Microsoft Taking Official Petitions to Keep XP Alive

[this_is_gav]

There is a kind of pattern here... look back a few years when NT4 was the system of choice in the business world... 2000 comes along and there's hardly any drivers for it, and it's patchy.... it starts getting a bit more support, people start writing stuff for it within the year and lo! XP is on the horizon...

Seems it's going the same way with Vista - XP is firmly the choice of business workstations, running off 2003 servers - yes, Vista is starting to get more support and drivers etc and people are adopting it more and more, but Windows 7 is now on the horizon, only a couple of years since Vista first hit the shelves... How long was XP the "norm" before Vista was announced?

I can see where you're coming from, but most devices shared drivers for 95, 98 & ME, while 32-bit 2000, XP & 2003 usually do, so I wouldn't expect much difference with Vista & Windows 7. They've already said that it's built on Vista, so I expect Vista drivers will work on Windows 7 in the majority of cases. Program compatibility should be even less problematic.

[Saracen]

Other than your work and ability to find stuff in super-dense filesystem hierarchies, because unlike XP, indexing actually works on Vista, as well as many other neat cravats. Seriously, Vista RTM was a *LOT* more polished off than XP RTM was. And if you do more than picking your nose and typing letters all day, 4GB or more memory can easily be used productively.

Well, given an element of poetic licence in your use of "if you do more than picking your nose and typing letters all day", that's exactly what a lot of corporate machines do.

For example, go into most major supermarkets. If they're using touch-sensitive screens at the checkout, and most are or are in the process of changing over, they're running P.O.S. systems under XP. How many checkouts do they have? Or, what about PCs in admin offices, accounts offices, credit control, secretarial uses, etc. As Peterb points out, they're running relatively undemanding applications - WP, maybe spreadsheets, accounts packages, perhaps presentation software. I have a machine here doing that, perfectly satisfactorily, with 512MB under XP on an Athlon XP1800. It does not need Vista, or a powerful processor, or bucketloads of RAM to do it.

Yes, there are uses for more RAM than that, but the vast majority of corporate PCs don't run that 'power-user' stuff.

[Saracen]

Is W2k a "Dead Duck", and why? Answers on a postcard please.

Ask Dangel - it's his term so he can define what he meant by it.

For my part ....

- is it still sold? Nope.
- is it likely to be something anyone would buy? Probably not.
- is it still capable of doing a satisfactory job. Hell, yes, and on a fair few PCs, no doubt still does.

I'm personally aware of one large company that was, until very recently, using it on hundreds of servers. That, now, has been phased out, but it's very recent (by which I mean, months).

[aidanjt]

And I completely agree, in many cases trivial amounts of RAM are only needed for trivial workloads. Looking at other Operating Systems like *BSD and Linux, they can do an absurd amount of work on a machine such as you describe. But given the price of DDR2, and enabling the ability of serious power-user apps, it's not a bad thing either. But I do see your point of view from a business prospective, and the hassle of updating machines and operating systems and having to fiddle with new licenses and such isn't needed, although, with new machines, there's no point in doing a backward step to a nearly depreciated OS.

[Saracen]

And I completely agree, in many cases trivial amounts of RAM are only needed for trivial workloads. Looking at other Operating Systems like *BSD and Linux, they can do an absurd amount of work on a machine such as you describe. But given the price of DDR2, and enabling the ability of serious power-user apps, it's not a bad thing either. But I do see your point of view from a business prospective, and the hassle of updating machines and operating systems and having to fiddle with new licenses and such isn't needed, although, with new machines, there's no point in doing a backward step to a nearly depreciated OS.

For you and me, memory prices are pretty trivial. But, even just looking at memory prices, what would it cost Tesco to upgrade just their checkouts from 1MB to 4MB?

Firstly, they'd have to do a lot of testing to make sure there were no unexpected gotchas, because you don't want to find out about it after you've upgraded tens of thousands of machines. Then, you've not got the £20, or £40, it cost to do your machine, but tens of thousands of lots of £20-£40.

