would have been irresponsible in 2005 for example to be spending a week of engineering time updating a video driver for Trident 8900 video hardware from 1994.

This is just a small personal example, but every company out there will have their own finite resources and their own list of tier 1 operating systems and hardware to support, with the majority of engineering resources going to what the majority of their customers are using. It's not really about whether it is theoretically possible to support some old OS or hardware device, it's more about whether it is economically feasible best allocation of finite engineering/research and development/quality control/technical support man power and the money that gets spent on paying the wages and necessary infrastructure to do that.


XP wont vanish, not as an absolute. People out there still use MSDOS including many of us via DOSbox packaged games for example, but I was at the bank a few weeks ago and they were having problems with their system and the clerk mentioned that it ran on MSDOS. Old software and hardware is never completely 100% eradicated, but that's not particularly relevant of course either. There is no one number that one can state as being insignificant outright without qualification, and there is no right or wrong answer either. It boils down to how much budget a company has to allocate manpower and other resources, what they wish to accomplish with those resources and then considering how to spend those resources on the biggest bang for the buck considering a great many factors which may be very unique from one company/project to the next, one product to the next, and one month to the next. One company may have the resources to spend on supporting their product on an older OS for many years, perhaps that product just works and does not really need any or many resources/manpower assigned to it at all, so they keep advertising it as supported. Another company with a different product may discover that continuing to support Windows XP for example means they are going to have to dedicate 2 of their 10 engineers working full time on it at a cost of $60,000-120,000 USD per year on an ongoing basis and decide that spending 20% of their engineering resources to support this legacy operating system is not going to drive 20% of their revenue so it makes no economic sense to do so. Instead, lopping off XP support means those 2 engineers that make up 20% of their team can now work on new features for their product that they perceive will benefit 98% of the existing users and cashflow.

Just a theoretical example, of which there are an infinite number that one could create to demonstrate that it varies greatly and every company and project needs to determine these things individually regardless of what some other company or project might be doing because everyone's circumstances are different.


Yes, but Microsoft is not coming out with SP4 for Windows XP, the product entered end-of-life in April 2014. :) Everyone decides to upgrade (or not to) for different very personal reasons, none of which are outright right or wrong, but are right or wrong for each on an individual basis.

I'd like to share some thoughts about 64bit Windows also since you mentioned it, but I'll write a separate post to generalize it as I think I've got some useful information I can share that might benefit others on understanding a bit more about why a 64bit OS and applications are almost always superior for performance. There are a lot of misconceptions about this out there, so I'll whip up a new post on the topic which others might find useful.

When deciding to create their gaming client, like any company that would consider doing this no doubt GOG explored all of the available options out there and not just Google's embedded Chromium engine. To undertake such a large project without considering all possible approaches would be short-sighted to say the least.

Both Google and Firefox's respective codebases are used by a number of 3rd party projects out there so there is precedence for using both. If however one does research on the pros and cons of each, they're likely to find that trying to embed Firefox's gecko engine is somewhat of an ever changing nightmare of sorts. The engine changes significantly with each release of the web browser, and while it's made available for other software to use it has turned out to be more of a gigantic headache for 3rd party projects than anything. Mozilla is aware of that and it has been a bit of a thorn in their side for some time now. Firefox the browser is an amazing piece of technology, and I'd almost rather not have a web browser at all if I couldn't use Firefox. But the way that it is all put together is not specifically designed cleanly to be used by 3rd parties in the sanest manner.

This is particularly true even for other Mozilla born projects that use xulrunner/gecko internally such as Mozilla Thunderbird, Seamonkey, and in the past projects like Songbird which is now discontinued. Other 3rd party projects like Miro have had to deal with this constantly moving target as well. In the end these projects waste a lot of resources simply trying to keep up with the ever moving ever changing Mozilla codebase. In the case of Mozilla Thunderbird, Mozilla refers to this as a "tax on their resources", and it is why for example Firefox is at version 43 right now and Thunderbird is at version 38.x, because they just can not keep up with the rapid rate of change of the entire Firefox codebase. The Thunderbird project for years went from struggling to keep up with each Firefox release to just sticking with one release and updating it with security fixes and major bug fixes, and then trying to rebase to a newer codebase from time to time, but the amount of resources spent doing that proved to be too much for the amount of manpower they had available and ultimately put resources on trying to keep up with Firefox rather than actually developing new features and functionality in Thunderbird itself.

