More Vista nonsense

[ddrueding]

A great analysis Tannin, but your software list is downright ancient. I'm not saying that it doesn't do what you want it to, likely better than newer software, but your customers won't stand for a DOS-based word processor or Quattro. I didn't switch to Vista for me, I did it for them. Now that I am well versed in it, I switched back to 2003. I'm not saying that you will want it for yourself, but I'm sure your clients will "need" it soon.

 

[Tannin]

It does indeed do it better, Dave, otherwise I'd change it. Let's take an example: Quicken for DOS. Simply, it ticks every single one of the important boxes:

• Speed: instant. No waiting, ever.
• Ability: excellent. It performs every required function that you need to run a business;
• Data entry: best-in-class. I can enter my accounts into Quicken for DOS faster, with fewer mistakes, and without ever, ever waiting for the application. Everything is on the keyboard, instant. No scrolling, pointing, clicking, no bringing up menus, it's just all there.
• Data security. Excellent. All you need is a copy of the data folder. Make that copy any way you like.
• Cost: zero. This was paid for almost 20 years ago, and every year that I continue to use it it saves me the price of a current version.
• Reliablity: 100%. I have never, ever, seen it crash or even glitch.
• Multi-tasking. Excellent. It runs in a window on my OS/2 desktop; I can cross-reference with spreadsheet data, invoices, whatever I want. If I want more screen estate, no problem: I run it over the network from my WinXP Thinkpad, freeing up the accounts machine's screen for other tasks.
• Ongoing cost: zero. Compare with the current version which requires you, one way or another, to hand your credit card over to Intuit regularly. Either it times out and you have to renew, or else (if you pay extra for the non-time-out version, they invalidate it with a new release anyway. (Trust me: I see customers grappling with those greedy, dishonest bastards regularly.)
• Installation: 10 out of 10. Copy the executables over to the folder of your choice on the machine of your choice, drag a shortcut wherever you want it. Click and go.
• Aggrivation: zero. No registration, no activation, no phone calls, no waiting on hold, no product codes to key in, nothing. Just use it.
• Appearance: excellent. The program itself is utilitarian, but I have finished my accounts and am staring out the window looking at the clouds blow past the treetops while you are still waiting for your Aero GUI to finish loading.

OK, I know that's not the point you were making, but I run my ancient software for a reason: that's it's best-in-class. When a newer product is better than whatever I'm using now, I switch to it. Hence my latest versions of PMView, Downloader Pro, Neat Image, Photoshop, Thunderbird, and various browsers. But where the newer version isn't as good as the old version, I keep the old one. If anyone ever produces a spreadsheet better than the ancient Quattro Pro I run, I'll switch to it. Indeed, I've bought and paid for newer versions, only to discover that they were inferior to the one I have, which is badged by Novell but in reality the last of the Borland versions before Corel turned it into one of their their normal slow-loading, gimmick-riddled bugfests.

(This post doesn't seem to be making any progress towards addressing your actual point. Maybe I better give Tea a go at it instead.)

 

[Tea]

A great analysis Tannin, but your software list is downright ancient. I'm not saying that it doesn't do what you want it to, likely better than newer software, but your customers won't stand for a DOS-based word processor or Quattro. I didn't switch to Vista for me, I did it for them. Now that I am well versed in it, I switched back to 2003. I'm not saying that you will want it for yourself, but I'm sure your clients will "need" it soon.

Tannin's view on this, Dave, is that no-one is holding a gun to his head, so why should he?

He gets cynical, recalcitrant, and very stubborn when people try to tell him what to do. Always has - since shortly after birth according to his mother (who is a nice lady and probably deserved a more malleable son).

(I'm not saying you are trying to tell him what to do here, well maybe you are, or at least offering a little friendly advice. That's not what triggers Tannin's famous pigheadedness: it's when powerful people or powerful companies think they can steamroller him that he digs his heels in.)

As of today, we have quite a nuber of Vista customers. I can think of at least two, but I'm sure that there are more, probably as many as five or six. We have almost as many Windows 95 customers as we have Vista customers, and many more that run Win 98. 2000 remains reasonably common, but the vast majority of people run XP. Off the top of my head, around 90%, maybe a little higher.

(We would have Mac customers too, maybe six or ten, but Tannin sneers at them until they go away.)

