Probably that thingy qualifies as semi-weird only, but my Philips (equivalent to Yujin Robotics Iclebo Smart) runs a full Linux, on a somewhat weirder Samsung-based ARM (S3C2460?? NOT bog-standard 2440!!) platform w/ 64MB flash and 64MB RAM and a USB port for firmware upgrades with an RT73 WLAN driver available which does work with e.g. a hama no. 00062744 type WLAN stick (plus audio hardware even, but not software-supported by default unfortunately). I have to admit that I haven't found the time yet to morph that machine into an OpenWrt-based platform, but I sure as hell do plan to.
So damn right. Put that URL in my ToDo list to check on later, only to then realize that SFD had happened on that very day. Thus no possibility to even visit it, not to mention actively help out. Why is it that such events frequently are so poorly advertised for?
Oh yeah? Then start buying one of those very well-known Logitech webcams (or almost any other of all the UVC-based cams out there), or try to get proper r/w NTFS support working (which package to choose for Snow Leopard? Tiger? Mountain Lion? And should you manually install macfuse? osxfuse? fuse4x? and which custom-tailored community-provided helper shell scripts to use to get GUI mounts working properly?), or try getting UDF support that is on par with other choices (10.4 sucked majorly in that regard, and I'm sure 10.6 has deficiencies there as well), or even buying a scanner (that one is arguably sort of ok), or try getting decent integrated SMART support (yes indeed, my MacBook HDD died...), or try getting write/*read* access to your HDD power management settings. And last time I heard some Linux graphics wiz described OS X as completely sub-par on the graphics system layer front (maybe does not apply to 10.8, though). And the things that Torvalds said about HFS+ were not all too pretty either... Or just try figuring out where exactly in the GIMP Applications installation tree its Python plugins are supposed to be placed (almost all docs here are Linux-focused).
And don't tell me a twice-broken palm rest and balloon-bulged battery (after two years!) is "quality hardware".
That company has been obsessed with their oh-shiny childs play hardware for too long methinks, with development of their professional products woefully abandoned. And don't get me started on their developer policy. All things to-be-paid-for, otherwise you'll end up with very outdated versions. Do they want people to do software development to have their under-represented infrastructure improved, or not???
I admit it: I'm an Apple customer. Once.
(not that Mafia$oft is much - if at all - better, admittedly)
You mean things such as: - the eternal stdafx.h vs. StdAfx.h, resource.h/Resource.h fire-fight? - the eternal "when will Visual Studio finally learn to handle reloading of *externally* modified project files in a comfortable way rather than a user having to manually acknowledge each of a hundred automatically regenerated projects" fire-fight? (see CMake suffering) - the annoying "when will SvnBridge finally be even remotely resembling something less buggy and the public project be doing even minimal amounts of urgent project maintenance"? And this in the face of originally independent and very enthusiastic founding developers... - the annoying "when will TFS finally have Web Access for other platforms that's not taking 2 (yes TWO!) minutes to login" (after a "your session timed out" forcibly-to-be-acknowledged non-hideable message box having appeared after some 10 seconds, to add insult to injury! So yeah you're FORCED to actively wait a full two minutes every time)
Buying into Mafia$oft "ecosystem" "infrastructure"!? HELL NO NEVER AGAIN!!
[reposting old semi-finished copy since Slashdot ate my entire edit window cold and UNRECOVERABLE when messing with Options dialog - and that's not the first time either of severe data loss with Slashdot website design!]
