Mac or PC ?

[Jamesds]

Which platform would I want to play civ on? To me it wouldn't matter as long as Civ ran nicely. If it was down to cash, then run Civ on a PC using windows or Linux, your experience will be the same, just a little cheaper then on a Mac.

Don't forget the time it takes to port Civ and it's expansions to the Mac - I couldn't wait for it to be ported, knowing I could use it had I only been using Windows.

I think the main thing in "which OS do you like" is personal preference. Everyone is different (thank goodness) and so we all feel more comfortable with a certain way of working, and these different operating systems appeal to different people, since they are worked differently. All the essays of arguments above written by Asher and Zetetic Apparat show this, like 'regedit' being unorganised, illogical and illegible to some, but to others it is structured logically and can be read and understood when needed.

I started my computer experience on an old 486, running Windows 3.1. Having since progressed through nearly all of the following Microsoft OS, I've been brought up to understand and be very familiar with Windows and the way it works. It makes sense to me.

That's my theory - anyone else stuck with the OS they started with?

 

[Asher]

It has to do with the registry, because when I finally did fix it, the weird use of paths involving drive letters to determine user profiles and such like required the use of the registry

Why, though? You can modify drive letters and user profile paths from a GUI.

But why doesn't regedit make it obvious that it's an alias?

I sincerely hope this a joke designed to bait me. This is hardly a startling innovation or insight.

Regedit is a last-resort tool -- a crude, debugging frontend to access the system registry that is intended to be written to, and read by, applications. Why would regedit waste time with niceities such as shortcut icons when they simply don't matter? Why waste the time when the user messing with regedit SHOULD know the basics of how it is set up and organized?

Screwing around with your registry is not unlike screwing around with your car engine. You should only do it if you know what you're doing. And even if you do fool around with it, you should not whine when there's not fluorescent lights built in with big, bold signs on where everything is -- you consult the documentation.

I'm just quoting the frontpage of the Microsoft Developer's Vista minisite:

When webmasters and marketing droids dictate technological evolution, I'd accept that as a passable counter.

To be fair, that's because any book title refers to it as "Programming Windows" or at most "WinAPI". Besides which, it's not really like anyone had to write another one after '98 or at most '02.

Have you seen the table of contents for the latest "Programming Windows" books?

Win32 should be considered obsolete. But it's not. Not even by teams developing within Microsoft.

Some teams at Microsoft (indeed, most) still deal with products built with Win32. You do not typically convert all of your existing code for massive products to a new API.

New Microsoft programs are in .NET, and slowly many main ones are being converts to .NET (eg, Office, Explorer). The use of Win32 still today means nothing, it's deprecated and obsolete in terms that new applications should not use it -- but existing ones still, obviously, use it.

 

[Zetetic Apparat]

Why, though? You can modify drive letters and user profile paths from a GUI.

OOI, which one?

Regedit is a last-resort tool -- a crude, debugging frontend to access the system registry that is intended to be written to, and read by, applications. Why would regedit waste time with niceities such as shortcut icons when they simply don't matter? Why waste the time when the user messing with regedit SHOULD know the basics of how it is set up and organized?

Every application should be built correctly and made as helpful as possible. It's not like this would have been a major task. (Although, given the multiple delays, I guess I have to forgive a little.)

Screwing around with your registry is not unlike screwing around with your car engine. You should only do it if you know what you're doing. And even if you do fool around with it, you should not whine when there's not fluorescent lights built in with big, bold signs on where everything is -- you consult the documentation.

Rubbish. This is an unnecessary analogy between physical and digital objects. You have comments in config files, right? I do not understand the mentality that "consulting the documentation" should be the normal procedure. Once basic concepts and the most simple of hard knowledge about a system have been learnt, everything else should follow intuitively. This can be the case a lot more easily with a computer system than with a car engine. Obviously somethings are just not simple enough for this to work, but most things really, really, really just need a decent UI. Most people can understand how most things work with good enough metaphors.

When webmasters and marketing droids dictate technological evolution, I'd accept that as a passable counter.

I'm not sorry for expecting prominent parts of MSDN to bear some relationship to the technical truth. Where the hell am I supposed to get information? Conceptual Tech Docs claim that "The .NET Framework can be hosted by unmanaged components that load the common language runtime into their processes and initiate the execution of managed code", implying that unmanaged code doesn't use .NET itself.

Have you seen the table of contents for the latest "Programming Windows" books?

The 5th and final edition (1998) is, not surprisingly, entirely devoted to Win32.

Some teams at Microsoft (indeed, most) still deal with products built with Win32. You do not typically convert all of your existing code for massive products to a new API.

Apple, Cocoa & Carbon, 1999-2001. As much Cocoa as possible. Having said that, Apple also did the sensible (but obviously risky and in some ways damaging) thing and relegated the old Macintosh Toolbox to a VM environment.

Nevertheless, I don't understand why by the time that .NET 3.0 and Vista were rolled out, stuff like Office and Explorer don't use .NET; why XP did not have .NET forced into it by a service pack or Windows Update and why Win32 was not properly deprecated in Vista. Backwards compatibility could have been maintained (as required by Microsoft's customers; they did a demo where VisiCalc ran under Vista which is impressive but revolting). Presumably this partly results from Vista not actually being a rewrite...

This would have been possible, and it would have been good for Windows. See, I understand that without backwards compatibility being perfect (and 'Classic' for Mac OS X was not an absolutely, perfectly complete solution but it was pretty good) then Vista becomes a lot less attractive to business. But sooner or later, MS has to spin off the older APIs (if they truly are destined to die) into a VM environment, right? And Vista was the most sensible time to do this to my mind. Yes, it would slow biz adoption, but it has to happen sometime...

New Microsoft programs are in .NET, and slowly many main ones are being converts to .NET (eg, Office, Explorer). The use of Win32 still today means nothing, it's deprecated and obsolete in terms that new applications should not use it -- but existing ones still, obviously, use it.

Fine, it's not like I totally disbelieve you. But where is the statement from MS? I simply cannot find it.

 

[Speedo]

OOI, which one?

diskmgmt.msc
lusrmgr.msc

 

[Asher]

Every application should be built correctly and made as helpful as possible. It's not like this would have been a major task. (Although, given the multiple delays, I guess I have to forgive a little.)

And there should also be world peace. Pick your battles.

Rubbish. This is an unnecessary analogy between physical and digital objects. You have comments in config files, right? I do not understand the mentality that "consulting the documentation" should be the normal procedure.

Not all config files have comments...even then, such comments are known as "documentation". Extensive regedit documentation is available. Only the most extreme powerusers should modify the registry themselves, and in such cases, documentation is ample on how to do so. Wanting a UI that encourages anyone and everyone to toy with it is a very bad thing. Ditto for a car engine. It's a perfect analogy...

Once basic concepts and the most simple of hard knowledge about a system have been learnt, everything else should follow intuitively.

This is true about using the system. This is not true about modifying the system.

It's one thing to know how to use a car, it's another thing to know about how to modify it. Again, the analogy is sound.

This can be the case a lot more easily with a computer system than with a car engine. Obviously somethings are just not simple enough for this to work, but most things really, really, really just need a decent UI. Most people can understand how most things work with good enough metaphors.

The problem is what the registry actually DOES is far more complicated than the interface it currently has. Making the interface pretty and attractive does nothing to actually improve the usefulness of regedit. What it does do is perhaps encourage people who should not be meddling with it, to meddle with it. Next thing you know you've got people changing an innocent DWORD in their registry ("DWORD? What a complicated term! Who cares that it is the proper term to refer to the value as that's how it's stored by the applications who use the registry, let's call it something nice and generic -- perhaps with a picture!"). Such an innocent change can FUBAR their whole installation.

I'm not sorry for expecting prominent parts of MSDN to bear some relationship to the technical truth. Where the hell am I supposed to get information? Conceptual Tech Docs claim that "The .NET Framework can be hosted by unmanaged components that load the common language runtime into their processes and initiate the execution of managed code", implying that unmanaged code doesn't use .NET itself.