OK, that will change, we wil gradually get more people running Vista, but not in any hurry. Right now, the ecomomics of supporting Vista are not there. Tannin says it would take more non-chargable ape-hours to learn the damn thing than we could possibly justify in terms of the trivial number of chargable jobs that would eventuate. Time enough to learn it when we actually have some measurable customer demand for Vista skills.

The longer we delay, the cheaper and easier it becomes:

• first because the hardware gradually gets more compliant and the drivers are less difficult - as with any Windows version
• second because even Vista will probably improve a little as time goes by and Microsoft try to fix it
• and third because the longer Vista is around for, the more expertise on it becomes readily available for instant, painless consumption in a time efficient manner: here, through Google, and a through a host of other sources.

Another reason to delay, Tannin says, is that we do not yet know for sure that Vista will ever become a mainstream OS. Tannn says that there is every chance that Vista will disappear from the installed base within a very short time, just as ME did: smart people refused to install it and stayed with Win98 until they could migrate direct to 2000 or XP, while dumb people who got lumbered with ME either downgraded to 98SE again or (before too long) upgraded to 2000 or XP. We barely had to touch ME - it never accounted for more than a very small percentage of the systems we had to work on. Any time spent learning ME would have been time wasted.

We have, however, become expert at upgrading Vista to XP. We see a steady stream of dissatisfied customers (from other shops) with machines, mostly notebooks, that they can't stand using any longer. We have had lots of practice at ferreting out obscure drivers from the lousy, error-riddled websites that pox-ridden screw-the-customer companies like ASUS and Toshiba provide, and although we find the horrible standard of customer support from (e.g.) Toshiba painful, we mostly succeed in the end and it's quite a profitable business. It costs our customers quite a lot to upgrade - typically ~$250 by the time they buy an XP Home licence as well as pay us for our time, but they go away happy, with a system that performs as you have every right to expect a modern computer should.

Finally, if we are going to have to develop Vista skills, Tannin sees no reason whatsoever why he should be the one to do it. Damien can become our in-house Vista specialist. He is younger than Tannin, and has a greater tolerance for bad products. In any case, Tannin is transitioning to a part-time role at the workhouse. He is taking on various other things, such as web design jobs, and aims to wind up doing no more than 2 or 3 days a week in the shop.

Nevertheless, though he is is rarely thoughtful enough to say so, he is grateful that people like you are prepared to buggerise about endlessly with excreta-encrusted things like Vista so that, when the time comes, he can instruct me to ask you damn-fool questions in the Tech Support forum.

 

[Tea]

A great analysis Tannin, but your software list is downright ancient. I'm not saying that it doesn't do what you want it to, likely better than newer software, but your customers won't stand for a DOS-based word processor or Quattro. I didn't switch to Vista for me, I did it for them. Now that I am well versed in it, I switched back to 2003. I'm not saying that you will want it for yourself, but I'm sure your clients will "need" it soon.

Tannin's view on this, Dave, is that no-one is holding a gun to his head, so why should he?

He gets cynical, recalcitrant, and very stubborn when people try to tell him what to do. Always has - since shortly after birth according to his mother (who is a nice lady and probably deserved a more malleable son).

(I'm not saying you are trying to tell him what to do here, well maybe you are, or at least offering a little friendly advice. That's not what triggers Tannin's famous pigheadedness: it's when powerful people or powerful companies think they can steamroller him that he digs his heels in.)

As of today, we have quite a nuber of Vista customers. I can think of at least two, but I'm sure that there are more, probably as many as five or six. We have almost as many Windows 95 customers as we have Vista customers, and many more that run Win 98. 2000 remains reasonably common, but the vast majority of people run XP. Off the top of my head, around 90%, maybe a little higher.

(We would have Mac customers too, maybe six or ten, but Tannin sneers at them until they go away.)

OK, that will change, we wil gradually get more people running Vista, but not in any hurry. Right now, the ecomomics of supporting Vista are not there. Tannin says it would take more non-chargable ape-hours to learn the damn thing than we could possibly justify in terms of the trivial number of chargable jobs that would eventuate. Time enough to learn it when we actually have some measurable customer demand for Vista skills.

The longer we delay, the cheaper and easier it becomes:

• first because the hardware gradually gets more compliant and the drivers are less difficult - as with any Windows version
• second because even Vista will probably improve a little as time goes by and Microsoft try to fix it
• and third because the longer Vista is around for, the more expertise on it becomes readily available for instant, painless consumption in a time efficient manner: here, through Google, and a through a host of other sources.