Fast forward to today and Mozilla is abandoning the Thunderbird project and not investing any more resources in it, leaving the project to fend for itself. As an avid Firefox and Thunderbird user I am both saddened by this, but also in a way happy about it too. For over a decade Mozilla's approach to do everything themselves may have served them well, but it is not without cost, and now they're feeling some of the cost of that today more than ever. In response they're "trimming the fat" as it were, and they are redesigning the entire browser as we speak. The new browser that will be released in a year or more (likely) is designed with their Rust programming language, and their new multi-process enabled engine (Electrolysis/Servo or whatever it's called this week). This is a dramatic departure from the past and they're moving forward and leaving the old legacy design behind. This is well known for anyone that follows Firefox development, which anyone considering using a web browser engine as a basis for a new project would likely be aware of, and which would have a definite impact on the decision making process involved in evaluating browser technologies for use in new projects.

So when people see that Steam, Opera web browser, GOG Galaxy and a variety of other projects and commercial products have chosen to use the Google embedded Chromium engine as the basis for their own web browser based technology, it is not a decision made by the flip of a coin. While the Mozilla Firefox browser is a fantastic piece of technology, it was never designed specifically with focus on it being an easy solution to use for 3rd party embedded projects even though they have made it able to be used for such. In simpler terms: "Thar be dragons in thar."

Aside from that however, you've suggested that they could have used Firefox as a basis and just built Galaxy as an addon. This is technically impossible. Browser addons do not technically have access to the computer and local filesystem to do things like installing and updating video games or various other things that something the nature of a full blown video game client needs to do. Even if browsers did expose such capabilities (which would be a massive security weakness I might add), it would be an extremely horrible idea as they would have to try to make their entire video game client platform work across an endless number of multiple versions of Firefox on Windows/Linux/Mac over time, and addons regularly break when new versions of the browser come out. That would set GOG's tech support guys on fire every 6 weeks when a new version of Firefox comes out. Fortunately though it's not actually possible to do what you suggest so it never would have been a possibility to consider.

Aside from that though, as part of the Mozilla Firefox redesign, Mozilla is shit-canning the old legacy XPCOM addon interfaces that have been around forever and which multitudes of existing addons still use to this day. The reason they're doing this is multi-fold. One is that the old interface is highly flexible but exposes addons to too many browser internals which when addons use them ultimately causes those addons to break every time Mozilla puts out a new version of Firefox if the Firefox code has changed in any way that those addons were relying upon. Everyone who uses Firefox has experienced this. "I updated Firefox and now my _____ addon is broken." Just imagine how well that would go over with GOG technical support if their entire flagship gaming client was dependent on random changes in each new release of Firefox.

So ultimately, it's an interesting thought but it is completely impractical from a technical perspective to consider developing a full fledged gaming client on top of Firefox, both considering doing it as an addon due to addons not having the technical capability to actually do what a gaming client needs, and embedding firefox inside of a custom solution since it is a constantly moving target without a stable API/ABI as it never intended to advertise it as having such.

So if a developer is exploring what pre-existing web technologies to consider basing their project on with a new development, they will most likely rule out Firefox rather quickly. Looking elsewhere there are not a lot of other choices out there and Google's Chromium bits are already quite popularly used by a lot of other projects and products and are arguably much better suited to the task, even though Chromium comes with its own set of challenges.

From the perspective of choosing an actually feasible solution that is technologically sensible GOG chose the right decision to go with the Chromium engine to base GOG Galaxy on. The only other option would have been to more or less write all of their own code completely from scratch, which would probably have meant we might have ended up seeing Galaxy in 2025 or so. :)

Your premise is false, in that you (at least appear to be) assuming that the only benefit that a 64bit program running on a 64bit operating system has is that it can access more than 4GB of memory, and that if a program does not need or use that much memory it has no benefit of being 64bit or running on a 64bit operating system. That is a false premise and false assumption.