That's not the implication at all. This is just Logic 101.

The sentence you quoted is a very, very high-level overview for a summary webpage. It doesn't encompass nuances or details, otherwise it'd be rather difficult to comprehend. If you've attended any of MS' developer evangelism meetings, or been a subscribed to MSDN, it'd be a much different discussion right now.

The 5th and final edition (1998) is, not surprisingly, entirely devoted to Win32.

Actually, there are much more recent versions. For example, "Programming Windows (with C#)". Same author as the 1998 edition and all the ones before it, but about .NET -- not Win32. There've been no book released from MS about Win32 programmming since 1998 -- 9 years ago. There've been plenty on .NET. Is that not a sign?

Apple, Cocoa & Carbon, 1999-2001. As much Cocoa as possible. Having said that, Apple also did the sensible (but obviously risky and in some ways damaging) thing and relegated the old Macintosh Toolbox to a VM environment.

Sensible? I strongly disagree. Using MacOS was a nightmare for years as apps transitioned. Their userbase declined during the period. If Windows did what Apple did, it'd be a disaster on a much larger scale due to the MANY more applications and MANY more users. Apple's solution was quick, cheap, and lazy -- sensible, I think not.

Nevertheless, I don't understand why by the time that .NET 3.0 and Vista were rolled out, stuff like Office and Explorer don't use .NET;

Are you serious? Here's a thought: They're MASSIVE an Win32 and .NET are virtually nothing alike? For Office, switching to .NET fully (note: Office supports .NET extensively for programming) is a HUGE, HUGE undertaking. Work was underway for Vista to include a .NET explorer, but this work was not progressing fast enough to be included in Vista. It'll likely come in a service pack, or in Vista's successor. Why do you think Vista was delayed and then work kickstarted? Everything was going to .NET but it was an incredibly huge burden to do in any reasonable timeframe for the next OS release, so it was postponed.

why XP did not have .NET forced into it by a service pack or Windows Update

Err. It did. Do your homwork.

and why Win32 was not properly deprecated in Vista. Backwards compatibility could have been maintained (as required by Microsoft's customers; they did a demo where VisiCalc ran under Vista which is impressive but revolting). Presumably this partly results from Vista not actually being a rewrite...

Define "properly deprecated"? Win32 works as a backwards compatibility tool. MS no longer advocates it or guarantees its presence in future versions of Windows -- don't bother asking me to google that, they've said it repeatedly in MSDN newsletters and in conferences. Vista is a transitional OS from Win32 to .NET, both work, but new apps should be in .NET since Win32 won't be around forever.

But sooner or later, MS has to spin off the older APIs (if they truly are destined to die) into a VM environment, right? And Vista was the most sensible time to do this to my mind. Yes, it would slow biz adoption, but it has to happen sometime...

It did happen in Vista. Try using Vista 64-bit (the future base of new operating system). It breaks off all 16-bit compatibility. 32-bit compatibility will stay for quite some time -- because, let's face it, the reason MS is where it is instead of Apple was Apple was way too stupid to understand the importance of backwards compatibility. That's why they're a niche player now.

Fine, it's not like I totally disbelieve you. But where is the statement from MS? I simply cannot find it.

- Subscribe to MSDN
- Attend a developer conference
- Try to find a recent Win32 book vs a recent .NET book

 

[Asher]

diskmgmt.msc
lusrmgr.msc

Screenshots for our non-Vista using friends.

As I said, there should be no reason for you to be modifying the system registry. It's a "do not touch" area, like your car engine. If you want to drive it, use the interfaces provided to you in the cabin.

 

[Zetetic Apparat]

Firstly, management console snap-ins don't work if your drive letters are fubar'd (or at least that's how Vista decides to see it). MMC would start but they'd simply refuse to load. I have no idea why, but then again I have no idea why user profiles are linked to drive letters either. I'm sure there's a point to the whole drive letters system but I've yet to work it out...

And there should also be world peace. Pick your battles.

I can more easily achieve a system with helpful interfaces that world peace. I hope. Otherwise I've been wasting my time, one way or another.

Not all config files have comments...even then, such comments are known as "documentation". Extensive regedit documentation is available. Only the most extreme powerusers should modify the registry themselves, and in such cases, documentation is ample on how to do so. Wanting a UI that encourages anyone and everyone to toy with it is a very bad thing. Ditto for a car engine. It's a perfect analogy...

Comments are an example of self-documentation. Metaphors in GUIs count as self-documentation. Everything that I have ever changed in the registry has been by working out the meaning of the keys and whatever. It's not difficult (since I'm not massively intelligent or divinely inspired) and nor is regedit's interface, by-and-large even if it is remarkably unpolished.

This is true about using the system. This is not true about modifying the system.

It's one thing to know how to use a car, it's another thing to know about how to modify it. Again, the analogy is sound.

No, that's a spurious line to draw. Every app should have a helpful interface.

The problem is what the registry actually DOES is far more complicated than the interface it currently has. Making the interface pretty and attractive does nothing to actually improve the usefulness of regedit. What it does do is perhaps encourage people who should not be meddling with it, to meddle with it. Next thing you know you've got people changing an innocent DWORD in their registry ("DWORD? What a complicated term! Who cares that it is the proper term to refer to the value as that's how it's stored by the applications who use the registry, let's call it something nice and generic -- perhaps with a picture!"). Such an innocent change can FUBAR their whole installation.

That's not analogous to making it clear that something is a link. I'm prepared to go and mess about with stuff in the registry (say changing paths for where an app is installed, again, I have no idea why some apps need to store stuff like this in the registry, but they seem to). I won't touch things that I can't work out what they are. However, I can often work out things I need to alter and anything which helps me get there is appreciated. I don't consider myself a (Windows, at the very least) 'poweruser' (and you complain about marketing droids ) but I'm prepared to go an change things that are very slightly the under the hood.

That's not the implication at all. This is just Logic 101.

The sentence you quoted is a very, very high-level overview for a summary webpage. It doesn't encompass nuances or details, otherwise it'd be rather difficult to comprehend. If you've attended any of MS' developer evangelism meetings, or been a subscribed to MSDN, it'd be a much different discussion right now.

Quite probably. I don't often come across developer documentation that doesn't at least mention in passing that it is actually lying. It quite normal to point out that something is inaccurate outline as far as I am concerned. So I assume that most things in technical documentation are accurate.

Actually, there are much more recent versions. For example, "Programming Windows (with C#)". Same author as the 1998 edition and all the ones before it, but about .NET -- not Win32. There've been no book released from MS about Win32 programmming since 1998 -- 9 years ago. There've been plenty on .NET. Is that not a sign?

Oh, sure, but that's got brackets and everything. The sign as far as I can tell is that programming with the Win32 API under Windows hasn't fundamentally changed massively since 1998.

Sensible? I strongly disagree. Using MacOS was a nightmare for years as apps transitioned. Their userbase declined during the period. If Windows did what Apple did, it'd be a disaster on a much larger scale due to the MANY more applications and MANY more users. Apple's solution was quick, cheap, and lazy -- sensible, I think not.

Apple's solution was not entirely elegant by any means, effectively involving the release of two public betas before 10.1 arrived and was vaguely usable as a day-to-day operating system. However, I think this is better than the release of Vista which involved an excessive amount of feature chopping and so forth to try to force out a semi-ready product to sell to businesses. (As a sidenote, I doubt that the Macintosh userbase declined in the period 1998-2001. It's barely possible for one thing... )

Are you serious? Here's a thought: They're MASSIVE an Win32 and .NET are virtually nothing alike? For Office, switching to .NET fully (note: Office supports .NET extensively for programming) is a HUGE, HUGE undertaking. Work was underway for Vista to include a .NET explorer, but this work was not progressing fast enough to be included in Vista. It'll likely come in a service pack, or in Vista's successor. Why do you think Vista was delayed and then work kickstarted? Everything was going to .NET but it was an incredibly huge burden to do in any reasonable timeframe for the next OS release, so it was postponed.