Another reason to delay, Tannin says, is that we do not yet know for sure that Vista will ever become a mainstream OS. Tannn says that there is every chance that Vista will disappear from the installed base within a very short time, just as ME did: smart people refused to install it and stayed with Win98 until they could migrate direct to 2000 or XP, while dumb people who got lumbered with ME either downgraded to 98SE again or (before too long) upgraded to 2000 or XP. We barely had to touch ME - it never accounted for more than a very small percentage of the systems we had to work on. Any time spent learning ME would have been time wasted.

We have, however, become expert at upgrading Vista to XP. We see a steady stream of dissatisfied customers (from other shops) with machines, mostly notebooks, that they can't stand using any longer. We have had lots of practice at ferreting out obscure drivers from the lousy, error-riddled websites that pox-ridden screw-the-customer companies like ASUS and Toshiba provide, and although we find the horrible standard of customer support from (e.g.) Toshiba painful, we mostly succeed in the end and it's quite a profitable business. It costs our customers quite a lot to upgrade - typically ~$250 by the time they buy an XP Home licence as well as pay us for our time, but they go away happy, with a system that performs as you have every right to expect a modern computer should.

Finally, if we are going to have to develop Vista skills, Tannin sees no reason whatsoever why he should be the one to do it. Damien can become our in-house Vista specialist. He is younger than Tannin, and has a greater tolerance for bad products. In any case, Tannin is transitioning to a part-time role at the workhouse. He is taking on various other things, such as web design jobs, and aims to wind up doing no more than 2 or 3 days a week in the shop.

Nevertheless, though is is rarely thoughtful enough to say so, he is grateful that people like you are prepared to buggerise about endlessly with excreta-encrusted things like Vista so that, when the time comes, he can instruct me to ask you damn-fool questions in the Tech Support forum.

 

[LunarMist]

DPP and NX both work in XP and even 2000 for the time being. I still have PS 6.0 on the current notebook. Perhaps I'll put CS2 on the new one if it isn't hopelessly slow.

 

[mangyDOG]

I'm with Tannin on this, I remember when XP first came out it was abit of a PITA but it was substantially better than Windows 95/98/ME on new hardware and only marginally slower than Windows 2000. Vista however sucks the life out of even the fastest gear, and running it on a notebook is an exercise in futility. Interestingly I have just sold my fifth PC running Ubuntu this year, thats two more than Vista! Still small potatoes compared to XP, but it will be interesting when XP is finally withdrawn from the market.

Cheers,
mangyDOG

 

[Tannin]

Time I found the time to investigate and prepare a Linux pre-load option, MangyDog. I think it might sell quite well, in a modest sort of way, and unlike Vista, the skills required are likely to be both portable and of long-term value.

I have sold ... er ... I have the exact number here somewhere ..... ah, found it: zero Vista systems. Well, tell a lie: I once sold a system and the customer came back before it was ready to pick up and said his daughter had decided she wanted Vista on it. So I told him to buy it somewhere else.

(Not quite the cavilaer decision it sounds like: this was a system I'd sold at a rock-bottom price (for reasons which are not relevant here) and which was likely to require quite a bit of hand-holding. They had several printers, some arcane accounting software, and bugger-all computer skill. I could see a train-wreck coming up - I probably shouldn't have sold them even an XP machine, given the likely time wasted doingafter-sales support - so I just said "to hell with it, someone else can have the headache".

And there you hav it: my one and only Vista sale.

Two other people have asked for it, and then asked if they could have XP instead after 5 minutes worth of considering the pros (nil) and cons (lots) with me.

Vista has been out, what, 14 months now? Round figures, that's one customer wanting Vista every four months. To give you a sense of proportion, I sold 7 new XP systems this week alone. Oh, and two or three major upgrades involving a copy of XP, either to replace Win98 (2) or a pirated version (1). That's a busier week than average, but you expect to be busy a this time of year.

 

[Tannin]

Interestingly, nearly half of the new systems were Thinkpads. That's an unusually high proportion of notebooks vs desktops. Possibly because of the usual back-to-school thing, possibly because many of the other places seling notebooks try to lumber you with one or other of those worst of all possible Vista versions, Home Basic and Home Premium, neither of which includes upgrade to XP rights, and possibly becaise Lenove are giving us particularly good value for money just at present.