Then, when do you do it? You certainly aren't going to close a store to do it, 'cos it'll cost several hundred grand of lost turnover, per store, to do it. So, you've talking overnight. Now, you have to get engineers to go in, take the PCs out from their housing, open them up, monkey with the motherboard, perhaps remove the PSU, risk some machines having problems restarting, and put it back in it's home, then restart and test. You certainly don't want checkout staff coming in in the morning and finding their checkouts don't work. And you might have thirty, forty or more checkouts per store in the bigger ones, and how many thousand stores? And it needs careful project management, just to get the right people to the right stores on the right days, and for non-24 hour stores, arranging night security staff to be on hand, and so on.

That memory upgrade just became a multi-million pound project, that will probably takes months.

And that's just the RAM, which is trivial compared to an OS change.

[aidanjt]

And I'm not disagreeing with you there. But the main point is, why buy a new machine, and revert it to an Operating System that's already starting to showing signs of bit-rot?

[Saracen]

And I'm not disagreeing with you there. But the main point is, why buy a new machine, and revert it to an Operating System that's already starting to showing signs of bit-rot?

Because, perhaps, what's "bit-rot" to you is "tried and tested" to a corporate that's putting custom (or custom-tweaked) software on it.

Take that Tesco example .... and it's just an example. It'll take them a couple of years to design, write (or adapt, or get adapted) PoS software, then test it with a hardware base (and bear in mind, that means non-standard devices like chip-and-pin readers, scanner-scales, receipt printers, etc. And it's a project that's got to be rolled out across hundreds or thousands of stores. It takes a LONG time to do a project like that, and it tends to only get done when it has to. I've done that type of operation running on hardware that's 20 years old, using software written for proprietary operating systems.

And the whole thing has to be done as part of an on-going, and hopefully growing business. So there'll be new stores coming along, and refurbs of older ones. They need very careful change control to be able to support that kind of operation, and NO change to that structure is done lightly, once they start rolling it out. And, as like as not, they have a contract with a main supplier ensuring that hardware availability is assured for some years, as is parts availability.

So if you're looking at a project that could well take several years from inception to completion, at some point you have to take a decision and set it in stone. So any project that's competing now (like the three-year project I mentioned some posts back) will have reached that set-in-stone point long before Vista was available.

And given that, you're going to have major customers like that with decisions already made and half-implemented. It's a huge undertaking to go back now and change to Vista. And it will ONLY happen if there is a very convincing reason for it, because it's the kind of thing that will add months of work, cost millions and probably delay completion by a year, by the time you've tested.

And, of course, no store can just change a system and shove it in a customer-facing role in a shop. Those systems are covered by Weights and Measures legislation, and the software subsystem (and hardware performance) has to be certified by Trading Standards (or a self-cert authority) as being compliant with type approvals. Software has a checksum, and software installed in store MUST comply with a known checksum. Change the OS and if it require ANY changes at all to that software suite, you risk that checksum changing, and if it does, you now also need to get EVERY checkout with the new software recertified for weights and measures compliance.

Which comes back to "bit-rot", and the point I've made over and over again in this thread. As far as corporates are concerned, upgrading is potentially a very different experience from you or me doing it at home. Obviously, those weights and measures requirements don't apply to lot of corporate PCs, but the background principles do.

So, if it ain't broke, corporates will not spend what it would cost to change unless there's a very convincing reason. And, if part way through a project, or even idf they've just got past the test and decision-making point, they'll need sound reasons why changing OS now makes commercial sense. Such reason might exist, but even given that Vista is a better OS (and in many ways, it is), will those improvements be reflected in sufficient benefits to make those potentially very substantial costs worth absorbing. Everything I'm seeming suggests that usually they're not, and it's certainly my opinion, even for my own infrastructure of ten or so machines.

[Splash]

Ask Dangel - it's his term so he can define what he meant by it.

For my part ....

- is it still sold? Nope.
- is it likely to be something anyone would buy? Probably not.
- is it still capable of doing a satisfactory job. Hell, yes, and on a fair few PCs, no doubt still does.

I'm personally aware of one large company that was, until very recently, using it on hundreds of servers. That, now, has been phased out, but it's very recent (by which I mean, months).