I simply can't believe that this is the case. Microsoft has resources greater than small countries if it wants to use them.

Err. It did. Do your homwork.

Sorry, it's late. Then there's no excuse. XP should have been the transition system. Vista should have been what Longhorn was going to be.

Define "properly deprecated"? Win32 works as a backwards compatibility tool. MS no longer advocates it or guarantees its presence in future versions of Windows -- don't bother asking me to google that, they've said it repeatedly in MSDN newsletters and in conferences. Vista is a transitional OS from Win32 to .NET, both work, but new apps should be in .NET since Win32 won't be around forever.

I've tried Googling for Microsoft statements on the obsolescence or deprecation of Win32 and I can't find anything. I suspect that I don't know what MS marketing vocab for 'deprecated' is.

It did happen in Vista. Try using Vista 64-bit (the future base of new operating system). It breaks off all 16-bit compatibility. 32-bit compatibility will stay for quite some time -- because, let's face it, the reason MS is where it is instead of Apple was Apple was way too stupid to understand the importance of backwards compatibility. That's why they're a niche player now.

Less of a niche player than they were in 1999.

The reason why Microsoft understands backwards compatibility is because it feels it has to. I can't risk making a release which it can't immediately push on businesses. I don't understand that. Profitability can't really be a concern, nor can installed userbase. It seems unnecessary when Vista could have been a mucher cleaner break with WinAPI and presumably a much cleaner codebase. I think that it is a real shame that Vista has turned out to be far less than it should have been. It, and .NET, is putting some pressure on Apple to do better with Mac OS X, and Cocoa, but not nearly the amount that it should have done. The UI is improved, but I still find it jarring a lot more often than I do Mac OS X's.

 

[Asher]

Firstly, management console snap-ins don't work if your drive letters are fubar'd (or at least that's how Vista decides to see it).

Neither does launching regedit.exe...

MMC would start but they'd simply refuse to load. I have no idea why, but then again I have no idea why user profiles are linked to drive letters either. I'm sure there's a point to the whole drive letters system but I've yet to work it out...

Driver letters are simply mounting points. No different than /dev/hdb1, really, but with a simpler name. Of course, there are lots of unlabeled shortcuts you hate that direct, say, /home to something in /dev/hdb1 too.

No, that's a spurious line to draw. Every app should have a helpful interface.

Agree. However, some apps which are not supposed to be used (nay...dangerous to use) should not become user friendly to the point of encouraging people who should not be using it, to use it. My grandmother should not touch regedit under any circumstances, and the second you try to pretty it up with a "helpful" interface the more likely she is to try it -- and screw up her computer, subsequently.

That's not analogous to making it clear that something is a link.

Yes, it is. In cars, there are no diagrams posted identifying which cable goes where. The analogy is still sound.

I'm prepared to go and mess about with stuff in the registry (say changing paths for where an app is installed, again, I have no idea why some apps need to store stuff like this in the registry, but they seem to).

This is a pretty obvious use-case, isn't it? Back in the days before the registry was used, when you applied a patch, for instance, your entire HDD was searched for blah.exe, or you had to find it yourself. When the install path is stored in the registry, a simple function call returns IF the program is installed and WHERE it is installed. Vastly simplified application maintenance.

I won't touch things that I can't work out what they are. However, I can often work out things I need to alter and anything which helps me get there is appreciated. I don't consider myself a (Windows, at the very least) 'poweruser' (and you complain about marketing droids ) but I'm prepared to go an change things that are very slightly the under the hood.

I'll tell you this much: manually changing install paths in the registry is a dumb idea because you'll miss one or two and suddenly you're loading bad DLLs and have an unstable system. Again, another instance of the user meddling with the registry when he/she should not.

Quite probably. I don't often come across developer documentation that doesn't at least mention in passing that it is actually lying. It quite normal to point out that something is inaccurate outline as far as I am concerned. So I assume that most things in technical documentation are accurate.

That webpage is in no way a technical document. It's an overview that provides access to technical documents.

For what it's worth, most technical documents have countless errors, ambiguities, and omissions -- always assuming the documentation is correct is not a very good idea either. Especially with technology and its rate of change.

Oh, sure, but that's got brackets and everything. The sign as far as I can tell is that programming with the Win32 API under Windows hasn't fundamentally changed massively since 1998.

Those fundamental changes have happened, which is precisely why the new system is not called Win32 now.

Apple's solution was not entirely elegant by any means, effectively involving the release of two public betas before 10.1 arrived and was vaguely usable as a day-to-day operating system. However, I think this is better than the release of Vista which involved an excessive amount of feature chopping and so forth to try to force out a semi-ready product to sell to businesses.

Well, let's see here. Vista worked out of the box with a more-than-competent predecessor, Apple had an unusable operating system for several years and a completely obsolete and painful predecessor. I don't think that's a solid argument.

(As a sidenote, I doubt that the Macintosh userbase declined in the period 1998-2001. It's barely possible for one thing... )

Marketshare steadily declined until ~2003. The revitalization has more to do with iPod than OS X.

I simply can't believe that this is the case. Microsoft has resources greater than small countries if it wants to use them.

Spoken like a true manager. Throw resources at an engineering problem and it goes away, right? Nothing is farther from the truth. There's been no software engineering project in history as massive as Windows Vista was. They were essentially breaking new ground on many fronts, from experimental technology to experimental development environments to experimental software engineering techniques. After a couple years of slow progress they re-evaluated what was working and what wasn't, and chopped it up. This is completely unrelated to lack of resources or how much resources they had. There are some excellent software engineering whitepapers that've come out of the Vista development process that are actually being integrated into upcoming textbooks for software engineers.

It was not a resource problem.

Sorry, it's late. Then there's no excuse. XP should have been the transition system. Vista should have been what Longhorn was going to be.

XP is a transition system. Why do you think .NET was included in Windows XP's service packs? Vista is a further transition system: it not only added the .NET technologies, but the new Windows-specific APIs such as WPF, WCF, WWF, WCS, etc. The whole of these is what makes .NET 3.0 and the full Win32 replacement.

Again, I'm not sure if you're a technologist, but you can't replace the whole hog at the same time without a jarring hiccup. That's what OS 9 - OS X was for years, and it cost Apple billions of dollars in lost revenue and lost potential clients.

This whole discussion is absurd. You're talking as if Windows Vista is a failure in comparison to MacOS X, when MacOS X's transition period is effectively a "what not to do" business decision (along with their infamous 68k -> PPC switch), while Vista's approach is more gradual and doesn't result in the loss of users.

Less of a niche player than they were in 1999.

That's beside the point, their marketshare shrank in 1999, 2000, 2001 etc as OS X was a joke and OS 9 was archaic.

The reason why Microsoft understands backwards compatibility is because it feels it has to. I can't risk making a release which it can't immediately push on businesses. I don't understand that. Profitability can't really be a concern, nor can installed userbase. It seems unnecessary when Vista could have been a mucher cleaner break with WinAPI and presumably a much cleaner codebase.

This is just absurd. Simply put, you apparently don't have the background to make an educated argument here.

There is ZERO business sense, ZERO technological sense for Microsoft to make a "clean break" in APIs with Windows. What's the best way to lose your userbase? Cut off any tie to its predecessor. This is just common sense, I thought.

What happens when you make clean breaks? With the arguably superior MacOS back in the 80s, Apple turned out to be a 4% marketshare niche player to Windows' 95% marketshare behemoth. I think that alone tells you why "clean breaks" in APIs are a bad idea.

I think that it is a real shame that Vista has turned out to be far less than it should have been. It, and .NET, is putting some pressure on Apple to do better with Mac OS X, and Cocoa, but not nearly the amount that it should have done. The UI is improved, but I still find it jarring a lot more often than I do Mac OS X's.