- no atomic checkins ("goddamn, somebody forgot to check in another file again" - "err what in fact I did"), no consistent changesets - 5 GB max (official Microsoft recommendation), severe index size limitations - may corrupt easily - branching is ""supported"" but stay far away from it (said to be #1 source of corruption) - NO BLAME MECHANISM (perusing the Changelog thus takes about 5 minutes each to hit the relevant spot) - no 3-way merge AFAICS (and TFS still does not have that either) - cross-platform support is laughable (there's SOS, but you easily hit trouble on each of RHEL, Ubuntu and Mac, and of course there's no 64bit platform anyway, and it's x86 only) - performance is abysmal (depending on the moon phase, current mood of the Windoze server and number of busy colleagues, waiting for - yes, indeed - 1.5GB of source repository TAKES AGES (15 min. with VSS on Windows, up to 2 hours on SOS when synching database and subsequently fetching the entire tree); dito simply switching folders in GUI client treeview on SOS may take up to 5 minutes sometimes; this all as opposed to waiting for at most one minute for a kernel.org update to arrive from an internet server - terrible check-out mode of operation, leading to that dreaded "all files readonly" standard way of operation as opposed to good SCM - large slowdown of MSVS operation (lock contention!?!?) when additionally running the standalone VSS client - keeps asking the same stupid questions over and over again rather than simply fetching files - no distributed SCM (minor item) - sos command line client is NOT parallelization-safe (needs addition of manual flock/lockfile mechanisms in a wrapper script for reliable use within a massively parallel build system) - VSS DOES NOT LOCALLY DELETE FILES REMOVED FROM REPO (that one is a killer)
And please do me a favour and don't go cold-turkey to a TFS "upgrade" but have a worthwhile attempt at reaching an informed decision of which ALM/SCM to actually use. TFS, while most likely a lot better than VSS, is said to still have lots of rough spots. See various Internet blogs (e.g. "TFS the Lotus Notes of ALM").
Suitable alternative names may be Mercurial, git, and ALM components such as Trac.
xp, vista, or 7, or even 9x still? windows variant: starter, home, professional, enterprise, ultimate,...? (7 versions I believe, with broad repercussions on networkability etc.) xp sp1, 2 or hopefully 3?.net package installed, which version? av, firewall, all security updates ok? using xp, or a newer version i.e. at least moderately tolerable usb / wifi support? trying to make use of SATA on challenged xp versions?
Trying to use braindead legacy applications on newer systems (UAC, registry/file system shadowing,...)?
And don't get me started on laughable localization support shortcomings in Windows... (admittedly off-topic)
I had thought that there were some ways to overcome ntlmv2 stupidities...
Some minor search for "ntlmv2 tunneling linux" or some such showed e.g. http://cntlm.sourceforge.net/
Even if it works, still doesn't quite solve the problem that probably only the more knowledgeable part of the users would know about such solutions and know to use them, though...
Have you considered that by giving such a final answer, YOU might actually be(come) the roadblock to properly working non-Windows/Mac access?;-)
I applaud your efforts to try to get the support situation changed, but instead of stating a disillusioning "you can't", you should definitely switch to "I've tried a lot and so far I think you can't, but definitely feel free to show me otherwise.".
There might always come along someone who is brilliantly smarter than the entire rest and either a) figures out the specialty solution, or b) simply source-hacks together EVERYTHING required to "make it work anyway". At which point there's a nice solution ready for the taking...
And how many victims have seen their crucial personal search&rescue
efforts hindered/inhibited (translation: indirect deaths) by that
nuclear emergency and corresponding high risks within radiation-affected
areas?
Incidentally, if anyone out there has suggestions on how to reliably test for race conditions, please speak up.
Simply using Valgrind's Helgrind is one way to do that - available only on the best operating system that money can _not_ buy, of course;)
And the expert method here would be to Helgrind-run the unit tests themselves;)
(plus standard Valgrind memcheck runs)
But generally spoken, if people happen to have severe trouble getting their code free from race conditions, then that easily suggests that they have a plenty weird relationship with threading and should rethink their model (think isolated worker threads with cleanly passed _work units_ data and resulting "completed" notifications).
Having some 30 different threads in an app, all with liberal - and then not even properly locked - access to global data is __NOT__ the way to do things.
Enable XP Accessibility setting "reduced colors" in control panel.
Voila, all Excel cells are white, with none of the actual colors to be seen anywhere and, more importantly, no notice to be seen anywhere _either_.