Let me rephrase my question then... should Microsoft be pressured into starting to sell W2K again?

[Saracen]

Let me rephrase my question then... should Microsoft be pressured into starting to sell W2K again?

A rather loaded way of putting it.

You could equally well say that those that refuse to buy Vista because they want XP are simply refusing to be pressured by MS into buying Vista.

If people want to buy W2K, they're entitled to refuse to buy other OS's just because MS want them to. On the other hand, MS are entitled to decline to sell it, if it doesn't make sense to them to do so.

No, MS should not be pressured into starting to sell W2K again, but if enough people would buy it, it might make sense for them to do so. And the same applies to XP. The "pressure", however, is simply people telling MS that, for whatever reason, they don't want and won't buy Vista.

MS are trying to pressure people into Vista, by making XP unavailable. Their problem is that large numbers of people are refusing to bow to that pressure. And the pressure on MS is therefore of their own making, because while MS can, if they wish, decline to continue to supply XP, it's their own fault if Vista isn't good enough to convince those that still want XP to buy that instead. You can lead the buyer to water, but you can't make him buy Vista. And, unless or until someone comes up with a good enough reason to spend money on Vista (and nobody has yet), I won't be doing so, and it seems a lot of other people won't either.

[mikkri]

Business will look into Vista as soon as all these special support agreements with Microsoft expire. No big company is going to run OS that doesn't have commercial support on desktops.

[aidanjt]

Because, perhaps, what's "bit-rot" to you is "tried and tested" to a corporate that's putting custom (or custom-tweaked) software on it.

No, bit-rot is the degradation of software over time due to lack of maintenance of the various sub-systems, bit-rot factor is as true to me as it is to the biggest corporations.

Take that Tesco example .... and it's just an example. It'll take them a couple of years to design, write (or adapt, or get adapted) PoS software, then test it with a hardware base (and bear in mind, that means non-standard devices like chip-and-pin readers, scanner-scales, receipt printers, etc. And it's a project that's got to be rolled out across hundreds or thousands of stores. It takes a LONG time to do a project like that, and it tends to only get done when it has to. I've done that type of operation running on hardware that's 20 years old, using software written for proprietary operating systems.

And the whole thing has to be done as part of an on-going, and hopefully growing business. So there'll be new stores coming along, and refurbs of older ones. They need very careful change control to be able to support that kind of operation, and NO change to that structure is done lightly, once they start rolling it out. And, as like as not, they have a contract with a main supplier ensuring that hardware availability is assured for some years, as is parts availability.

So if you're looking at a project that could well take several years from inception to completion, at some point you have to take a decision and set it in stone. So any project that's competing now (like the three-year project I mentioned some posts back) will have reached that set-in-stone point long before Vista was available.

And given that, you're going to have major customers like that with decisions already made and half-implemented. It's a huge undertaking to go back now and change to Vista. And it will ONLY happen if there is a very convincing reason for it, because it's the kind of thing that will add months of work, cost millions and probably delay completion by a year, by the time you've tested.

And, of course, no store can just change a system and shove it in a customer-facing role in a shop. Those systems are covered by Weights and Measures legislation, and the software subsystem (and hardware performance) has to be certified by Trading Standards (or a self-cert authority) as being compliant with type approvals. Software has a checksum, and software installed in store MUST comply with a known checksum. Change the OS and if it require ANY changes at all to that software suite, you risk that checksum changing, and if it does, you now also need to get EVERY checkout with the new software recertified for weights and measures compliance.

The Tesco/PoS example is flawed, point of sales machines are considered embedded devices, they're also intentionally isolated from the regular intranet/intardwebs for security and reduced maintenance, which is a whole different ball game from a regular workstation/business PC. Also, the point about custom software breaking on the new OS is flawed as well, any correctly designed code written for NT4 will still run on Vista 99.99% of the time, Microsoft very very rarely strips old functions out of the WinAPI.

Which comes back to "bit-rot", and the point I've made over and over again in this thread. As far as corporates are concerned, upgrading is potentially a very different experience from you or me doing it at home. Obviously, those weights and measures requirements don't apply to lot of corporate PCs, but the background principles do.