The problem with Vista is it's real changes are obscured because most apps are still Win32. Ignorant users, early in Vista's lifecycle, don't see anything but a new interface and hardware incompatibilities.

When you start using the .NET 3.0 technology stack you start seeing some amazing technology that simply wasn't available before on MacOS X or Linux or Windows XP. Developers start seeing how much easier it is to develop their applications. Gamers start playing DirectX 10 games with far less system overhead. Consumers can watch HD-DVD and Bluray movies on their computer (which they cannot do, legally, on MacOS X or Linux). Multi-core users have access to the world's most advanced multi-threaded network stacks and schedulers. The system is FAR more secure by design than MacOS X or Linux (I'm willing to discuss why in another thread if people care). There's a whole stack of stuff in Vista that is not visible to users yet -- but it will be. Win32 applications in Vista are essentially just like running MacOS 9 in MacOS X, but in a more elegant way. The changes in Vista come when .NET 3.0 is used, just like how in OS X they came when Cocoa was used.

Right now, Vista to most users is nothing but a prettier XP. And considering XP's market penetration, I think you'd have a tough time arguing that's actually a bad thing. In the future, it'll be far more than XP.

Apple actually has a lot of catching up to do, and they know it -- users don't (yet).

 

[Zetetic Apparat]

Neither does launching regedit.exe...

Apart from Regedit worked. Which certainly does strike me as weird.

Driver letters are simply mounting points. No different than /dev/hdb1, really, but with a simpler name. Of course, there are lots of unlabeled shortcuts you hate that direct, say, /home to something in /dev/hdb1 too.

/dev/hdb1 is never a mount point, it's a file representation of a device. '/' is a mount point, '/mnt/myDrive' is a mount point. The drive letter thing only seems crazy to me when a system can't boot even itself properly because it's assigned it's own boot drive a different letter.

Agree. However, some apps which are not supposed to be used (nay...dangerous to use) should not become user friendly to the point of encouraging people who should not be using it, to use it. My grandmother should not touch regedit under any circumstances, and the second you try to pretty it up with a "helpful" interface the more likely she is to try it -- and screw up her computer, subsequently.

So you're saying that some interfaces should be intentionally crippled to reduce usability and prevent people from using it? The point is interesting (and not wholly without merit) but largely... stupid in my opinion. It's not like regedit automatically starts every so often as a test to whether the user is going to touch it or nor. I certainly haven't seen your suggestion in many HIG documents...

Yes, it is. In cars, there are no diagrams posted identifying which cable goes where. The analogy is still sound.

It's not a car. The possibility for hinting and self-documentation goes a long way. Even in a car, the wires will be colour-coloured (pre-supposing a knowledge of what the colours mean; in a computer UI, you can actually make it more obvious). Obviously, the entire system can't be self-documenting, but they could at least make an effort. To me, documentation indicates failure, a failure to design a decent interface. Obviously that's an absurd stance taken to its practical conclusion, but it's still the point of view that any interface designer should take.

This is a pretty obvious use-case, isn't it? Back in the days before the registry was used, when you applied a patch, for instance, your entire HDD was searched for blah.exe, or you had to find it yourself. When the install path is stored in the registry, a simple function call returns IF the program is installed and WHERE it is installed. Vastly simplified application maintenance.

Or the system uses caching as regards bundle IDs (whenever the bundle is loaded, cache its location) so you just call a function asking for the correct bundle ID.

I'll tell you this much: manually changing install paths in the registry is a dumb idea because you'll miss one or two and suddenly you're loading bad DLLs and have an unstable system. Again, another instance of the user meddling with the registry when he/she should not.

Few problems so far. Bad DLLs? I presume you mean something along the lines of apps not loading different versions to system versions. Urgh, what a mess...

That webpage is in no way a technical document. It's an overview that provides access to technical documents.

Conceptual documents count as technical documents to my mind. I don't expect them to lie still.

For what it's worth, most technical documents have countless errors, ambiguities, and omissions -- always assuming the documentation is correct is not a very good idea either. Especially with technology and its rate of change.

Yeah, I know, my father is technical author. But this is a pretty important thing they're describing. I expect a certain level of veracity.

Well, let's see here. Vista worked out of the box with a more-than-competent predecessor, Apple had an unusable operating system for several years and a completely obsolete and painful predecessor. I don't think that's a solid argument.

The problem is see is that this will continue. Windows 7 will sacrifice more and more features and it will still support Win32 natively. They'll say it won't, like they did with Longhorn, but eventually management will cave, and they'll do it.

Marketshare steadily declined until ~2003. The revitalization has more to do with iPod than OS X.

Well, except for when it rose from 2001-2002 (Q1) to be the first example in Google.

Spoken like a true manager. Throw resources at an engineering problem and it goes away, right? Nothing is farther from the truth. There's been no software engineering project in history as massive as Windows Vista was. They were essentially breaking new ground on many fronts, from experimental technology to experimental development environments to experimental software engineering techniques. After a couple years of slow progress they re-evaluated what was working and what wasn't, and chopped it up. This is completely unrelated to lack of resources or how much resources they had. There are some excellent software engineering whitepapers that've come out of the Vista development process that are actually being integrated into upcoming textbooks for software engineers

It was not a resource problem.

I'm not claiming Vista was. Office 12, however? That could have been written in .NET.

XP is a transition system. Why do you think .NET was included in Windows XP's service packs? Vista is a further transition system: it not only added the .NET technologies, but the new Windows-specific APIs such as WPF, WCF, WWF, WCS, etc. The whole of these is what makes .NET 3.0 and the full Win32 replacement.

And Windows 7 will be a transition system again. And this is sad.

Again, I'm not sure if you're a technologist, but you can't replace the whole hog at the same time without a jarring hiccup. That's what OS 9 - OS X was for years, and it cost Apple billions of dollars in lost revenue and lost potential clients.

Rubbish. It was not smooth but it did not damage Apple, as a company, in any way.

This whole discussion is absurd. You're talking as if Windows Vista is a failure in comparison to MacOS X, when MacOS X's transition period is effectively a "what not to do" business decision (along with their infamous 68k -> PPC switch), while Vista's approach is more gradual and doesn't result in the loss of users.

Vista is not a failure by Microsoft's standards. It will sell (eventually )and they'll push it on businesses. What Microsoft could have achieved, what I'm sure the Longhorn team thought they were going to be allowed to achieve, would have been much, much, more impressive technologically. Right now Vista adoption is small. It would be even smaller, I admit, adding software incompatibility to the semi-temporary hardware incompatibilities.

This is just absurd. Simply put, you apparently don't have the background to make an educated argument here.

There is ZERO business sense, ZERO technological sense for Microsoft to make a "clean break" in APIs with Windows. What's the best way to lose your userbase? Cut off any tie to its predecessor. This is just common sense, I thought.

What happens when you make clean breaks? With the arguably superior MacOS back in the 80s, Apple turned out to be a 4% marketshare niche player to Windows' 95% marketshare behemoth. I think that alone tells you why "clean breaks" in APIs are a bad idea.

It makes perfect technological sense: members of the Longhorn team bemoan the horrible amounts of copy-pasta Win32 code that exist in Vista. They didn't want it there and it's obviously not healthy to have it there.
It makes some business sense. Windows is not going to be dislodged from an installed base of the kind it has now, in the next half-a-decade. Vista would not sell as well, correct. However, it would provide a much more competitive and sensible base for future development.

(Sidenote: DOS and then Windows crushed Mac OS because they were cheaper (lack of licensing & poor use of funds from Apple II success), and in Windows' case much better for use in corporate environments. There were no clean breaks with APIs in the 80s. 68k-PPC should have really only involved rewrites where assembler was involved (with FAT binaries allowing dual-compatibility)).

The problem with Vista is it's real changes are obscured because most apps are still Win32. Ignorant users, early in Vista's lifecycle, don't see anything but a new interface and hardware incompatibilities.