Tell a lie: it was 4 Thinkpads this week, so it must have been more than 7 new XP-based systems in total. Luckily, Damien does most of the work now, I just answer the phone and count the money.

 

[Clocker]

Looks like SP1 was out for the 64-bit versions....accidentally.

http://www.news.com/8301-13860_3-9876775-56.html?tag=nefd.pop

 

[Bozo]

Vista SP1 breaks some programs.

http://www.crn.com/software/206801221


Bozo :joker:

 

[sechs]

It is quite likely that CS4 will be a Vista-only poduct (knowing the poxheads at Adobe)

FYI, the "poxheads" at Adobe actively dislike Vista. Nearly all of their developers run XP for their main machines. Vista is sometimes tolerated in a virtual machine for testing purposes.

As long as no one in sales gets involved, and the Vista adoption curve remains definitively flat, I can see no reason why CS4 won't support XP.

If I could just get them to officially support x64....

 

[Tannin]

I really hope you are right, Sechs.

My reasoning is based on the fact that several Adobe products in the Photoshop family are XP-specific. (One or two of the Elements family, and Photoshop CS3 itself).

There is no valid technical reason for a Photoshop-type product not to run on Windows 2000 if it runs on XP (they are, after all, practically the same thing), so one can only assume that the developers (or their marketing-influenced bosses, of course) either did it deliberately (perhaps, like certain games developers, because Microsoft paid them to do it), or else that there was some trivial XP-specific hook that they chose to employ, presumably knowing full well that it would cripple the product for W2000 users.

Now it so happens that the machine I use for Photoshop runs XP (as opposed to several of my other boxes that are W2K) and this absurd restriction didn't bother me ....... but if they did it once, what is to stop them doing it again?

 

[Fushigi]

Maybe they choose to support the OS only while the OS developer offers mainstream support. W2K is out of mainstream support and is in the extended support phase where only critical security bugs are addressed. No design changes or enhancements are being developed. MS support lifecycle.

 

[Tannin]

That's a fair-sounding reason, Big F. But, seriously, what support do they actually get anyway? Maybe the programmers I know are pretty small beer, but I get the clear picture from those guys (full-on professionals, I'm talking about, not your home hacker with a copy of Visual Basic) that the most time-effective way of dealing with Windows bugs is just to program your way around them, use a different function call or whatever.

Of course, the reality is that there are never any design changes or enhancements to any windows version once a newer and hopefully more expensive version is out. For example, how difficiult would it have been to back-port a working USB driver to the earlier Win9x versions? A trivially easy task, seeing as the code was already written for Win ME. But they chose not to do it, as there was no commercial gain in it for them. And, let us remember, there was a compelling reason for them not to do it: lack of USB support was the single biggest reason people upgraded from Win95 or 98 to something newer. It has often occurred to me that a third-party USB interface for Win9X would have sold like crazy, and then immideately afterthat it would occur to me that this would be something MS would come down on like a ton of bricks, as (at that time) there was no other compelling reason to switch to ME or 2K.

Later on, of course, other reasons emerged. But at the time USB was pretty much the only thing.

What are they going to find to break Win XP? There will be something. Bet your house on it.

 

[ddrueding]

They were hoping DX10. And it will become a bigger reason in the future.

 

[Bozo]

Tannin, most of the motherboards that I remember using for '95 and'98 had no USB ports on them, nor any sockets to plug an adaptor to. Some motherboards had USB ports late in '98 life cycle.
Then again, I wasn't looking for a USB port at the time and could have missed them altogether.

Bozo :joker:

 

[Tannin]

Bozo, you did indeed miss them. USB ports came along years before the event. It was maybe 2 years before there was any USB hardware, 1 or 2 before the was even broken USB software support in Windows, and then there was a futher long wait before the hardware became practical or common. That's where the name comes from: for years - probably three years or so - everybody thought it stood for Useless Serial Bus.

But be all that as it may, lots of motherboards had USB, and an add-on card was always pretty cheap, and in any case, many (most?) people have two or three motherboards per Windows version. Did back then, often still do; if not to upgrade then because of hardware failure.

