Probably because you gain a *lot* more functionality/convenience from a two button mouse, and arguably quite a bit less from three. I think they're insane if they don't add a wheel, but it wouldn't really have to be clickable (lots of folks really don't get that the wheel is a button).
I'm firmly in the 3 button camp (UNIX/Linux user) but two is better than none, and I can see why they're doing that - especially given Apple's "interesting" notions about mice.
I'm sure I'm not the only one out there who will buy an iBook the minute they have a two-button (or more) trackpad...
What about the "why should I switch from my extremely fast IMAP server that I can access from anywhere, with optional webmail if I'm somewhere without ssh or IMAPs support" crowd.
Gmail doesn't interest me because (a) it doesn't have enough space and (b) it doesn't let me manage my mail how I want to.
I'd like its search features and awesome spam filter, but not enough to trust google with my mail, be forced to do archiving somewhere else, and have to manage my mail the way they think I should.
If you're wondering why I have a multi-gigabyte mailbox - time, mostly. I keep archives of most of the lists I'm on too as its faster and more convenient to search them this way than via their web search most of the time. That gets big after a while.
I'm not entirely convinced about the defensive patent argument. I'm sure it's true for many companies, but I'm not at all inclined to believe MS is one of them.
- They've tried to actively enforce some patents - admittedly rather old and crappy patents (FAT32 anybody?), but that's still hardly a defensive move. - It appears - I have no firsthand knowledge - that they want at least some degree of software patentability in the EU. This doesn't really fit with the purported view that the patent system is broken and they're just playing by the rules until it can be fixed. - If they wish to focus on defense with patents, why would they not even reveal details about their patent claims on the now-defunct Sender-ID, let alone license them?
Now, they're a big company and they do a lot of things - but some of this appears comes from a very high level, the EU patent push in particular. I don't think that's consistent with someone who's just playing along until things can be fixed.
Because their customers are used to having it, and don't want to upgrade anything they don't have to. They'll often have heavily customised CDE environments, or be an ISV that integrates apps into the CDE, etc.
I imagine that once they have good management tools for GNOME they'll push it a bit harder, but even so they can't offer the same sort of stability (in terms of compatible changes only, etc) for GNOME.
Personally, I logged into CDE only to go "gah!" and log back out... and I like XFCE. However, I'm sure Sun do have good reasons for keeping it around.
Well, partly or largely it's hypocrisy, but there's also the issue that most people who're infringing the copyright on music and films, even if they distribute them publically, do not claim they personally created them.
Sometimes, I think it's quite fine for the company to repair the situation and be done with it. If the wronged developers are OK with that, and think it's a reasonable option, that's just fine.
One one hand there's the desire to "send a message" but on the flip-side, some OSS developers feel that it's important not to totally alienate the commercial devel world, and understand that sometimes screwups and misunderstandings happen. Especially in a world of cheap contractors, offshoring, MCSEs who think they're developers, and limited knowledge of OSS and its licensing.
You may have guessed that I fall on the side of playing things cautiously. If someone stole my code, I'd either (a) laugh at them and tell them how utterly stupid they are for actually bothering to steal code that bad, or (b) get rather pissed off... but attempt to find out what's going on and why before flying off the handle.
I'm not at all convinced that much tolerance would be extended to really blatant offenders who deny it and continue to offend. Unfortunately, I'm not aware of any cases like that that've come to court.
That's the issue. Not everybody knows. I saw a post on the local Mac user group list (I support macs at work so I need to keep track) about CherryOS. They had no idea of the issues involved and the article they referred to did not mention them.
It's worth making some noise even about something this blatant.
While this is true, the chances are that most services will not be started by default on such a system. I was quite impressed with the default FC3 install - (almost) no services running by default, and a packet filter in place anyway.
While it is better not to have the services installed at all, it makes relatively little difference since the attacker would need some form of local access in order to use them if they do not run by default.