When you start using the .NET 3.0 technology stack you start seeing some amazing technology that simply wasn't available before on MacOS X or Linux or Windows XP. Developers start seeing how much easier it is to develop their applications. Gamers start playing DirectX 10 games with far less system overhead. Consumers can watch HD-DVD and Bluray movies on their computer (which they cannot do, legally, on MacOS X or Linux). Multi-core users have access to the world's most advanced multi-threaded network stacks and schedulers. The system is FAR more secure by design than MacOS X or Linux (I'm willing to discuss why in another thread if people care). There's a whole stack of stuff in Vista that is not visible to users yet -- but it will be. Win32 applications in Vista are essentially just like running MacOS 9 in MacOS X, but in a more elegant way.

I believe you. Which is why, I'd rather that Win32 applications were spun off and away. Except massive numbers of system apps are still Win32-based, not just 3rd party apps, not just Microsoft apps.

The changes in Vista come when .NET 3.0 is used, just like how in OS X they came when Cocoa was used.

I don't think you mean that comparison. (Since many things are still written in Carbon and feel like Mac OS X apps rather than Classic/Toolbox apps.) Carbon isn't as to Win32 as Cocoa is to .NET; that's what you've been trying to convince me of and I believed you.

So, you're essentially saying that when all the Win32 stuff in Vista is actually dead then Windows will be much more impressive for it.
Uh.. Snap.

I'm sorry if I'm being stupid. I do not have the same background as you, I am relatively ignorant. I will admit that. But I am not totally absolutely ignorant.

 

[Asher]

Apart from Regedit worked. Which certainly does strike me as weird.

That strikes me as implausible as well, seeing as both launching an .msc and launching an .exe both are essentially launching a command from the boot drive. Even if this did happen, it's such an obscure scenario that it's not even worth discussing in general terms.

/dev/hdb1 is never a mount point, it's a file representation of a device. '/' is a mount point, '/mnt/myDrive' is a mount point.

Thank you for the semantical clarity, but you get the idea. C:\ is no different than /home/. They are functionally the same. Hell, you can use C:\ to be root "\" and then mount harddrives as folders "C:\home" if you really want to emulate *nix. Windows supports that.

There's nothing inherently puzzling of inferior about drive letters, they're the same idea behind mount points in *nix.

The drive letter thing only seems crazy to me when a system can't boot even itself properly because it's assigned it's own boot drive a different letter.

I think you should really give this example a rest. Aside from being unverifiable, it's a ridiculously obscure issue that doesn't do us any good in general discussions -- if your C:\ mounting failed, it'd be no different than if your "/" mounting failed in *nix. You're drawing a big distinction here where there is not one, at least with NTFS.

So you're saying that some interfaces should be intentionally crippled to reduce usability and prevent people from using it? The point is interesting (and not wholly without merit) but largely... stupid in my opinion.

Crippling is taking a functional interface and disabling it. That's an entirely different case of having a functional tool intended for developers. The use case for regedit is developers who have programs using the registry, therefore they don't need to be handheld into its use. It's not a matter of crippling it, it's a matter of overcomplicating and wasting resources prettifying a tool that should not be used by most people. You'll find that most developer tools are not pretty and grandmother-friendly, because that is not the point of them. And make no mistake, regedit is a developer tool.

It's not a car. The possibility for hinting and self-documentation goes a long way. Even in a car, the wires will be colour-coloured (pre-supposing a knowledge of what the colours mean; in a computer UI, you can actually make it more obvious).

On my car (and '04 model), the wires are all encased in black plastic cabling.

Obviously, the entire system can't be self-documenting, but they could at least make an effort. To me, documentation indicates failure, a failure to design a decent interface. Obviously that's an absurd stance taken to its practical conclusion, but it's still the point of view that any interface designer should take.

No, this isn't true. Interfaces are to be designed with the target audience in mind. When I design trading systems at work, there's no way to make it "intuitive" or "self-describing" because the application domain is very complex and specific. The users know how to use it because they know the application domain, and the interface makes sense to them.

The second you start "dumbing down" the interface to appeal to any and all clients that don't know what they're doing, you begin to add frustrations to the real user of the system. While some minor changes like adding a shortcut overlay to HKEY_CURRENT_USER to indicate it's a shortcut obviously don't overcomplicate things, they are not necessary either. Again, the user toying with regedit should know what he is editing before he goes to edit it. The registry is the easiest way to FUBAR your computer -- and I'm beginning to suspect the reason you got your obscure drive-letter error is you played with the registry.

Some components are meant to be foreboding to casual users while simple and intuitive to its designed users -- such is the case with regedit.

Or the system uses caching as regards bundle IDs (whenever the bundle is loaded, cache its location) so you just call a function asking for the correct bundle ID.

This sounds like a great idea. What could we call such a system? How about the system registry?

Few problems so far. Bad DLLs? I presume you mean something along the lines of apps not loading different versions to system versions. Urgh, what a mess...

Yeah, incompatible DLLs. The registry is not meant to be handled by hand except as a last resort when a program error prevents you from modifying it as it was intended. This is yet another reason why users should leave it well alone.

Conceptual documents count as technical documents to my mind. I don't expect them to lie still.

No one is saying they were lying -- just that in the one-sentence overview it was overly simplistic. Which is the purpose of overviews. They are not unequivocal truths.

The problem is see is that this will continue. Windows 7 will sacrifice more and more features and it will still support Win32 natively. They'll say it won't, like they did with Longhorn, but eventually management will cave, and they'll do it.

Ahhh, sacrifice features? I think this is why your argument is baffling to me. Which features were sacrificed so Windows 2000, XP, and Vista could handle Win32?

I'm not claiming Vista was. Office 12, however? That could have been written in .NET.

What? How do you figure? This is absurd. Do you know how huge Office 12 is? Do you know how old it is? Do you know how massive an undertaking it is to rewrite it to use a new API?

Parts of Office 12 do use .NET. The new parts. The new GUI, for instance, is not Win32-based. There are massive new programming APIs in Office 12 for .NET use. You can embed .NET programs inside Office applications. Office 12 is partially .NET but not entirely, because that is a massive undertaking.

And Windows 7 will be a transition system again. And this is sad.

Rubbish. It was not smooth but it did not damage Apple, as a company, in any way.

Rubbish? Absurd. No other word for that. OS X' "transition period" was marked by flagging sales and shrinking marketshare. Did not damage Apple in any way? Hardly.

Vista is not a failure by Microsoft's standards. It will sell (eventually )and they'll push it on businesses. What Microsoft could have achieved, what I'm sure the Longhorn team thought they were going to be allowed to achieve, would have been much, much, more impressive technologically. Right now Vista adoption is small. It would be even smaller, I admit, adding software incompatibility to the semi-temporary hardware incompatibilities.

This is an interesting point. Vista did make a clean break in some areas from XP and before -- notably the driver model. There are far less drivers than applications people are used to, and this has caused enough of a headache already.

Unfortunately, from a security standpoint, there was no way to enforce backwards compatibility with drivers.

It makes perfect technological sense: members of the Longhorn team bemoan the horrible amounts of copy-pasta Win32 code that exist in Vista. They didn't want it there and it's obviously not healthy to have it there.

I'll stop you right here.
1) Who said they didn't want it there? Without it there, adoption would be abysmal. They know that. They know their old programs would cease working without a suitable replacement ready. They know that.
2) "Obviously not healthy to have it there" -- huh? This makes no sense either.

It makes some business sense. Windows is not going to be dislodged from an installed base of the kind it has now, in the next half-a-decade. Vista would not sell as well, correct. However, it would provide a much more competitive and sensible base for future development.

Nonsense. Without support for Windows applications of past, there is no incentive to use Windows Vista vs switching to OS X or Linux. In fact, OS X and Linux would have more complete application support than Windows. Marketshare would plummet, OR, the world would stay on XP forever which works with all of the programsthey need.