 

[sechs]

That's a fair-sounding reason, Big F. But, seriously, what support do they actually get anyway? Maybe the programmers I know are pretty small beer, but I get the clear picture from those guys (full-on professionals, I'm talking about, not your home hacker with a copy of Visual Basic) that the most time-effective way of dealing with Windows bugs is just to program your way around them, use a different function call or whatever.

I've been involved in the development of big (and generally messy) bit of software that could make or break the company, and something as easy as support of an older operating system is an enormous amount of work.

Even if it only needs to be tested, every key case and good amount of general testing has to be done on that operating system. That's time and money. And if you find a bug that operating system specific, you have to put resources on a fix, and then testing the fix on every platform. The dollar signs add-up quick.

I can tell you that it is the sales and accounting folks that generally decide these kinds of things. If the work isn't going to sell enough additional copies, it certainly won't get done. In the case of x64 support, for example, they simply didn't see enough additional money coming in to justify the effort -- even though I've seen few critical issues on the whole CS3 suite.

 

[Tannin]

That would be a fair point, Sechs ..... if we were talking about a $149 (less with coupon) package. But we ain't. We are talking about a $1200 package, and at that price for something that is, after all, just a glorified PainShop Pro with 90% fewer bugs and an even more primitive and disorganised user interface, I expect the little extra that usually comes with an 800% price premium - notably, in this context, better OS compatibility. I also expect it to be hand delivered by the head of department in person and come with a bottle of Moet and a free laptop computer.

Of course, what I expect and what I get is different animals.

sigh

 

[Tea]

Tannin, there is a typo in your post.

 

[Tea]

woops - double pozt

 

[Tannin]

No there isn't. I said Painshop Pro and I meant Painshop Pro.

 

[Fushigi]

Of course the developer's revenue might be roughly the same if 8x as many people buy the product @ $149 v. 1x $1200.

Where I work we're about to start to undertake a process of upgrading the underlying tools for an ERP package we run. The version number will go from x.x0 to x.x2. For a ".02" version upgrade the process is expected to take 4-7 months as every single bit of customization - probably around 15 man years worth - must be reviewed against the new tools. Once that's done we can upgrade the main ERP product. Once that's done we can upgrade the system's OS. Once that's done we can do what I want and upgrade the hardware. So many I'll be able to upgrade by this time next year.

 

[Bozo]

Microsoft lowers some Vista prices:
http://www.microsoft.com/presspass/features/2008/feb08/02-28BrooksQA.mspx

Getting desparate?

Bozo :joker:

 

[Bozo]

Vista's activation has been cracked. According to Slashdot.


Bozo :joker:

 

[Tannin]

They must be in panic mode. Microsoft never lower Windows prices. Never, ever. For them, tjhis is thin edge of the wedge stuff. It's like the Pope saying birth control is OK on Tuesdays ..... what happens next, we wonder.

 

[Chewy509]

They must be in panic mode. Microsoft never lower Windows prices. Never, ever. For them, tjhis is thin edge of the wedge stuff. It's like the Pope saying birth control is OK on Tuesdays ..... what happens next, we wonder.

Could it be that MS has finally realised that the public perceives Vista as a not-so-great product and are discounting accordingly... in an effort to boost sales.

Hopefully the price won't jack up again with the release of the next (and rushed) version of Windows.

PS. Let's not forget that the SP1 release has done nothing to dampen the ill-will towards Vista either...

 

[Gilbo]

Vista SP1 still hasn't been made available on Windows Update for any of my computers .

 

[Fushigi]

Gilbo, one of my three Vista machines was in the group that got SP1 during the recent slip-up. To my knowledge MS is still on track to release SP1 via MS/Windows Update this month. And it should be available now to some or all MSDN subscribers.

Other than improved performance - my Windows Experience Index went up by .2 - I haven't really noticed any changes. It's still fast, stable, and works with my hardware & peripherals just fine.

 

[Bozo]

The only thing that I noticed after installing SP-1 was more bloat on the hard drive.


Bozo :joker:

 

[LiamC]

...For example, how difficiult would it have been to back-port a working USB driver to the earlier Win9x versions? A trivially easy task, seeing as the code was already written for Win ME. But they chose not to do it, as there was no commercial gain in it for them.
...

Ah no. The driver models (which is what you program to) were radically different in 95 OSR2, 98, 98SE and W2K. ME was a bastard case. Even XP's driver model is significantly different (1) from W2K and Vista is a whole new ball game. Don't get me started on 16 v 32 v 64 bit.