Not to mention that this is really nice if for the life of it you cannot remember ever having enabled this Accessibility setting (I suspect it may have been enabled by SP3 update or some other quirky software), on an office computer that no kids have been getting at at any time.
luser "solution" of the problem? REINSTALL (yes, that's the same kind of reinstall as in the Ubuntu screen resolution case - which some users above even confirmed to _also_ happen with XP in certain situations - maybe driver-dependent?).
An experienced user (me) would be able to solve this very puzzling issue of lost colors on XP after >= 10 minutes of hardcore googling.
wxWidgets sockets: Linux: broken. Mac: broken. Windows: works (talking about multi-threading race conditions within the 3 layers of socket classes).
Morale of the story: use wx for GUI parts, but consider using something else for those non-GUI parts which it has trouble with or which are less likely to be properly served by a _GUI_ framework.
For networking stuff, libcurl is VERY hard to beat (binaries for perhaps up to 50 (FIFTY) platforms, used by "all fortune 500",...).
Spoken by someone who's been using wx all the time for 5 years now and generally likes it very much, but... well...
you wouldn't let a butcher (GUI) take care about your dents (sockets) either, right!?!?
Umm, I don't think you realized that your two postings were just about polarly opposite statements...
He wasn't talking about having Windows booted, he was mentioning 8 seconds _up to_ having Windows boot, whereas you...
Yup, by now it should be well-known that they are unable to design anything worth any crap.
That M company somehow managed to spoil and mentally disrupt decades of programmers.
See http://www.flounder.com/guiowner.htm etc. for radio button stuff and many other scary things.
Yup, on a universally open system that tracks your every move, chews your left finger and kills your dog on moonless nights.
On systems that are worth coding for, these things might be called inotify or dnotify or some such.
Anyway, thanks for documenting the way to do it on a certain platform.
And I'd say people who go ahead implementing it in such a low-bottom miserable way without any second thought still have lots and lots and lots more things to learn still.
Debian system, 1995, Athlon XP, eternally upgraded, gone through 1, 1.2, 4.3, 13, 30, 40, 80, 120GB. Yes, that's 8 HDDs copied over each time!
Now do that with Windows. Judging from anecdotal evidence it will easily quit working with just one peripheral not aligned to the moon phase.
I don't even remember whether any part in that box is still original...
Sorry for spoiling your party;)
An ALC268_ACER_ASPIRE_ONE entry is available in 2.6.28-rc2 patch_realtek.c, too, IOW ALSA git pull request obviously has been handled and I'll now have to test 2.6.28-rc2 to see whether it really works.
Why then does my Acer One A110L (oh, look, that's the same machine even! Unless you've got 150L...) have issues on 8.10 with permanently non-working internal microphone (no matter what you try) and spotty ath5k wifi (it sorta works, but then sometime later you experience weird connection delays or Even OOPSes or so) and very, very, very problematic SD card reader support (resume may kill your SD partition table even on kernels as new as 2.6.27, and God Help You if you dared putting a swap partition on it even) and sharply reduced SSD performance after resume? (UDMA/33 instead of 66)??
_Usable_ sound support consists of both output _and_ input, you know... (so many people in the Acer One league seem to forget this).
And especially given a nice builtin webcam...
I'm going to try ALSA source using module-assistant (that one is said to improve sound situation, but if it isn't fully fixed then I guess I'm going to take the valuable time to augment the realtek-specific patch file for intel-hda to properly support the seemingly unusual Acer One sound peripheral setup).
OK, the situation could certainly be much worse (especially considering an actually usable WLAN), but it's far, FAR from perfect.
Several issues listed here aren't Acer One specific actually (and I expect them to be fixed relatively soon), and the A110L actually is a very nice machine (minus stupid LCD display coating and somewhat lame performance of an SSD, especially in combination with relatively insufficient 512MB).
