The whole point of a "Personal Computer" is that all of the computing is under control of the person using it.
In the early days, even the idea of a protected mode OS was abhorrent to PC users because of the loss of control. Now the industry is basically proposing a timesharing system where some foreign entity has root on your system and is using your cycles and your databus in a way that you have no control over. Good bye DOS, hello Multics.
They are doing this out of fear of closed-box media consumption terminals from AOL and Sony. It's a terrible day when longtime personal computing advocates like Bill Gates have decided they must destroy the PC in order to save it.
Yes, in a recessionary economy, management will cut all costs down to the bone. And frankly, after 7 years of bloated ever-expanding IT budgets, this makes perfect sense -- after Y2K, ERP, B2C, B2B, CRM, and so on, what's left to do?
The last recession was in the early-90s. Established vendors like IBM, Lotus, and Novell got shoved out the door in favor of Microsoft because they were the cheap solution. A similar process happened at the high-end where mainframes and VAXes were replaced by cheap UNIX systems.
When the economy picked up, the cheap-o vendors (MS, Sun) reaped most of the rewards.
Now we're entring another cost-cutting cycle. Linux will be adopted because it's cheap, but once adopted, it will be there to stay. Sure, the culture of "Free" is depressing if you are in the IT business, but that's just the logical extent of what's been happening all along.
You don't get it. The product being sold by spam isn't Herbal Viagra or College Diplomas -- it's the spam itself.
It's a pyramid scheme. It's not about selling the product. It's about convincing people to pay you to sell their product through spam, to buy your address lists, or buy your spam software.
It's not about the people stupid enough to buy, it's about the people stupid enough to think "With all this spam, someone out there must be buying."
A large percentage of spam doesn't even have a valid contact address/url/phone. It's purely about claiming to prospective clients that you can deliver X messages or have Y valid addresses.
So, go ahead and convince grandma not to buy any spam prodcuts. Great. Meanwhile these guys are on a sales arms-race that will eventually render standard netmail useless.
Easier fix would be to configure Mozilla to disallow all Mouseover events
Since mouseover menus are arguably one of the two "good" uses of JavaScript (along with form validation), at that point you've gotten into a arms race and pretty much neutered the feature you are trying to save.
The real solution is to disable Javascript entriely and only enable it for a whitelist of sites where it's really needed (a few ecomm sites).
Oddly, it's much easier to do this in IE than Mozilla.
However, as Office becomes more and more of a client-server^W^Wweb-services product, the file itself will become less and less accessible.
With a little setup magic, the usual shitty workflow of mailing each other DOC files could easily be be replaced by mailing each other URLs to a centralized content/licence mgmt server.
I got my mother one of these about 3-4 years ago (because I hate providing tech support), and guess what? I have NEVER had to provide tech support. Whoop! So in my book these things are great.
(Once the thing borked, but the WebTV people talked her through reseting it, which I gather involved redownloading the OS.)
So, in my book these devices are great. Sure Mom is "permanently challenged" (is that PC for "retarded"?), but she gets her e-mail when the little light blinks and that's all that really matters.
Meanwhile her friends are scared shitless of their iMacs given to them by their 'helpful' sons and are always taking them in for "repair".
The one downside to these things is that the WebTV Internet is completely spamvertised. Everything about it is designed to encourage you to go to MS's marketing "partners" -- you can't set your home page, you can't set your search page, it's difficult to enter URLs, etc. But mom uses it mostly for e-mail, so it's OK.
The other downside is that MS takes a big loss on these things, they haven't proved very popular, dot-com stuff like this ain't cool anymore, and eventually the service will probably be disconnected (or there will be a XBox Forced Upgrade).
The thing to understand is that WebTV is a Dumb Terminal for End lUsers, which one would think fits the Unix jockey model of the world perfectly.
I agree that the "Personal Computer" is pretty much done.
However, I'm not sure if entertainment needs will drive the next wave in personal computing. It's just as likely that specialized PC-like Media Consumption Terminals (such as Tivo and XBox) will become the next adoption wave, and due to specialized software, the 'wait' that we've conqured in PCs won't be a problem. PCs then become an extention of the broader consumer electronics market, and that includes demands placed on CPUs, hard drives, and so on.
(Note that I say "PC-like", because depite the hw, these things will be DRMed to hell and (worse) will not be end-user programmable.)
On the other hand, for what the 80s lacked in Moore's Law, it made up for in creative hardware designs. Hopefully a hardware stagnation will inspire the range of stuff we saw back then, rather than just turning everything from your TV to your Oven into a casterated 80%-IBM-compatible PC.
I don't think this is true. this "commoditisation of the hardware market" will only make the percentage of pro-quality equipment drop but the raw number will probably increase.
That was true in the boom years of the 1990s, but it wasn't true in the stagnent 1980s. Then, the high-end (think UNIX Workstations, Macs, etc) got faster but also more and more expensive. Meanwhile, the low-end (think 8086 or 286-based PCs) never got any faster and only gradually got cheaper.
The fact that you could buy a new XT-type 8086-based PC from IBM in 1990, tells you that without compelling power-hungry software, there's no real nead for the hardware to progressively faster. I think this $200 Via chip thing is likely the start of a trend.