There's a lot more to games than the graphics. That's the point. Tracking objects, running AI, and huge amounts of other things are handled by the CPU, not the GPU.
If your game is GPU limited and you massively upgrade your graphics, you may find the CPU is unable to keep up. It may be unable to feed data to the GPU fast enough, for example, or unable to do housekeeping duties to keep up.
The two approaches are orthogonal and can be combined. For example, Red Hat FC3 has both SELinux and ExecShield (which includes library address randomisation and W^X-style checks like what MS calls "data execution protection").
No, the "PC guys" cannot guarantee that every single allocation and buffer access in a 100,000 line program is done correctly. Hell, the "PC guys" find it very hard just to make the program work at all.
I'm one of said guys, so I'm well and truly familiar with this pain.
Part of it is culture - features and fast development above correctness. A project I'm involved with now is a horrifying mass of spaghetti that barely works at all, yet they're still adding new features with little focus on cleanup. There is no test suite.
As for libraries, etc... not really, no. Nothing universal, anyway. Higher level toolkits do help a lot though.
IMO the part of the solution is use a language better suited to writing software on the scale we do it on PCs. Not C, suitable for OS kernels and microcontrollers, or C++, but something with AUTOMATIC MEMORY MANAGEMENT. No more (programmer-visible) buffer manipulation, no more pointer ownership problems, double free()s, and forgotten `delete's.
Use of tools like valgrind is also important, as is coding carefully and considering getting it right as more important than getting it "working".
Finally, though it's not as useful for security issues, test suites are a really important thing in software development IMO.
The author's point with the CPU issue is almost certainly that you've removed your video card bottleneck, and are quite likely to hit a CPU bottleneck as a result. This seems entirely sensible to me, and matches my experience when upgrading video cards in the past.
You won't need a faster CPU for SLI, but if you don't get one (or already have a really fast one) you're unlikely to get the full benefits of the enormous truckload of money you dropped on video cards. That's my understanding, anyway.
As for dual GPU cards, there is almost certainly a difference even if they present the same interface to the drivers etc. If nothing else, a dual GPU card would have only one bus interface, not two - a shared, slower link to the CPU and main memory, but faster communication between the cards themselves. At least in my limited understanding.
For me, that's just it - the interface. Higher resolution helps, as does my pro quality 19" monitor, but it's the interface that's the killer.
Show me a mouse that ships with a keyboard and mouse so console developers can *rely* on them being present, and I might care about console games. Hell, just the mouse would do, though mouse-and-controller would be more than a tad clumsy:S
As it is, I find most games I care about (RTS, strategy games, and games like Deus Ex and System Shock II) either don't exist for consoles, or are pathetic hollow shells of their former selves. Deus Ex II: Invisible War, anybody?
The upgrade cycle and low starting price would be attractive, but I don't find it too bad with my desktop. Then again, I don't buy top end gear, so I'm usually in the lower middle of the requirements and performance bracket. This is helped by the fact that I often play older games.
When it comes to developers building for consoles first, I'm painfully aware of that. I often fire up a game demo for something really interesting, and it has a bloody console interface or is written with the assumption that the user will be confined by console input devices. Especially in FPS games, this is *incredibly* annoying.
Call me a bigot, but if all I can get is console games, I'll just stop buying games. There are a lot of old, good games out there - and I'm gaming less these days anyway. I play games for fun, and I don't find console UIs fun.
That article is a bit light on details like the actual questions used. It also suggests that not just Google was covered. Can I tell what is an ad on Yahoo or MSN ? Hell, no.
I'm no Google fanboy, but I think their success is in part due to making it clear what are ads and what aren't, and making those ads things the searcher is likely to want to see.
I'm not sure that's really accurate. Slashdot isn't exactly known as the guiding light in OSS circles, and I'd be extremely reluctant to view posts here as representative. Even here, I think the idiots may be in the minority.