Programs that do require more than 3GB of memory, or that could benefit from accessing more than 3GB of memory may benefit from the increased amount of memory available on a 64bit OS, but to think that this is the only benefit that a 64bit CPU and OS provides is incredibly short sighted. The overwhelming majority of computer software out there uses only a fraction of 4GB of memory, and until recent years not many games used or would have benefited from breaking the 4GB memory barrier. In recent years some games have started to push the envelope and now include huge amounts of data, mostly in the form of 3D textures but other data as well.

Some games initially provided both 32bit and 64bit binaries for compatibility with older systems and to not lose customers using 32bit systems, while providing an enhanced experience for people with newer 64bit systems. Nowadays we're seeing more and more games take the full leap to 64bit however such as Grand Theft Auto V, not only because of the massive amount of textures and other data the game uses, but because of the actual performance gains that running under AMD long mode provides.

The technical fact of the matter is that any CPU or memory bound application running in 64bit long mode on an AMD64 class processor can experience a dramatic performance improvement over the same code running as a 32bit application on the same system, and this is completely regardless of how much memory the program uses. In fact, the performance benefits have absolutely nothing whatsoever to do with the program or operating system being 64bit, but rather they have to do with the fact that the AMD64 architecture includes new features that are only available in 64bit long mode that are not available to programs running in 32bit legacy or compatibility modes by the very design of the processor itself. Namely, long mode provides double the number of general purpose registers, a whole host of new addressing modes and other features that are extremely beneficial to compilers for the purposes of code optimization. These features give CPU/memory bound applications up to about a 30% performance boost in 64bit long mode compared to the same application running in 32bit compatibility or legacy modes. The actual amount of performance gain will depend on the specific nature of the program and what it is doing, however video games in particular fit very much square into the "CPU/memory" bound type of application space and are thus among the class of software that greatly benefits from this newer architecture.

As it turns out, the fact it is 64bit architecture is kind of irrelevant. The primary performance benefits apps get have nothing to do with it being 64bit. Not only that, but 32bit applications also get a potential performance benefit running on a 64bit OS because when they call upon the OS to do something, that happens inside the OS largely with 64bit code which benefits from the same performance benefits I described above.

In fact if we temporarily ignore the performance benefits I mention above, 64bit applications actually are more likely to run slower than 32bit ones on the AMD64 architecture because 64bit address pointers consume twice as much space in the cache lines as 32bit pointers do, which means you can fit less of them in the cache which in turn means that the system experiences a higher number of cache misses in 64bit long mode overall which means the processor has to go out to main memory more often which is more costly. Likewise, similar happens when using 64bit data types instead of 32bit types, because they take up double the amount of memory and have double the cache footprint. That only affects data that is explicitly declared one way or another to result in a 64bit data type being used, or a type is used that the compiler/platform chooses a 64bit type instead. On Windows 64bit however Microsoft chose the default integer types to be 32bit, so 64bit applications will use 32bit integers by default unless the developer specifically chooses to use 64bit integers. The end result is that the performance hit due to increased cache pressure in 64bit applications running in long mode is reduced to be practically negligible especially when we take into the account the actual real world optimization/performance benefits that we temporarily ignored above.

There are very few programs that will not benefit from being 64bit running on a 64bit OS, and for the most part they are programs that are disk/network I/O bound so hardware is the bottleneck and the CPU is twiddling its thumbs waiting for the hardware to respond, and the long mode performance optimizations end up doing their job but not mattering in the end due to hardware latency. Another type of program that wont benefit or may even run slower, is a program that has custom hand written assembly language tweaked for x86 but which does not have equivalent 64bit AMD64 assembly code present, causing the compiler to fallback to a version written in a higher level language that is not as optimized. This is only a problem where an application is being rebuilt to be a 64bit app where the same amount of low level optimization has not been done as it was in the 32bit build. That's a rare occurrence nowadays really, and with many apps it would likely be unnoticeable anyway as modern computers are so fast to begin with.

In short, thinking that applications only benefit from being 64bit on a 64bit OS if they need more than 4GB of memory (3GB actually) is extremely short-sighted and just technically incorrect right across the board. :)

No matter how you slice it, in the year 2015 using PAE to extend memory on a legacy OS like XP is just bad advice all around. This is well documented both on Microsoft's own website and across the larger web. There are a very small number of cases where PAE is theoretically beneficial but performance is not one of them, and moving to a 64bit OS and applications is a vastly superior solution. In the context of playing video games, PAE offers zero practical benefit, in particular because there are almost no 32bit-only personal computers out there that implement PAE at the motherboard chipset level that have more than 4GB of memory. Sure, one can claim and attribute all kinds of wishy washy benefits to it and not actually do any actual scientific benchmarking or even research pre-existing information on the web on the topic if it makes one feel good about it I suppose, but it doesn't give any benefits except perhaps at a psychological level of wishful thinking really.