Which is true, until they need to roll out a new software package for whatever reason, only its broken because the OS is missing vital subsystems or specific functions of particular subsystems that's required for running it. Backwards compatibility is a lot easier to work with than trying to backport OS subsystems without having the source code, even with the source code, it's a serious pain in the ass.

So, if it ain't broke, corporates will not spend what it would cost to change unless there's a very convincing reason. And, if part way through a project, or even idf they've just got past the test and decision-making point, they'll need sound reasons why changing OS now makes commercial sense. Such reason might exist, but even given that Vista is a better OS (and in many ways, it is), will those improvements be reflected in sufficient benefits to make those potentially very substantial costs worth absorbing. Everything I'm seeming suggests that usually they're not, and it's certainly my opinion, even for my own infrastructure of ten or so machines.

Nobody is saying they need to 'upgrade now! (TM)', just that there's no point in shoving an antique on a new computer when the new OS works with legacy infrastructure, and this way companies can slowly migrate to a new OS as they replace dead or depreciated machines.

[Salazaar]

It's not so much what's so 'goddamn special' about xp, but more a case (for me at least) of what's so 'goddamn special' about Vista?

For me, it's like this. I have XP, and I don't have Vista. XP is legit, and paid for. If Microsoft are prepared to give me the ten or so copies of Vista I'd need to bring all my systems up to Vista, and to pay for the time it'll take to reinstall those systems, and for me to get used to Vista, I'll do it. Or, if someone .... anyone ..... can point out what's so 'goddamn' good about Vista that it justifies the money, and time, and effort, involved in upgrading, I'll even pay for it myself.

But so far, nobody has been able to come up with a convincing reason for me to change. Feel free to have a try at it though.

Is XP perfect? Nope, but neither is Vista.

Is Vista better than XP? Even in my rather jaded view, yup, probably. It just isn't better by anything like enough to justify the cost of me switching to it.


So .... what's so 'goddamn special' about XP? I've got it, it's installed, working and does what I need of it, that's what. It's that old adage in practice ..... if it ain't broke, don't fix it.

I agree 100% with this point of view.

No, XP isn't perfect but it does a damn good job and I don't see any compelling reason to spend quite a lot of money 'upgrading' to Vista.

[aidanjt]

Who's talking about spending money on upgrading anything?.. We're talking about people trying to force Microsoft to continue to support what is in computing, a museum piece, for *new* machines.

[Saracen]

Who's talking about spending money on upgrading anything?.. We're talking about people trying to force Microsoft to continue to support what is in computing, a museum piece, for *new* machines.

I've been talking about spending money on upgrading, and what OS to get on new machines, and why in many cases there are good reason for people refusing to go to Vista and wanting XP on new machines.

And Salazaar was quoting me.

So rather than ask who's talking about spending money on upgrading, try reading the thread, because I've been talking about it since since my first post, both in terms of me spending money, and why many corporates want XP not Vista..

[Saracen]

.....

Nobody is saying they need to 'upgrade now! (TM)', just that there's no point in shoving an antique on a new computer when the new OS works with legacy infrastructure, and this way companies can slowly migrate to a new OS as they replace dead or depreciated machines.

Yes, there is a point in doing that - only supporting one platform and knowing what to expect. Not every company will either want or need to do this, and those that don't can move to Vista if they want. And probably will or have. But those that do want to do this want XP because it's what they've got and it's doing the job just fine. They don't want to move to a new platform just because the OS vendor wants them to, and in many cases, are telling hardware vendors than a new order is contingent on it being XP.

[aidanjt]

Windows is Windows, it's all the same platform regardless of version, they all use WinAPI, and knowing what to expect from Windows?.. aghm. All they're doing is delaying the inevitable and screwing themselves later on when they really are forced to upgrade their entire infrastructure in one go, both financially, and in terms of workload for their IT staff, and lost productivity for staff waiting while IT frantically runs around all the machines with a stack of unattended install DVDs or enabling PXE for remote deployment.