I modify software I work with quite a bit, actually.
I usually do so unhappily, bitching and moaning the whole time, as I'd prefer not to have to - but if I need a cusomisation for my site that's not configurable, I'll still modify the product if necessary.
I also fix the odd problematic bug and provide a patch with my bug report. As somone who does OSS development work, I *know* how happy that makes the developers.
That said, I'm working under different constraints than apply to a company buying MS software. We "pay" some of the gains we make on licensing in fixing the bloody software so it does what we need and does it properly. For my workplace this turns out well, for others it probably doesn't. I had someone to scream at for support who would actually fix things, I'd prefer to do that.
I do think the ability to modify MS products would be useful for some. Look at the extent of modifications many outfits already do to their SOEs... I rather doubt code mods would be a big step above the often rather major surgery they already do. I've certainly heard enough people swearing about being unable to change a particular setting via group policy....
That's the way. Write code that doesn't need comments to be understood, then comment the damn thing anyway;-)
There's very little more helpful than comments scattered through the code explaining the "bigger picture" -/why/ this routine does this this way, etc. Those are usually things you can't impart through the code its self anywhere near as easily.
I've been known to click on non-offensive ads that are of interest, usually related to things I need at work. "Hmm, another supplier of SATA based storage servers, you say?..."
That said, I click on ads only extremely rarely and NEVER on annoying ones. Interesting, really, as I'm one of those folks who is really only reached by 'net adversising - I watch no TV (cant't stand the ads and the drivel) and listen only to non-commecial radio (similar reasons).
I wonder how many more people there are who don't mind inoffensive, low key advertising and may actually view it as useful, but avoid anything more intrusive? I suspect "a lot".
Frankly, I don't think there _is_ a speed issue with X11. There are performance issues on XFree86/XOrg with some (many) drivers, AFAIK mostly due to limited developer time and limited access to hardware. The fact that the current software RENDER implementation is not signficantly optimised, and few drivers implement RENDER hardware acceleration, does not help.
Working on my NVidia equpped box here (GeForce Ti, nvidia drivers, but for 2D 'nv' is almost as good) X is much snappier than I usually find WinXP to be. Turning on RENDER acceleration has helped a lot.
I'm sure folks will bring up the "because of the network" myth up here, so let's get this straight - any slowness in X is not because of network support. Go ask Keith Packard, I'm pretty sure he's been rather clear on the matter more than once. My personal, very much non-expert understanding is that most performance issues peope experience are due to limited hardware acceleration and inferior drivers.
If you don't believe me about how much difference the hardware and drivers make, go find an S3 based system, preferably S3 Trio32/S3 Trio64, and compare it to a PCI-based (to keep it fair) NVidia GeForce 4 MX on the same hardware. It's like they're two totally different computers - the change is jaw-dropping. I use thin clients a lot, so I care strongly about video performance and tend to notice these things.
It's also worth noting that hopefully many of these plans will lead indirectly to performance improvements, by making RENDER acceleration and RENDER optimisation pretty much mandatory.
Actually, Samba doesn't need to be Windows compatible to be useful either. Aside from the many other devices that speak SMB (many NAS units, etc), there's also the not insignficant matter of the POSIX extensions in Samba.
I'm increasingly inclined to view CIFS and Samba as an NFS alternative for Linux that does't suck. I'm hoping NFS4 turns out well, but I'm not holding my breath. We do need something to replace NFS, and we need it about ten years ago.
My advice, and that of serious Windows support pros I've worked with: Do it over the network. All of it. Even OS installs. Slipstream service packs and fixes into your build image, along with your base software etc. Install packages automatically on login using AD. You can do all this... and it'll save you a lot of pain. Hell, you don't even need to worry about your CD key, you can do that as part of the automated network install script.