Yes it could have been done. Trivial it wasn't.

(1) A driver programmed to the W2K driver model will run on XP, but it's "not the done thing" (TM). If you do it, you run into not-signed driver issues and WHQL. It's possible, but not worth the bother.

98SE had reasonable USB support I thought, but the peripherals themselves weren't (in a lot of cases) up to scratch. I don't think I went fully USB until 2003, and by that time I was running XP (or was that 2004?).

 

[Clocker]

I'm dual booting my legit copy of Vista Ultimate x64 for a couple weeks now on my C2D machine with 4GB of RAM. I just downloaded and installed SP1:

http://www.dailytech.com/Windows+Vista+SP1+Now+Available+to+the+Public/article11138.htm

Now Vista thinks my computer has 18GB RAM. I think I will hold off on activating.

 

[mubs]

M$ fesses up: Vista feature (UAC) designed to 'annoy users'

 

[ddrueding]

At last they succeeded at something....

 

[Chewy509]

The other side of the table, is that UAC is simply more than a Prompt, it alters how Windows operates.

The number of applications that either fail to install or not work correctly with UAC enabled is quite large (yes even with full local admin rights, apps still break with UAC enabled). Turn off UAC and all of a sudden everything starts working again.

Or sales manager just got a new laptop with Vista on it, and I had to turn UAC off to make our ERP software work (MS Great Plains, if your interested). He had already had enough of UAC, especially when it prompted him to "Allow" him to delete a file 3 times. (The file was on the local file server).

OT: I went to a Windows 2008 Server Jumpstart 2day course last week, and I openly stated that I disliked Vista. This opened a class discussion on Vista, and the instructor asked why I didn't like Vista. Gave him about 8 reasons all with evidence (and demonstrations as most people on the course had Vista on their laptops), and funny enough he even agreed that they were all known faults with Vista and we had to live with them. (And yes UAC was one of them, not because of the annoying prompts, but because it broke software without good reason).

 

[Fushigi]

IMO UAC doesn't break software. It exposes poorly written software for the crap that it is. Software that stood a decent chance of failing on a properly hardened XP system. That some of that software would come from MS is no great surprise considering their heinous past (and Great Plains was an acquisition anyway).

 

[ddrueding]

IMO UAC doesn't break software. It exposes poorly written software for the crap that it is. Software that stood a decent chance of failing on a properly hardened XP system.

But on a hardened XP system, there was a significant degree of granularity to the controls and security. If program x was poorly written and required access to system file y, you could change just that bit. With UAC it is either on or off.

 

[Chewy509]

IMO UAC doesn't break software. It exposes poorly written software for the crap that it is.

Case scenario 1: Software updates. Most software use a secondary *.exe to auto-update software, which download the updated exe/dll/data from an internet based server, and then copy said updates files over the top on the files in C:\Program Files\<Application>\*. Even though the user account which the software updater runs may have full local admin rights, UAC will break this software by DENYING the updater from copying over the files even if you click "Allow".

Case scenario 2: UAC has been proven to re-route registry requests from HKLM to another location, which will break ~50% of applications if they don't have any fault-tolerance or self-repair system in place.

Case scenario 3: A piece of software I've written (an software development IDE) does not work with UAC active. The IDE doesn't use any registry entries, can run from any location, and will work fine when running as the "Guest" account on 2K/XP/2K3 even with DEP enabled for ALL applications. The only conclusion is that it's a .NET based application that calls some unmanaged code in a DLL (which was coded in 64bit Assembly for performance), and UAC doesn't like this aspect.

All the above 3 cases are from when the user account you are using has full local administrator rights.

 

[Fushigi]

Case 1: Changing a file is probably the issue. Delete the old file, then write the new file. Changing/overwriting programs is something that malware does, not something that properly designed applications do.

Case 2: No ability to recover from registry issues is easily a symptom of a poorly written application.

Case 3: Not using managed code is probably the issue, although that's a guess.

One of your three examples represents a potential UAC issue while the other two are symptomatic of poorly written applications/installers.

 

[Chewy509]

Case 1: Changing a file is probably the issue. Delete the old file, then write the new file. Changing/overwriting programs is something that malware does, not something that properly designed applications do.

So how are applications meant to update themselves? (The overwrite a file is really a delete, copy File I/O transaction).