If anyone doubts the reality of this claim however, I strongly encourage anyone and everyone to do extensive research on Google about the pros and cons of using PAE on 32bit systems to break the 4GB memory barrier, and more importantly I encourage everyone to research the details behind Microsoft's PAE support for XP which they later dumbed down on XP client systems due to the problems they themselves discovered and have well documented on their website.

Regardless of all of that though, the overwhelming majority of people out there have no idea what PAE even is and they are not using it and are not likely to ever use it, so it isn't terribly interesting other than an extremely unimportant part of ancient computer history. I'd classify it on the "matters in 2015" scale right alongside MSDOS using XMS/EMS. :)

It's nice to know your background is in video driver development.

I hear what you are saying and of the remaining discrete video card or iGPU makers big boys of today it is Intel, AMD, and nVidia. These guys make the current graphics chips you are going to need drivers for and because of this monopoly they control what OS you are going to use.

Back in the day you are correct that there were a multitude of different video card vendors and having parents who ran a computer store back in those days which I called the golden era of computing (1980-1995) I was privileged to have assimilated all that hardware and software knowledge and yes I still I've seen my share of Monochrome, CGA, EGA, and VGA graphics cards come and go and still have a majority of these in my possession as well as a bunch of these still sealed brand new and unused. I can spout a few names like IBM, Genoa, Tseng Labs, Orchid in the early ISA days. In the late 90s 3Dfx, ATI, Diamond, and nVidia ended up dominating for a good while in the discrete graphics cards and Intel stuck with making CPUs and very lower powered iGPUs.

But if you consider the differences between the older OSs compared to XP there hasn't been any other OS other than DOS that has stayed around as long in the PC industry. So while you are correct that a company which may be capable of creating a XP driver may choose not do so because of their finite resources but in that case I would hope the company (Intel, AMD, nVidia) would allow some allow some sort of open source driver code sharing so people who use XP, Linux, or MAC OS could reverse engineer and adapt it for their OS. Obviously I wouldn't suggest any company to make a graphics driver for DOS or Windows 3.1 (XP actually can actually run Windows 3.1 apps) because hardly anyone uses it and XP would still be a superior choice to spend the effort due to its massive software and hardware support infrastructure. They could at least they could open up the possibility of charging more for the CPU to get the driver support of their choosing that way the finite resources you mentioned, engineers who would work on the driver could be compensated. But if you were talking about Intel specifically I just find that with such huge resources and money that even producing a measily XP graphics driver would actually put a dent into their work force in comparison to say a team developing one for Windows 10. They have been entrenched in Microsoft code since the beginning way before Microsoft became a contender that IBM underestimated.

However being that Intel's iGPU is built into the CPU and the Intel HD 1000 is practically given away for free with most of their current low end CPUs it's a shame they won't give people the necessary tools to create XP drivers for it even if their own company wishes not to expend any resources to it for their own reasons. I'm referring to only Haswell and later processors that Intel chose to drop driver support. These iGPUs are still not up to par with discrete graphics cards but getting closer by the time the Cannonlake iGPU is released. I know most people on here have a negative view of XP on GOG but I enjoy testing out the limits of Intel's newer iGPU's on XP to see how powerful they can run older games. The advantages of not requiring a graphics card for one is bonus and the lower wattage makes these a great HTPC candidate for a very slim profile chassis.

The only issues with the Ivy Bridge iGPU that I found comparing the XP versus Vista version is the lack of HDCP support. In discrete graphics cards I have no HDCP support issues under XP. But Intel purposely did not add this into their XP driver to force people to Vista or Windows 7.

At the Haswell stage they dropped Vista driver support as well which didn't make sense since Microsoft hasn't dropped support for that OS. The Ivy Bridge being 22nm and Haswell as well I just see how hard it would have been for them to even include an XP driver. If they had another die shrink going to Broadwell then I can see maybe they have to reinvent the wheel. But can you honestly tell me each iGPU they release with their CPU requires a completely new driver from scratch or is it just merely minor tweaks of the driver to support the extra execution units?