I'm using Linux thin clients for most of my basic needs users at work. They're getting pretty good now, but I'm still running into a frustrating number of stupid bugs. I think I spend about an equal amount of time supporting them and the win98 users - at "near zero". Ditto our one and only XP user now that I've got the bugs ironed out. Most of my time is wasted supporting the MacOS 9 desktop publishing staff due to the nightmarish OS and apps involved there.
If you think Windows is hard to manage, try MacOS. ARRGGGGHHHHHH. MacOS/X is a little better, but still pretty awful IMO.
Microsoft is also pretty reasonable with CD keys etc compared to many companies. QuarkXPress and Adobe Photoshop both scan the network for other copies, interrogate them for their CD key, and refuse to run if they find it's the same. This makes image based installs impossible since they don't provide any way to install and configure the app, then "de-personalize" it so all you have to do to get it working is enter the CD key. (You can do this with Windows, BTW). Those apps are a nightmare and in comparison Windows looks absolutely lovely to manage.
I'm also finding my trials with OO.o and GNOME for our journalists pretty dismal so far. All sorts of weird bugs keep on turning up and I'm about to give up and get them Windows boxes. I use Linux at home without issues, but these uses can and do break stuff all the time.
In the end, I guess it comes down to picking the right tool for the right job. MS desktops, managed well, are OK. I don't like them, but they work. Especially if you lock IE down so hard the user can't even run it, and if they figure out how to run it anyway, can't visit anywhere or do anything. Too bad they cost so bloody much and still insist on bundling IE, Outlook Express (Yes you can remove it, but it'll be back every time you patch the damn OS), etc.
Slashdot Mirror
Probably because you gain a *lot* more functionality/convenience from a two button mouse, and arguably quite a bit less from three. I think they're insane if they don't add a wheel, but it wouldn't really have to be clickable (lots of folks really don't get that the wheel is a button).
I'm firmly in the 3 button camp (UNIX/Linux user) but two is better than none, and I can see why they're doing that - especially given Apple's "interesting" notions about mice.
I'm sure I'm not the only one out there who will buy an iBook the minute they have a two-button (or more) trackpad...
What about the "why should I switch from my extremely fast IMAP server that I can access from anywhere, with optional webmail if I'm somewhere without ssh or IMAPs support" crowd.
Gmail doesn't interest me because (a) it doesn't have enough space and (b) it doesn't let me manage my mail how I want to.
I'd like its search features and awesome spam filter, but not enough to trust google with my mail, be forced to do archiving somewhere else, and have to manage my mail the way they think I should.
If you're wondering why I have a multi-gigabyte mailbox - time, mostly. I keep archives of most of the lists I'm on too as its faster and more convenient to search them this way than via their web search most of the time. That gets big after a while.
I'm not entirely convinced about the defensive patent argument. I'm sure it's true for many companies, but I'm not at all inclined to believe MS is one of them.
- They've tried to actively enforce some patents - admittedly rather old and crappy patents (FAT32 anybody?), but that's still hardly a defensive move.
- It appears - I have no firsthand knowledge - that they want at least some degree of software patentability in the EU. This doesn't really fit with the purported view that the patent system is broken and they're just playing by the rules until it can be fixed.
- If they wish to focus on defense with patents, why would they not even reveal details about their patent claims on the now-defunct Sender-ID, let alone license them?
Now, they're a big company and they do a lot of things - but some of this appears comes from a very high level, the EU patent push in particular. I don't think that's consistent with someone who's just playing along until things can be fixed.
Because their customers are used to having it, and don't want to upgrade anything they don't have to. They'll often have heavily customised CDE environments, or be an ISV that integrates apps into the CDE, etc.
... and I like XFCE. However, I'm sure Sun do have good reasons for keeping it around.
I imagine that once they have good management tools for GNOME they'll push it a bit harder, but even so they can't offer the same sort of stability (in terms of compatible changes only, etc) for GNOME.