You should have mentioned that you have tried Wine 1.0 and that
it certainly didn't work for you, especially that it didn't support Photoshop despite
that particular one being as more or less DIRECTLY supported by it as it can
realistically get. Not to mention that the "open source Windows OS project" is
pretty much based on Wine anyway...
Plus, Vista has such a stunningly _high_ appcompat rating percentage that Wine would
"NEVER" be able to come even distantly close to this extraordinary amount of compatibility
with Windows applications...
Now please let me go die in peace from my lethal sarcasm overdose...
Linux has: - completely standardized/open formats and protocols as opposed to the "completely brand new" file formats that Microsoft loves to create each year - had fully working x86 64bit support 3(!) years earlier than Windows (well... minus a ton of 64bit-incapable proprietary offerings, which never get their things done in time anyway) - had nice 3D desktops at least a year earlier than Vista - enjoyed a nice Bluetooth stack as the very first operating system ever - and several others I don't recount right now
So, pray tell me, this is "completely brand new" (implies lots of bugginess) how?
So, let's see how well Windows does in comparison: - "completely brand new" office suite (entirely redesigned GUI) with "completely brand new" DRM-crippled file format - "completely brand new" operating system with "completely brand new" yet currently less visible DRM crippling which extends all the way down to hugely increasing hardware development costs and complexity to make sure users will have lots of fun with self-crippling hardware (protected audio path etc.pp.) - "completely brand new" Windows shell (oh, right, NOT available) and fully based on.NET (again, NOT) and WinFS (right, again NOT, has been in development for almost FIFTEEN years I think, since WinNT at least),... - "completely brand new" Windows build (how much time did it take them to come up with a final build of Vista after the semi-decade-long desaster that was Vista development? 6 months?? And you think this will turn out to be stable??)
Conclusion: please tell me where your point is?
Re:Why replace something that works well (PS/2)?
on
Leopard Vs. Vista
·
· Score: 1
You forgot another important reason: PS/2 uses its very own and special interrupt, thus wasting valuable interrupt space. This is less of an issue now that many systems use APIC by default, though.
Re:Why replace something that works well (PS/2)?
on
Leopard Vs. Vista
·
· Score: 1
OK, I'll tell you: PS/2 is easier.;)
Why? Easy, because PS/2 has a flat top on the round connector, whereas your usual bog-standard USB connector has an uncharacteristic design which won't tell you by hand which side to use (the only distinguishing "finger" is *internal* to the connector). Fine, now you could tell me that some PS/2 plugs are rotated by 90 degrees thus plugging it in is difficult, but then some USB plugs are rotated, too.
Slashdot Mirror
Probably that thingy qualifies as semi-weird only, but my Philips (equivalent to Yujin Robotics Iclebo Smart) runs a full Linux, on a somewhat weirder Samsung-based ARM (S3C2460?? NOT bog-standard 2440!!) platform w/ 64MB flash and 64MB RAM and a USB port for firmware upgrades with an RT73 WLAN driver available which does work with e.g. a hama no. 00062744 type WLAN stick (plus audio hardware even, but not software-supported by default unfortunately). I have to admit that I haven't found the time yet to morph that machine into an OpenWrt-based platform, but I sure as hell do plan to.
So damn right. Put that URL in my ToDo list to check on later, only to then realize that SFD had happened on that very day.
Thus no possibility to even visit it, not to mention actively help out. Why is it that such events frequently are so poorly advertised for?
Oh yeah? Then start buying one of those very well-known Logitech webcams (or almost any other of all the UVC-based cams out there), or try to get proper r/w NTFS support working (which package to choose for Snow Leopard? Tiger? Mountain Lion? And should you manually install macfuse? osxfuse? fuse4x? and which custom-tailored community-provided helper shell scripts to use to get GUI mounts working properly?), or try getting UDF support that is on par with other choices (10.4 sucked majorly in that regard, and I'm sure 10.6 has deficiencies there as well), or even buying a scanner (that one is arguably sort of ok), or try getting decent integrated SMART support (yes indeed, my MacBook HDD died...), or try getting write/*read* access to your HDD power management settings. And last time I heard some Linux graphics wiz described OS X as completely sub-par on the graphics system layer front (maybe does not apply to 10.8, though). And the things that Torvalds said about HFS+ were not all too pretty either...