I see what you mean by the cost incurred for the engineers but wouldn't it be better to get the Windows 10 and 7 driver out first as a priority and then for the team to iron out an XP driver afterwards rather than dividing the team's effort between a new and old OS? To me that would slow things down. Once they got the driver out for the dominant OS then releasing a XP driver would just be extending the hardware support. Again I'm not suggesting them to go back even further and add Windows 95 / 98, 3.1, and DOS driver support which as you know probably amount to an even smaller amount of users than Linux today and wouldn't be practical.

Yes I know XP ended support on that fateful day but Microsoft has been known to change their mind as they have that right and re-extend support. If I remember correctly wasn't it originally slated maybe 5 years prior to 2014 like around 2009 but when they realized XP was not going away after Vista they decided to extend it to 2014? However once W7 finally got its groove on they could comfortably push a 2014 support deadline.

And like I said if Microsoft did release XP Ultimate 128-bit using the same user interface of XP and classic mode and incorporating even Vista's Aero and W7-W10's user interface themes and make it true 32-bit and 64-bit backward compatible to run all their apps I believe they could redominate the market with just one OS version. If any company could do it would be Microsoft.

It's interesting you were using Thunderbird. I think I was using that on my P4 until I upgraded.

I was using Outlook Express 4.0 I believe it was. I know it's pretty dated but I tended to retain the older versions because of their speed or interface. Now I haven't played around with later OE versions since so there might be a better once released since then.

However the biggest disadvantage probably was this 2GB limit on the mail size and if you went over it corrupted the entire mailbox.

I think I ended up using Thunderbird to download all my email into one location because it could have large sized mailboxes. I'll have to reinvestigate that as I liked having all my POP email in one location without having to manually login to each email account.

This was my goto email client. I think I tried some other alternatives like Eudora and one other I have forgotten but I kept Thunderbird.

Whether GOG Galaxy continues using Google code it's really up to them but I still find Firefox code to be leaner and I despite the "GoogleUpdate.exe" that hangs in the background process even when I close Chrome. So even if Google might have their code more able to adapt you have to decide if you want Big Brother lurking and data mining everything about you. But Firefox version 22.0 seems to the fastest of the new versions and the most stable. It also is the last version to contain the option to disable and enable Javascript in the options. However I have a Javascript enable disable Firefox addon so I can toggle it on and off when the computer starts freezing up from bad Javascript code.

Also if you take note I believe it was around Firefox version 29 they tried to give it a Google Chrome look. So whatever departures or changes they've had I've mainly used older versions of Firefox since they worked or until some website like BofA requires you to update to a newer version.

I have Firefox Firefox 3.6.28 on another system just to browse Youtube videos and general internet research. I've not had any major problems yet and you can force ignore the Adobe flash update warnings that keep asking if you want to use Adobe Flash. Sometimes older versions actually run better than newer Adobe Flash versions and I noticed the newer versions added some features I disliked when watching Youtube clips.

LOL Good explanation on Firefox vs Google code choice. But yes the Firefox add-ons did allow a lot of control. I'm not sure if it was that limited where you couldn't create a network connection between two people as I think I saw in one of their later versions some Firefox Chat Beta so they must have introduced some new code. My idea was a way to have a buddy list within a browser interface and just network the gaming data between people playing the same game. If MSN Gaming at one point was able to make a gaming zone within Internet Explorer I just don't see something like the Firefox addon wouldn't work. If it's not possible to do it as an add-on then just use Firefox 22.0 source code. But if it truly is easier for the GOG team to use Google code then they have to make GOG Galaxy a reality now than wait another 10 years in 2025 if they wrote their own code from scratch as you stated.

At the moment even if they released GOG Galaxy Final, I just don't see enough titles that support multiplayer network play and I even suggested they deep discount them $2-$5 or make free and focus on some older games like Worms multiplayer, Doom 1 and 2 multiplayer, Warcraft 1 and 2 multiplayer, and recently Mortal Combat Arcade Kollection which would be some quick session type games with minimal download times to install instead of some of these larger GB sized games which cost around $22 to $50. Getting GOG Galaxy users is the hard part but once you got them pouring in then they would later start buying other GOG Galaxy network playable new games.