Personally, I logged into CDE only to go "gah!" and log back out
Well, partly or largely it's hypocrisy, but there's also the issue that most people who're infringing the copyright on music and films, even if they distribute them publically, do not claim they personally created them.
That makes a big difference.
Sometimes, I think it's quite fine for the company to repair the situation and be done with it. If the wronged developers are OK with that, and think it's a reasonable option, that's just fine.
... but attempt to find out what's going on and why before flying off the handle.
One one hand there's the desire to "send a message" but on the flip-side, some OSS developers feel that it's important not to totally alienate the commercial devel world, and understand that sometimes screwups and misunderstandings happen. Especially in a world of cheap contractors, offshoring, MCSEs who think they're developers, and limited knowledge of OSS and its licensing.
You may have guessed that I fall on the side of playing things cautiously. If someone stole my code, I'd either (a) laugh at them and tell them how utterly stupid they are for actually bothering to steal code that bad, or (b) get rather pissed off
I'm not at all convinced that much tolerance would be extended to really blatant offenders who deny it and continue to offend. Unfortunately, I'm not aware of any cases like that that've come to court.
That's the issue. Not everybody knows. I saw a post on the local Mac user group list (I support macs at work so I need to keep track) about CherryOS. They had no idea of the issues involved and the article they referred to did not mention them.
It's worth making some noise even about something this blatant.
Hmm... I thought most theatrical releases were dubbed, myself. Certainly almost all that ever make it here (Australia) are.
As someone who finds dubs, especially American Anime dubs, a bit like cleaning their ears with a chainsaw, I'm not convinced dubs are good for sales.
While this is true, the chances are that most services will not be started by default on such a system. I was quite impressed with the default FC3 install - (almost) no services running by default, and a packet filter in place anyway.
While it is better not to have the services installed at all, it makes relatively little difference since the attacker would need some form of local access in order to use them if they do not run by default.
There's a lot more to games than the graphics. That's the point. Tracking objects, running AI, and huge amounts of other things are handled by the CPU, not the GPU.
If your game is GPU limited and you massively upgrade your graphics, you may find the CPU is unable to keep up. It may be unable to feed data to the GPU fast enough, for example, or unable to do housekeeping duties to keep up.
The two approaches are orthogonal and can be combined. For example, Red Hat FC3 has both SELinux and ExecShield (which includes library address randomisation and W^X-style checks like what MS calls "data execution protection").
No, the "PC guys" cannot guarantee that every single allocation and buffer access in a 100,000 line program is done correctly. Hell, the "PC guys" find it very hard just to make the program work at all.
... not really, no. Nothing universal, anyway. Higher level toolkits do help a lot though.
I'm one of said guys, so I'm well and truly familiar with this pain.
Part of it is culture - features and fast development above correctness. A project I'm involved with now is a horrifying mass of spaghetti that barely works at all, yet they're still adding new features with little focus on cleanup. There is no test suite.
As for libraries, etc
IMO the part of the solution is use a language better suited to writing software on the scale we do it on PCs. Not C, suitable for OS kernels and microcontrollers, or C++, but something with AUTOMATIC MEMORY MANAGEMENT. No more (programmer-visible) buffer manipulation, no more pointer ownership problems, double free()s, and forgotten `delete's.
Use of tools like valgrind is also important, as is coding carefully and considering getting it right as more important than getting it "working".
Finally, though it's not as useful for security issues, test suites are a really important thing in software development IMO.
The author's point with the CPU issue is almost certainly that you've removed your video card bottleneck, and are quite likely to hit a CPU bottleneck as a result. This seems entirely sensible to me, and matches my experience when upgrading video cards in the past.
You won't need a faster CPU for SLI, but if you don't get one (or already have a really fast one) you're unlikely to get the full benefits of the enormous truckload of money you dropped on video cards. That's my understanding, anyway.
As for dual GPU cards, there is almost certainly a difference even if they present the same interface to the drivers etc. If nothing else, a dual GPU card would have only one bus interface, not two - a shared, slower link to the CPU and main memory, but faster communication between the cards themselves. At least in my limited understanding.