Or just try figuring out where exactly in the GIMP Applications installation tree its Python plugins are supposed to be placed (almost all docs here are Linux-focused).
And don't tell me a twice-broken palm rest and balloon-bulged battery (after two years!) is "quality hardware".
That company has been obsessed with their oh-shiny childs play hardware for too long methinks, with development of their professional products woefully abandoned.
And don't get me started on their developer policy. All things to-be-paid-for, otherwise you'll end up with very outdated versions. Do they want people to do software development to have their under-represented infrastructure improved, or not???
I admit it: I'm an Apple customer. Once.
(not that Mafia$oft is much - if at all - better, admittedly)
You mean things such as:
- the eternal stdafx.h vs. StdAfx.h, resource.h/Resource.h fire-fight?
- the eternal "when will Visual Studio finally learn to handle reloading of *externally* modified project files in a comfortable way rather than a user having to manually acknowledge each of a hundred automatically regenerated projects" fire-fight? (see CMake suffering)
- the annoying "when will SvnBridge finally be even remotely resembling something less buggy and the public project be doing even minimal amounts of urgent project maintenance"? And this in the face of originally independent and very enthusiastic founding developers...
- the annoying "when will TFS finally have Web Access for other platforms that's not taking 2 (yes TWO!) minutes to login" (after a "your session timed out" forcibly-to-be-acknowledged non-hideable message box having appeared after some 10 seconds, to add insult to injury! So yeah you're FORCED to actively wait a full two minutes every time)
Buying into Mafia$oft "ecosystem" "infrastructure"!? HELL NO NEVER AGAIN!!
No need for so many compliculationistics: simply spec the job to be done using VSS - problem solved.
[reposting old semi-finished copy since Slashdot ate my entire edit window cold and UNRECOVERABLE when messing with Options dialog - and that's not the first time either of severe data loss with Slashdot website design!]
- no atomic checkins ("goddamn, somebody forgot to check in another file again" - "err what in fact I did"), no consistent changesets
- 5 GB max (official Microsoft recommendation), severe index size limitations
- may corrupt easily
- branching is ""supported"" but stay far away from it (said to be #1 source of corruption)
- NO BLAME MECHANISM (perusing the Changelog thus takes about 5 minutes each to hit the relevant spot)
- no 3-way merge AFAICS (and TFS still does not have that either)
- cross-platform support is laughable (there's SOS, but you easily hit trouble on each of RHEL, Ubuntu and Mac, and of course there's no 64bit platform anyway, and it's x86 only)
- performance is abysmal (depending on the moon phase, current mood of the Windoze server and number of busy colleagues, waiting for - yes, indeed - 1.5GB of source repository TAKES AGES (15 min. with VSS on Windows, up to 2 hours on SOS when synching database and subsequently fetching the entire tree); dito simply switching folders in GUI client treeview on SOS may take up to 5 minutes sometimes; this all as opposed to waiting for at most one minute for a kernel.org update to arrive from an internet server
- terrible check-out mode of operation, leading to that dreaded "all files readonly" standard way of operation as opposed to good SCM
- large slowdown of MSVS operation (lock contention!?!?) when additionally running the standalone VSS client
- keeps asking the same stupid questions over and over again rather than simply fetching files
- no distributed SCM (minor item)
- sos command line client is NOT parallelization-safe (needs addition of manual flock/lockfile mechanisms in a wrapper script for reliable use within a massively parallel build system)
- VSS DOES NOT LOCALLY DELETE FILES REMOVED FROM REPO (that one is a killer)
And please do me a favour and don't go cold-turkey to a TFS "upgrade"
but have a worthwhile attempt at reaching an informed decision of which ALM/SCM to actually use.
TFS, while most likely a lot better than VSS, is said to still have lots of rough spots.