Actually your assumed premise of my premise is false as that was not what I was stating. You took my question to mean more than what I was simply asking. I was curious what programs you actually used that exceeded 4GB and it didn't refer specifically to only 32-bit or 64-bit applications. I was not stating running programs using 4GB or more as some inherent advantage only found on 64-bit. I was only curious if you used any programs that exceeded 4GB of memory usage out of curiosity as I myself can't think of any programs I've ever used on either 32-bit or 64-bit OSs that have used massive amounts of memory unless you count creating a 28GB RAMdrive.

Now on the gaming side I have installed Crysis on Vista 64-bit and did do some basic comparisons on a friend's laptop a few years ago between running 2GB and 4GB on it versus XP 32-bit with the same memory configuration. I did notice at the time that the XP 32-bit with 2GB Crysis did run faster than on Vista 64-bit with the same 2GB. Remember I create Multi OS boot setups so the hardware is exactly the same and I'm using cleanly installed brand new clean OS configurations with all drivers and service packs without any bloatware for the fastest set up. In fact I usually tweak the OSs to make them faster by disabling System Restore, unnecessary GUI Graphical Animations, and switch to Windows Classic Quick Launch Classic mode for one quick launching. Then I launched Crysis and pulled up the fps and testing different Graphics Detail Settings and did some game play testing to see if it was laggy or fluid.

However during that test I did when both had 4GB installed XP 32-bit was still faster however the extra 2GB of memory made Crysis run smoother on Vista than with just 2GB. As I said earlier from what I've read PAE is really 36-bit as they claim allowing up to 64GB of accessible memory which in any form is still better than just 4GB capped or less under XP 32-bit. Even if it turns out the 64GB is split into 4GB isolated memory regions when programs use them it is still better than not using that extra memory installed sitting idly. Again I haven't tested the patched kernel for 64GB PAE yet so I have no idea if that what it really does in functionality.

As for as your explanation regarding AMD64, not sure why you focused on AMD 64 bit processors and not Intel 64 bit CPUs but I stopped using AMD since the K6 days as at the time they were still compatible enough to Intel and cheaper. However I later switched back to the Pentium 1 line continuing to the P3 1GHZ - first major threshold back in the day. Then I followed that up with my P4 - 3.06 GHz - yet another threshold. Then the current system the Quadcore Ivy Bridge I'm on now. Yes back in those days speed was the benchmark of PCs: 4.77 MHz, 8, 10, 12, 16, 20, 25, 33, 40, 50, 66, 75, 90, 100, 133, 166, 200. The first leaps from 4.77-33 MHz were excruciatingly slow but I never was a full adopter of AMD due to backward compatibility concerns. I think the last truly compatible Intel CPU AMD made was the 486. After that I lost interest when compatibility was important. I remember some DOS games back then would not work on AMD even if there were only a handful. However when it came to Windows 98SE AMD CPU compatibility wasn't a deal breaker.

However I like your explanation of how 32-bit code would run faster on a 64-bit OS. However I've read about this before and it's similar to how 16-bit code runs on a 32-bit CPU. However as you know most 32-bit Windows code we would be running came predominantly from XP. If you were to run this XP 32-bit code in Vista 64-bit or later XP 32-bit compatibility with all software is not exactly 100%. This is like running 98SE code on Windows NT 4.0 and most of the time it may not even start or crash.

I haven't tested this thoroughly but from what you stated concerning running 32-bit XP code on an AMD64 CPU. I don't have any AMD64 CPUs so I don't know if this would matter in your statement.

I could however easily test your theory with the Intel 64-bit CPU by using Vista 64-bit running Crysis 1 32-bit version and compare it to the 64-bit version. Then do another comparison using XP 32-bit and compare all 3 results. Also I believe you can force DX9 on Vista to match XP compatibility testing results to make it a fair across the board as using DX 10 and later will be more taxing and compromise the Vista benchmark results.

Again, I never said only 64-bit applications only benefited on a 64-bit OS and you were wrong to assume that's why I asked you if you used any programs that used more than 4GB of memory and the only reason I asked was out of curiosity as I may be interested in trying some of these out. :)