Secondly, your argument seems to rest on the case that Win32 being in Windows Vista makes it less competitive (again, absurd) and not a sensible base for future development (again, absurd). You're making arguments that all depend on really absurd statements, so none of them follow at all.

(Sidenote: DOS and then Windows crushed Mac OS because they were cheaper (lack of licensing & poor use of funds from Apple II success), and in Windows' case much better for use in corporate environments. There were no clean breaks with APIs in the 80s. 68k-PPC should have really only involved rewrites where assembler was involved (with FAT binaries allowing dual-compatibility)).

You're missing the point. It's much simpler.

Users bought and liked software they had. They bought a new Mac (PPC). None of their old software worked. They had to buy new software. Users balked at this (higher perpetual operating costs), and looked to Windows which makes an active effort to maintain backwards compatibility. Mac userbase plunges, Windows rises.

Same scenario happened in the OS X transition.

I believe you. Which is why, I'd rather that Win32 applications were spun off and away. Except massive numbers of system apps are still Win32-based, not just 3rd party apps, not just Microsoft apps.

Yep! Welcome to the real world. Massive number of applications are Win32 still because the vast majority of the world uses Win32 right now. This will change as most users adopt Vista and more developers leverage .NET 3.0 (which makes their lives easier and allows more types of programs to be made easily).

I don't think you mean that comparison. (Since many things are still written in Carbon and feel like Mac OS X apps rather than Classic/Toolbox apps.) Carbon isn't as to Win32 as Cocoa is to .NET; that's what you've been trying to convince me of and I believed you.

So, you're essentially saying that when all the Win32 stuff in Vista is actually dead then Windows will be much more impressive for it.
Uh.. Snap.

This is true, why is it a "snap"? Old programs behave like old programs, Vista enables new types of programs developed with ease. If we got rid of Win32 in Vista and instantly replaced it with .NET 3.0, it'd be much more impressive. Unfortunately, unlike Apple, Microsoft's situation is far more complicated due to the massive software library and userbase.

 

[Zetetic Apparat]

On my car (and '04 model), the wires are all encased in black plastic cabling.

And, yes that's how computer systems are going. (Free software, in a sense, reverse this but the bar to entry is high). My parents grew up with NASCOMs and Ataris where messing around was relatively easy. I, as someone growing up with Classic Mac OS and Windows 9X don't like this trend. I acknowledge that some of it is inevitable – computer systems have become much, much more complex – but I don't feel that people need to be any more discouraged from messing with software. If you do the wrong thing with your car, you could kill yourself or other people on the road, with software on your home computer (aside from data loss which is notable but something of a strawman if you were to cite it) the risks really aren't there. And in a managed environment, they could always been disabled.

Presumably, if you were working at compiler divisions at IBM, you're lucky enough to come a generation a lot near to valves than myself.

No, this isn't true. Interfaces are to be designed with the target audience in mind. When I design trading systems at work, there's no way to make it "intuitive" or "self-describing" because the application domain is very complex and specific. The users know how to use it because they know the application domain, and the interface makes sense to them.

I hate you. Well, not you specifically but people who think like you. I've spent two week this summer trying to fix up order management and procurement-order-implementation systems' interfaces at a telecom's company. I was actually meant to be testing and documenting them, but eventually I became sufficiently irritated by the attitude that "oh it's complex and specific; that's why we have training". By and large the fixes that I got implemented were insignificant in terms of manhours or difficulty (MVC and all that), but really worthwhile in terms of hinting what the user should and had to do. I know systems are too complex, for many there is no perfect UI; but you are actually saying that the UI to your trading systems' users is intuitive, but only to people with knowledge of the background subject, presumably of trading itself. That's not really equivalent to knowledge of Windows system design.

The second you start "dumbing down" the interface to appeal to any and all clients that don't know what they're doing, you begin to add frustrations to the real user of the system. While some minor changes like adding a shortcut overlay to HKEY_CURRENT_USER to indicate it's a shortcut obviously don't overcomplicate things, they are not necessary either. Again, the user toying with regedit should know what he is editing before he goes to edit it. The registry is the easiest way to FUBAR your computer -- and I'm beginning to suspect the reason you got your obscure drive-letter error is you played with the registry.

No, I fubared my boot records. My fault, I did something without thinking. No excuse for making it so damn hard to fix... (I wouldn't have risked my actions if I knew they were going to turn into such an effort.)
I don't see your claim that every task involved using the registry requires 5 textbooks and a 4 day trawl through documentation (apologies, hyperbole) before you go ahead. Most things on a computer, even system configuration, are reversible. Yes, I am aware there are things on every system, including in the registry, which will break booting and so on. But there are plenty of things that won't. If people don't want to risk taking out their computer, then don't go and deliberately search out regedit. It's not like it's easy to find.

This sounds like a great idea. What could we call such a system? How about the system registry?

Because this occurs automatically without process intervention or 'registration' under Mac OS X. If you move the bundle and then start it, then it will update the location. Hell, with the mdserver in Tiger, it may update bundle locations automatically without them even having to be started. If I move Civ IV on my harddrive, then it breaks patching. Which to me seems crazy. Add or Remove Programs is not a package manager but nor is Windows truly drag-and-drop. (Mac OS X on the other hand appears to be drag-and-drop and could be if they sorted out support files left by bundles, when bundles were deleted; there are problems on every system...)

Yeah, incompatible DLLs. The registry is not meant to be handled by hand except as a last resort when a program error prevents you from modifying it as it was intended. This is yet another reason why users should leave it well alone.

God, I still hate Windows' system of handling libraries.

No one is saying they were lying -- just that in the one-sentence overview it was overly simplistic. Which is the purpose of overviews. They are not unequivocal truths.

And that should be made clear as it is in most Apple documentation and most community-based documentation. Most people don't lie in their documentation without making it clear that they are doing so and why.

Ahhh, sacrifice features? I think this is why your argument is baffling to me. Which features were sacrificed so Windows 2000, XP, and Vista could handle Win32?

Time must have been devoted to the re-implementation of Win32 into Vista. Microsoft claimed for a significant period of Longhorn's development that it would not have Win32 compatibility. The turnaround must have cost significant manhours, many of which could have been focused elsewhere. Not all, I'm sure, are transferable, but some would have been.

What? How do you figure? This is absurd. Do you know how huge Office 12 is? Do you know how old it is? Do you know how massive an undertaking it is to rewrite it to use a new API?

Given the time involved for the MacBU, a relatively tiny team, to move Office from Toolbox to Carbon, a significant but not insurmountable one.

This is an interesting point. Vista did make a clean break in some areas from XP and before -- notably the driver model. There are far less drivers than applications people are used to, and this has caused enough of a headache already.

Unfortunately, from a security standpoint, there was no way to enforce backwards compatibility with drivers.

I know. And these breaks are good in my view but bad from the point of view of a business buyer. Hence, presumably, relatively low sales of Vista so far.

I'll stop you right here.
1) Who said they didn't want it there? Without it there, adoption would be abysmal. They know that. They know their old programs would cease working without a suitable replacement ready. They know that.
2) "Obviously not healthy to have it there" -- huh? This makes no sense either.

1. Various members of the development team in ~2002 said they were happy to be able to see it go. Not members of the marketing team. No, I don't have quotes, sorry. Google? (Which actually strikes me as weird, as that is before the abandonment of XP in favour of Server 2003 codebase...)
2. Because it's a ton more code to support and patch. Which, is likely to be tending towards a mess (because everything does unfortunately). Just as Toolbox was becoming impenetrable by Mac OS 9.

Nonsense. Without support for Windows applications of past, there is no incentive to use Windows Vista vs switching to OS X or Linux. In fact, OS X and Linux would have more complete application support than Windows. Marketshare would plummet, OR, the world would stay on XP forever which works with all of the programsthey need.

Aside from the fact that programs would be rewritten on XP, for .NET, and the become usable on Vista.