See various Internet blogs (e.g. "TFS the Lotus Notes of ALM").
Suitable alternative names may be Mercurial, git, and ALM components such as Trac.
Trying to use braindead legacy applications on newer systems (UAC, registry/file system shadowing,
And don't get me started on laughable localization support shortcomings in Windows... (admittedly off-topic)
Even if it works, still doesn't quite solve the problem that probably only the more knowledgeable part of the users would know about such solutions and know to use them, though...
Have you considered that by giving such a final answer, YOU might actually be(come) the roadblock to properly working non-Windows/Mac access? ;-)
I applaud your efforts to try to get the support situation changed, but instead of stating a disillusioning "you can't", you should definitely switch to "I've tried a lot and so far I think you can't, but definitely feel free to show me otherwise.".
There might always come along someone who is brilliantly smarter than the entire rest and either a) figures out the specialty solution, or b) simply source-hacks together EVERYTHING required to "make it work anyway". At which point there's a nice solution ready for the taking...
And how many victims have seen their crucial personal search&rescue efforts hindered/inhibited (translation: indirect deaths) by that nuclear emergency and corresponding high risks within radiation-affected areas?
Simply using Valgrind's Helgrind is one way to do that - available only on the best operating system that money can _not_ buy, of course ;)
And the expert method here would be to Helgrind-run the unit tests themselves ;)
(plus standard Valgrind memcheck runs)
But generally spoken, if people happen to have severe trouble getting their code free from race conditions, then that easily suggests that they have a plenty weird relationship with threading and should rethink their model (think isolated worker threads with cleanly passed _work units_ data and resulting "completed" notifications).
Having some 30 different threads in an app, all with liberal - and then not even properly locked - access to global data is __NOT__ the way to do things.
Enable XP Accessibility setting "reduced colors" in control panel. Voila, all Excel cells are white, with none of the actual colors to be seen anywhere and, more importantly, no notice to be seen anywhere _either_.
Not to mention that this is really nice if for the life of it you cannot remember ever having enabled this Accessibility setting (I suspect it may have been enabled by SP3 update or some other quirky software), on an office computer that no kids have been getting at at any time.
luser "solution" of the problem? REINSTALL (yes, that's the same kind of reinstall as in the Ubuntu screen resolution case - which some users above even confirmed to _also_ happen with XP in certain situations - maybe driver-dependent?).
An experienced user (me) would be able to solve this very puzzling issue of lost colors on XP after >= 10 minutes of hardcore googling.
Morale of the story: use wx for GUI parts, but consider using something else for those non-GUI parts which it has trouble with or which are less likely to be properly served by a _GUI_ framework. For networking stuff, libcurl is VERY hard to beat (binaries for perhaps up to 50 (FIFTY) platforms, used by "all fortune 500", ...).
Spoken by someone who's been using wx all the time for 5 years now and generally likes it very much, but... well... you wouldn't let a butcher (GUI) take care about your dents (sockets) either, right!?!?
Real men do that with their entire GUI on the trip...
http://ldn.linuxfoundation.org/article/gtk2-let-application-follow-you
(NO, I haven't been using that, but it's damn tempting)
Yeah, and not much for the better. At least those things that I care about.
Umm, I don't think you realized that your two postings were just about polarly opposite statements... He wasn't talking about having Windows booted, he was mentioning 8 seconds _up to_ having Windows boot, whereas you...
Yup, by now it should be well-known that they are unable to design anything worth any crap. That M company somehow managed to spoil and mentally disrupt decades of programmers. See http://www.flounder.com/guiowner.htm etc. for radio button stuff and many other scary things.
On systems that are worth coding for, these things might be called inotify or dnotify or some such.
Anyway, thanks for documenting the way to do it on a certain platform.
And I'd say people who go ahead implementing it in such a low-bottom miserable way without any second thought still have lots and lots and lots more things to learn still.