For me, that's just it - the interface. Higher resolution helps, as does my pro quality 19" monitor, but it's the interface that's the killer.
:S
Show me a mouse that ships with a keyboard and mouse so console developers can *rely* on them being present, and I might care about console games. Hell, just the mouse would do, though mouse-and-controller would be more than a tad clumsy
As it is, I find most games I care about (RTS, strategy games, and games like Deus Ex and System Shock II) either don't exist for consoles, or are pathetic hollow shells of their former selves. Deus Ex II: Invisible War, anybody?
The upgrade cycle and low starting price would be attractive, but I don't find it too bad with my desktop. Then again, I don't buy top end gear, so I'm usually in the lower middle of the requirements and performance bracket. This is helped by the fact that I often play older games.
When it comes to developers building for consoles first, I'm painfully aware of that. I often fire up a game demo for something really interesting, and it has a bloody console interface or is written with the assumption that the user will be confined by console input devices. Especially in FPS games, this is *incredibly* annoying.
Call me a bigot, but if all I can get is console games, I'll just stop buying games. There are a lot of old, good games out there - and I'm gaming less these days anyway. I play games for fun, and I don't find console UIs fun.
That article is a bit light on details like the actual questions used. It also suggests that not just Google was covered. Can I tell what is an ad on Yahoo or MSN ? Hell, no.
I'm no Google fanboy, but I think their success is in part due to making it clear what are ads and what aren't, and making those ads things the searcher is likely to want to see.
"... from the open-source community ..."
I'm not sure that's really accurate. Slashdot isn't exactly known as the guiding light in OSS circles, and I'd be extremely reluctant to view posts here as representative. Even here, I think the idiots may be in the minority.
I modify software I work with quite a bit, actually.
... I rather doubt code mods would be a big step above the often rather major surgery they already do. I've certainly heard enough people swearing about being unable to change a particular setting via group policy....
I usually do so unhappily, bitching and moaning the whole time, as I'd prefer not to have to - but if I need a cusomisation for my site that's not configurable, I'll still modify the product if necessary.
I also fix the odd problematic bug and provide a patch with my bug report. As somone who does OSS development work, I *know* how happy that makes the developers.
That said, I'm working under different constraints than apply to a company buying MS software. We "pay" some of the gains we make on licensing in fixing the bloody software so it does what we need and does it properly. For my workplace this turns out well, for others it probably doesn't. I had someone to scream at for support who would actually fix things, I'd prefer to do that.
I do think the ability to modify MS products would be useful for some. Look at the extent of modifications many outfits already do to their SOEs
That's the way. Write code that doesn't need comments to be understood, then comment the damn thing anyway ;-)
/why/ this routine does this this way, etc. Those are usually things you can't impart through the code its self anywhere near as easily.
There's very little more helpful than comments scattered through the code explaining the "bigger picture" -
Some marketoids actually view annoying ads as the best, on the theory that they stand out. I really hope they're not right.
I know I've refused to deal with companies before because of their advertising, but I'm not sure the majority of folks will.
I've been known to click on non-offensive ads that are of interest, usually related to things I need at work. "Hmm, another supplier of SATA based storage servers, you say?..."
That said, I click on ads only extremely rarely and NEVER on annoying ones. Interesting, really, as I'm one of those folks who is really only reached by 'net adversising - I watch no TV (cant't stand the ads and the drivel) and listen only to non-commecial radio (similar reasons).
I wonder how many more people there are who don't mind inoffensive, low key advertising and may actually view it as useful, but avoid anything more intrusive? I suspect "a lot".
Frankly, I don't think there _is_ a speed issue with X11. There are performance issues on XFree86/XOrg with some (many) drivers, AFAIK mostly due to limited developer time and limited access to hardware. The fact that the current software RENDER implementation is not signficantly optimised, and few drivers implement RENDER hardware acceleration, does not help.