Secondly, your argument seems to rest on the case that Win32 being in Windows Vista makes it less competitive (again, absurd) and not a sensible base for future development (again, absurd). You're making arguments that all depend on really absurd statements, so none of them follow at all.

It does not make it less competitive, it does make it worse base for future development. Except that Windows 7 will (apparently) still have all the Win32 (and indeed Win16) APIs implemented. So it's a fantastic base if Microsoft never intends to actually replace Win32 with .NET.

You're missing the point. It's much simpler.

Users bought and liked software they had. They bought a new Mac (PPC). None of their old software worked. They had to buy new software. Users balked at this (higher perpetual operating costs), and looked to Windows which makes an active effort to maintain backwards compatibility. Mac userbase plunges, Windows rises.

Same scenario happened in the OS X transition.

68k->PPC transition was relatively smooth, given the ease of porting applications over. The low userbase was due to the high cost of Macintoshes compared to IBM-compatibles, not accessory costs involved with software.

This is true, why is it a "snap"? Old programs behave like old programs, Vista enables new types of programs developed with ease. If we got rid of Win32 in Vista and instantly replaced it with .NET 3.0, it'd be much more impressive. Unfortunately, unlike Apple, Microsoft's situation is far more complicated due to the massive software library and userbase.

"Snap" because I agree. It's just that I wish Microsoft would have taken that jump. Yes, there would have been a risk involved, but it would have produced a better OS for me. Do you really believe that breaking backwards compatibility would have really users to other OSs in significant numbers? Do you really think it would have impacted the profitability of Microsoft (particularly if they kept selling XP)?

You obviously believe that making Vista incompatible with Win32 apps, or spinning it off into a VM (sacrificing some compatibility), would have made it unsellable. I believe that given it's currently poor (but far from abysmal) sales numbers, it would make more sense to have continued selling XP while providing a considerably better operating system.

(Oh, and I've managed to find statements on Vista, Win32 and .NET 3.0, by referring to them as Longhorn, Win32 and WinFX. You definitely seem to be right about WinFX being a better name...)

 

[Asher]

I hate you. Well, not you specifically but people who think like you. I've spent two week this summer trying to fix up order management and procurement-order-implementation systems' interfaces at a telecom's company. I was actually meant to be testing and documenting them, but eventually I became sufficiently irritated by the attitude that "oh it's complex and specific; that's why we have training".

You misrepresent my position. I'm a HCI evangelist and guru (see my opinions in the internet browser thread). There are some developers, though, that mistake making good interfaces with dumbing down interfaces to make them "easy" as opposed to being "productive".

The product I'm working on right now is used by traders in a very, very fast-pasted, high-stress work environment. Is it the easiest program in the world to use? No. Should it be? No. Everything is about efficiency and speed in the program. There's a learning curve to it, but it's a VERY friendly interface in that it doesn't get in your way, it shows you what needs to be done unobtrusively (dynamic field highlighting and validation wizards), etc.

It's very important that the interface is designed with the users in mind. Users are not always the same. The people intended to use regedit are not people who don't understand the registry (if you don't understand it, why are you editing it?). They are developers who know the registry and how it works, and regedit is a very effective tool for them (hence terms like DWORD, HEX, etc).

It is not proper to dumb down every single program so your average joe can use them. Not only does this place a welcome mat for screwing up your registry, but such "niceities" for making the interface "intuitive" to people who've no idea how the system works obstruct and interfere with the real users.

I don't see your claim that every task involved using the registry requires 5 textbooks and a 4 day trawl through documentation (apologies, hyperbole) before you go ahead. Most things on a computer, even system configuration, are reversible. Yes, I am aware there are things on every system, including in the registry, which will break booting and so on. But there are plenty of things that won't. If people don't want to risk taking out their computer, then don't go and deliberately search out regedit. It's not like it's easy to find.

Seems to me like you're contradicting yourself here. You admit regedit is not easy to find (in fact, you need to know it exists to find it), then you complain that its interface isn't designed for people who don't know how to use it. If you know the basics about the registry, regedit is perfectly acceptable and easy to use to manipulate it. What you're asking for (shortcut icons for the aliased root keys, for example) is essentially to make regedit a tutorial on what the registry is. That's something that does not belong in a programmer's productivity tool.

Because this occurs automatically without process intervention or 'registration' under Mac OS X. If you move the bundle and then start it, then it will update the location. Hell, with the mdserver in Tiger, it may update bundle locations automatically without them even having to be started.

This IS a registry. Apple just doesn't have an equivalent program like regedit to modify it. But that's exactly what the registry is.

God, I still hate Windows' system of handling libraries.

It's really no different than any other shared, runtime-linked libraries. Library hell is far more of an issue on Linux and MacOS X than Windows -- try linking a gcc 3.3-compiled library to an application built in OS X 10.4.

In a lot of ways, Windows Vista is a lot more advanced in handling shared libraries. Automatically, Vista creates shadow copies of DLLs that are used when the application is installed. Even if you update your system copy that's global, in the event that fails Vista will automatically link with the original, "intended" DLL. Such intelligence I don't see on Linux or MacOS X.

And that should be made clear as it is in most Apple documentation and most community-based documentation. Most people don't lie in their documentation without making it clear that they are doing so and why.

They did not lie -- what they said is true from a high level. That is the purpose of the "overview". It is not true in every case imaginable, and the second you start putting asterisks and semantic sentences in brief overviews you destroy the whole purpose of the overview. And if you think Apple and most community-based documentation overviews lack sweeping generalizations, you're very wrong.

Time must have been devoted to the re-implementation of Win32 into Vista.

How do you figure? Why would it need to be re-implemented? Win32 is already implemented in the NT operating systems.

Microsoft claimed for a significant period of Longhorn's development that it would not have Win32 compatibility.

This is news to me. Link?

Given the time involved for the MacBU, a relatively tiny team, to move Office from Toolbox to Carbon, a significant but not insurmountable one.

I'm sure you're well aware that transitioning from Toolbox to Carbon was intended to be easy. There are no fundamental differences, as there are between Win32 and .NET 3.0.

1. Various members of the development team in ~2002 said they were happy to be able to see it go. Not members of the marketing team. No, I don't have quotes, sorry. Google? (Which actually strikes me as weird, as that is before the abandonment of XP in favour of Server 2003 codebase...)

This is the first I've heard of this, and I keep up with this stuff. Since google turns up no results, and I can't remember this either, I'm going to say you're mis-remembering...and in such case, it's hardly a pillar for an argument.

2. Because it's a ton more code to support and patch. Which, is likely to be tending towards a mess (because everything does unfortunately). Just as Toolbox was becoming impenetrable by Mac OS 9.

Win32 is pretty stable by now. And since MS, unlike Apple, supports their products long-term, these Win32 fixes need to be made anyway for Windows 2000, XP, and 2003.

Aside from the fact that programs would be rewritten on XP, for .NET, and the become usable on Vista.

Yes, but this takes time and foresight -- which most developers do not have. Plus many applications are no longer maintained but still very important for people's use.

Long story short, it'd be the dumbest mistake in MS' history to remove Win32 from Vista. Virtually nothing would work...Firefox, Opera, Office, games, etc. Who would buy such a thing? No one -- these transitions need to happen over time. It was only with Vista that MS removed support for 16-bit DOS applications, afterall. You view this as a negative, but it's really the reason MS dominates and Apple is a niche player.

It does not make it less competitive, it does make it worse base for future development. Except that Windows 7 will (apparently) still have all the Win32 (and indeed Win16) APIs implemented. So it's a fantastic base if Microsoft never intends to actually replace Win32 with .NET.

This is still nonsense. FWIW, Win16 is already removed in the 64-bit versions of Windows XP, 2003, and Vista (and Vista is the last 32-bit Windows). And I think it is likely that Windows 7 will still have Win32 too, because why shouldn't it? It's genuinely not hurting anything by being there, and improving the quantity of software people can run. New apps can and will be in .NET 3.0, but legacy apps will still be supported.