Debian system, 1995, Athlon XP, eternally upgraded, gone through 1, 1.2, 4.3, 13, 30, 40, 80, 120GB. Yes, that's 8 HDDs copied over each time! Now do that with Windows. Judging from anecdotal evidence it will easily quit working with just one peripheral not aligned to the moon phase. I don't even remember whether any part in that box is still original... Sorry for spoiling your party ;)
An ALC268_ACER_ASPIRE_ONE entry is available in 2.6.28-rc2 patch_realtek.c, too, IOW ALSA git pull request obviously has been handled and I'll now have to test 2.6.28-rc2 to see whether it really works.
Oh it does?
Why then does my Acer One A110L (oh, look, that's the same machine even! Unless you've got 150L...) have issues on 8.10 with permanently non-working internal microphone (no matter what you try) and spotty ath5k wifi (it sorta works, but then sometime later you experience weird connection delays or Even OOPSes or so) and very, very, very problematic SD card reader support (resume may kill your SD partition table even on kernels as new as 2.6.27, and God Help You if you dared putting a swap partition on it even) and sharply reduced SSD performance after resume? (UDMA/33 instead of 66)??
_Usable_ sound support consists of both output _and_ input, you know... (so many people in the Acer One league seem to forget this). And especially given a nice builtin webcam...
I'm going to try ALSA source using module-assistant (that one is said to improve sound situation, but if it isn't fully fixed then I guess I'm going to take the valuable time to augment the realtek-specific patch file for intel-hda to properly support the seemingly unusual Acer One sound peripheral setup).
OK, the situation could certainly be much worse (especially considering an actually usable WLAN), but it's far, FAR from perfect.
Several issues listed here aren't Acer One specific actually (and I expect them to be fixed relatively soon), and the A110L actually is a very nice machine (minus stupid LCD display coating and somewhat lame performance of an SSD, especially in combination with relatively insufficient 512MB).
Plus, Vista has such a stunningly _high_ appcompat rating percentage that Wine would "NEVER" be able to come even distantly close to this extraordinary amount of compatibility with Windows applications...
Now please let me go die in peace from my lethal sarcasm overdose...
Linux has:
.NET (again, NOT) and WinFS (right, again NOT, has been in development for almost FIFTEEN years I think, since WinNT at least), ...
- completely standardized/open formats and protocols as opposed to the "completely brand new" file formats that Microsoft loves to create each year
- had fully working x86 64bit support 3(!) years earlier than Windows (well... minus a ton of 64bit-incapable proprietary offerings, which never get their things done in time anyway)
- had nice 3D desktops at least a year earlier than Vista
- enjoyed a nice Bluetooth stack as the very first operating system ever
- and several others I don't recount right now
So, pray tell me, this is "completely brand new" (implies lots of bugginess) how?
So, let's see how well Windows does in comparison:
- "completely brand new" office suite (entirely redesigned GUI) with "completely brand new" DRM-crippled file format
- "completely brand new" operating system with "completely brand new" yet currently less visible DRM crippling which extends all the way down to hugely increasing hardware development costs and complexity to make sure users will have lots of fun with self-crippling hardware (protected audio path etc.pp.)
- "completely brand new" Windows shell (oh, right, NOT available) and fully based on
- "completely brand new" Windows build (how much time did it take them to come up with a final build of Vista after the semi-decade-long desaster that was Vista development? 6 months?? And you think this will turn out to be stable??)
Conclusion: please tell me where your point is?
You forgot another important reason: PS/2 uses its very own and special interrupt, thus wasting valuable interrupt space. This is less of an issue now that many systems use APIC by default, though.
OK, I'll tell you: PS/2 is easier. ;)
Why? Easy, because PS/2 has a flat top on the round connector,
whereas your usual bog-standard USB connector has an uncharacteristic
design which won't tell you by hand which side to use (the only distinguishing
"finger" is *internal* to the connector).
Fine, now you could tell me that some PS/2 plugs are rotated by 90 degrees
thus plugging it in is difficult, but then some USB plugs are rotated, too.