Working on my NVidia equpped box here (GeForce Ti, nvidia drivers, but for 2D 'nv' is almost as good) X is much snappier than I usually find WinXP to be. Turning on RENDER acceleration has helped a lot.
I'm sure folks will bring up the "because of the network" myth up here, so let's get this straight - any slowness in X is not because of network support. Go ask Keith Packard, I'm pretty sure he's been rather clear on the matter more than once. My personal, very much non-expert understanding is that most performance issues peope experience are due to limited hardware acceleration and inferior drivers.
If you don't believe me about how much difference the hardware and drivers make, go find an S3 based system, preferably S3 Trio32/S3 Trio64, and compare it to a PCI-based (to keep it fair) NVidia GeForce 4 MX on the same hardware. It's like they're two totally different computers - the change is jaw-dropping. I use thin clients a lot, so I care strongly about video performance and tend to notice these things.
It's also worth noting that hopefully many of these plans will lead indirectly to performance improvements, by making RENDER acceleration and RENDER optimisation pretty much mandatory.
Perhaps they were all reading the article? *gasp*
I know, it's more likely the comments were intercepted en-route by a pack of marauding ducks, but hey, it could happen.
Actually, Samba doesn't need to be Windows compatible to be useful either. Aside from the many other devices that speak SMB (many NAS units, etc), there's also the not insignficant matter of the POSIX extensions in Samba.
I'm increasingly inclined to view CIFS and Samba as an NFS alternative for Linux that does't suck. I'm hoping NFS4 turns out well, but I'm not holding my breath. We do need something to replace NFS, and we need it about ten years ago.
Well, Python integrates tolerably with C and fairly poorly with C++. It's clumsy, verbose, and annoying, but the end result does work well.
Libraries like Boost::Python and tools like Pyrex help a *lot* when trying to make Python/C integration a bit less excruciating.
My advice, and that of serious Windows support pros I've worked with: Do it over the network. All of it. Even OS installs. Slipstream service packs and fixes into your build image, along with your base software etc. Install packages automatically on login using AD. You can do all this... and it'll save you a lot of pain. Hell, you don't even need to worry about your CD key, you can do that as part of the automated network install script.
I'm using Linux thin clients for most of my basic needs users at work. They're getting pretty good now, but I'm still running into a frustrating number of stupid bugs. I think I spend about an equal amount of time supporting them and the win98 users - at "near zero". Ditto our one and only XP user now that I've got the bugs ironed out. Most of my time is wasted supporting the MacOS 9 desktop publishing staff due to the nightmarish OS and apps involved there.
If you think Windows is hard to manage, try MacOS. ARRGGGGHHHHHH. MacOS/X is a little better, but still pretty awful IMO.
Microsoft is also pretty reasonable with CD keys etc compared to many companies. QuarkXPress and Adobe Photoshop both scan the network for other copies, interrogate them for their CD key, and refuse to run if they find it's the same. This makes image based installs impossible since they don't provide any way to install and configure the app, then "de-personalize" it so all you have to do to get it working is enter the CD key. (You can do this with Windows, BTW). Those apps are a nightmare and in comparison Windows looks absolutely lovely to manage.
I'm also finding my trials with OO.o and GNOME for our journalists pretty dismal so far. All sorts of weird bugs keep on turning up and I'm about to give up and get them Windows boxes. I use Linux at home without issues, but these uses can and do break stuff all the time.
In the end, I guess it comes down to picking the right tool for the right job. MS desktops, managed well, are OK. I don't like them, but they work. Especially if you lock IE down so hard the user can't even run it, and if they figure out how to run it anyway, can't visit anywhere or do anything. Too bad they cost so bloody much and still insist on bundling IE, Outlook Express (Yes you can remove it, but it'll be back every time you patch the damn OS), etc.