68k->PPC transition was relatively smooth, given the ease of porting applications over.

You're looking at this from a developer POV. When talking marketshare, the developer doesn't matter. 68k->PPC was jarrnig and abrupt for users who had to rebuild their entire software collection. This is a major factor in MacOS losing marketshare to Windows, who even advertised they value backwards compatibility.

"Snap" because I agree. It's just that I wish Microsoft would have taken that jump. Yes, there would have been a risk involved, but it would have produced a better OS for me. Do you really believe that breaking backwards compatibility would have really users to other OSs in significant numbers?

Absolutely. For most users, the majority of programs they use already exist in refined, stable versions for Linux and MacOS X. The reason they use Windows is because they use Win32 software. Otherwise, MacOS X WOULD have far higher sales and more people switching. If Windows no longer uses Win32 software, the far and away #1 reason to use Windows -- vast application library and support -- disappears. Not only this, but it becomes an advantage of the GNU-based systems. Is this the real reason you think MS should've done it? You must know breaking the Windows applications would just drive people to other systems.

Do you really think it would have impacted the profitability of Microsoft (particularly if they kept selling XP)?

Absolutely. Windows revenue was stagnating before Vista's release. After Vista's release, MS posted record profits. This is a no-brainer.

You obviously believe that making Vista incompatible with Win32 apps, or spinning it off into a VM (sacrificing some compatibility), would have made it unsellable. I believe that given it's currently poor (but far from abysmal) sales numbers, it would make more sense to have continued selling XP while providing a considerably better operating system.

Define poor sales? It sold more than XP did at this time in its life. And when XP came out, Linux/MacOS people lauded its "poor sales" too. People are, frankly, stupid if they think even a slim majority of Windows users are going to rush out to retail and upgrade their OS just because a new one is out.

Windows users use Windows to use Win32 programs. That's what it's for. Windows XP runs that (and XP is still for sale, BTW -- again, this isn't Apple that forces upgrades post-haste), so they'll continue using XP. My dad bought a new high-end Dell system (Core 2 Duo 2.6GHz, GeForce 7900GTX, 2GB RAM, etc) back in September. It came with XP, and he could even upgrade for free to Vista, but he didn't. Why? Because all he uses it for is Office and Internet Explorer and his other work applications.

The vast majority of people are like him. They won't run out and buy Vista. They'll get Vista when their new PC ships with it. Commenting on Vista's "poor sales" just shows how little that person understands of the Windows user world.

 

[Gatekeeper]

*shrug* I've used Apple's Macintoshes in school, at work and at home for well over a decade now and have never had a problem with them.

Personally, I rather hope Apple never claims more than 10 to 25 percent of the PC market. If they get too big, it makes them a target for hackers and whatnot, IMHO. Plus, us Macintosh users would then lose the "catchet" of owning a classy, well-designed machine with a top-notch OS. We'll end up just being "another face in the PC crowd."

But, hey, if someone prefers a Microsoft-OSed PC, it's no skin off my nose. To each thine own ...

Gatekeeper

 

[GenMarshall]

I have been exposed to Apple's Macintoshes in School. The Earlyest was Apple ][ and the last Mac OS that I encountered in High School (Mainly in the Graphics Lab), was Mac OS 8. I have used my Aunt's iMac that has Mac OS X installed. However I perfer to use the Windows OS mainly because of the following:

1. It has a larger market share in which games are developed on a Windows OS first.
2. Ease of navigation. I can identify what driver I am (C:\, D:\, etc) using
3. Windows OS are now adays an Industrial/Bussiness standard (There are a few bussiness that uses Macs but it's rare)

The following OSes that I have used (Not counting Virtualization) at the time of this post are:

Windows: 95 (Home) ->98 (Home) ->NT (High School) ->2K (High School) -> XP (Home & College)
Mac (Mainly used in both Middle School & High School Graphics Lab): System 6.x -> System 7.x -> System 7.5.x -> Mac OS 8

I mainly perfer the older Classic Mac OSes due to it's ease in virtualization (Thanks with Sheepshaver to emulate a Power Mac and vMac to emulate Mac Plus) unlike Mac OS X where I will have to use PearPC plus extra steps on using Darwin to set up a partition and VMware does not have support for a x86 version of Mac OS X.

 

[dutchking]

Wanted a backlit LED display, and Dells an Sonys all suck for it.

It's only going to run Vista.

Ok, if you need help or have questions on installing Windows just ask me. And YOU MUST PRINT OUT THE APPLE GUIDE!!!! 28 pages, takes an hour if your lucky, but you will be totally SOL if you don't!!! Glad to have another Mac user on the forum. But one question: Why Vista? I decided to install XP because Vista is too new to be trusted to me, besides the fact that it is probably the buggiest system out there right now.

 

[Asher]

1. Vista's much better on multicore systems.
2. Work -- I work for a company making .NET 3.0 software
3. Games. Mac doesn't have any worth mentioning. XP doesn't do DirectX 10 (future gaming).

I've been using Vista since it came out on my main PC, I don't find it to be buggy and I appreciate a lot of the new features as well. Back when I got my last laptop, everyone looked at me weird when I put XP on it instead of the "un-bloaty, more robust" Windows 2000.

 

[Jaffa_lord]

I cant get enough Windows, Mac is crap!

 

[dutchking]

1. Vista's much better on multicore systems.
2. Work -- I work for a company making .NET 3.0 software
3. Games. Mac doesn't have any worth mentioning. XP doesn't do DirectX 10 (future gaming).

I've been using Vista since it came out on my main PC, I don't find it to be buggy and I appreciate a lot of the new features as well. Back when I got my last laptop, everyone looked at me weird when I put XP on it instead of the "un-bloaty, more robust" Windows 2000.

Ok, your choice. That's not mensioning how it looks... So what did you make your disk partition at? (assuming you already installed vista)

 

[kirstenburg]

to be litterally honest ... it doesn't matter which system is better so to speak ... all that matters is that the system as a whole meets YOUR requirements .....

if u want your computer to do something in particular buy the best one for that job .... does it matter which that is to anybody else ?


so in conclusion ... nethier is best ... both are as good as each other ... untill you add in different factors such as what you, as the user wants the system to do for you.

but then there will always be favoritism towards the system you own ... which is indeed pc's as they copme at a more affordable price and tend to match most ppl's requirements ... therefore it is no surprise that the majority of ppl are voting for pc's. I myself would vote for Mac's for the pure and simple reason that i own one .....

so is there any point in this pole ? because there techically isn't really a straight answer to the question.

 

[Jamesds]

to be litterally honest ... it doesn't matter which system is better so to speak ... all that matters is that the system as a whole meets YOUR requirements .....

if u want your computer to do something in particular buy the best one for that job .... does it matter which that is to anybody else ?


so in conclusion ... nethier is best ... both are as good as each other ... untill you add in different factors such as what you, as the user wants the system to do for you.

but then there will always be favoritism towards the system you own ... which is indeed pc's as they copme at a more affordable price and tend to match most ppl's requirements ... therefore it is no surprise that the majority of ppl are voting for pc's. I myself would vote for Mac's for the pure and simple reason that i own one .....

so is there any point in this pole ? because there techically isn't really a straight answer to the question.

Hey kirstenburg, let me first Welcome you to the CFC Forums!

What you say is very true, what matters is which suits your needs. Well, the poll is quite interesting to me because it confirms what I expected, but on a supposedly PC biased forum (this is the Civilization forums, after all) I would have expected a smaller percentage of Mac votes.

So, it seems Macs are well established in the market, and who knows what is yet to come - they have just released their new iMacs, and a new OS is coming soon too. I think Apple are going to become more popular in the near future.

It is also interesting to read of people who have experienced both, and from a technological view, find one better. However, again, that depends on how they 'think' and how they prefer to work.

It's been an interesting debate anyways!