I 'rolled my own' version of this many years ago. All you need is the "Network client for DOS" (see any NT/W2K Server CD), and the "NDIS2" driver disk for your card, and a little time on your hands.
"So my advice to you is give it a try. I gave up on the 9x series the day NT4 came out and never looked back."
Yeah, so did I, but there were several reasons that "not everyone" could do that.
Historically, NT's IDE drivers were shit (the printed docs even made references to "SCSI is the future"). It was also incompatible with a large number of Pentium-era chipsets, which meant you had to chose your hardware very carefully.
However, if you had 64MB and a supported SCSI system, NT4 just blew Win 98 out of the water.
Unfortunatly, MS torpedo'd plans for a NT4 "Home Edition" because they were waiting on ActiveDirectory and a bunch of other complex enterprise features in "NT5".
So they left NT DirectX, Plug-n-Play, decent IDE, and Power Mgmt on the shelf for 3 years. Windows 98 just should never have existed.
Since you could never get it to network to anything
Hey, just yesterday I used an old DOS Netboot disk to copy some files over to a machine I was setting up.
Microsoft can obsolete DOS, but as of yet they haven't introduced a replacement that can get a machine on the network with a single floppy disk. I doubt they'll ever get a version of NT working from read-only media.
Actually, when ActiveX first came out (IE3), they didn't even have the "Do You Trust?" dialog. Led to some very nasty exploits because it ran whatever was signed by anyone.
The main issue is that it's impossible to sandbox C/C++ code that's running in the same process space as your browser. The ONLY solutions are something VM-based like Java or.NET -- or to radically alter how memory-protection works on multi-user OSes (drumroll... Palladium).
This is equally true for Mozilla Plugins as it is for ActiveX controls, BTW, except it looks like that Mozilla is missing the signature layer that ActiveX has.
First of all the idea of using XML to describe formatted documents is not exactly a "Eureka!" idea -- it's one of the things XML is designed for.
Calling MSO's format a "ripoff" without examining the implementation is retarded.
Second, "VFolders" is a ripoff of a old old old Lotus Notes feature called "Views", the lack of which has been a complaint about Outlook for years.
Claiming that this feature was "lifted directly" from some obscure Unix mail client when it's been in one of the most-widely used corporate mail client for a decade or more just shows your ignorance.
There's plenty of ways to farm karma around here without resorting to making shit up. NET Server's ability to boot in console mode and be admined from the command-line is a good place to start when looking for UI ideas borrowed from Unix.
Another example in this vein is Microsoft Word, which still has the *exact same* crappy modal dialog box layout that it had in 1994.
Not because MS doesn't know that it's crappy (Word for Mac has nice pallettes), but because if anyone had to spend a nickle on corporate training on top of the new licencing program, they'd shit bricks.
So all of the "freedom to innovate" is spent on crappy Clippy features that everyone disables and never invested into basic improvements in the program interface. (I haven't seen the O11 beta, but the last 4 versions are enough of a track-record.)
No, MS did examine the "Big Picture", and the Big Picture is that MIME types suck as a filetyping mechanism for web clients.
None of the major operating systems (Mac/Windows/Unix) use MIME natively, and the predominant web server platform (Unix) generally doesn't know a single thing about common Windows/Mac filetypes and relys on fragile manual mapping under the control of a fat bearded guy who couldn't give a fuck. Client-control is the correct answer (and the one that's used in virtually all mail-clients).
Not being able to download a Mozilla-based browser from a Mozilla-based browser should tell you something -- that Mozilla needs to start thinking 'big picture' themselves.
The one part you got right is that any given Microsoft implementation is going to make everyone else look "broken", but when it really is broken, that's a good thing. Netscape hasn't examined crap in this situation. It's the rattled M$-bashing tone that landed you on my foes list, but I browse at -1, so don't worry about it.
Ultimately MS made the right decision in doing client-side MIME-mappings -- Your average (Unix) webserver will never have a completely comprehensive set of MIME-mappings, and from the user's perspective it makes the browser appear "broken".
It took years before common filetypes like DOC were setup correctly. I've seen DOC mapped to application/x-lotus-manuscript for example -- apparently that MIME mapping file was like 20 years old. If they can't get DOC right, what are the chances that a Moz user will ever get a MPEG4 or an AutoCAD file off that server?
Even recently, I couldn't directly download early Chimera builds because Chimera's own host didn't recognize the Mac SMI archive format.
The problem with MS's implementation is that it doesn't follow a sane policy, and uses some obscure combination of MIME-types, extentions, and magicbits sniffing. Confused filetype rules contributed to Nimba, but the real problem was/is a Open() call that didn't distinguish between EXEs and JPGs.
If mozilla could implement this with sane rules and no stupidity, it would make their browser appear much less "broken" to certain users.
Do some research into what.NET offer web developers. I will give you a hint: its not embedded "applets".
First of all, "applets" are completely possible through IE's normal ActiveX download mechanism. Unlike with Java, there's not much hype associated with it in.NET, but there's also very little client-deployment as of yet (did MS just start bundling it in OEM Windows?).
Second -- System.Web.UI.WebControls -- Note the non-HTMLish (instead VB-ish) object model used to render HTML. Note that many of the controls duplicate System.Web.UI.HTMLControls. Right now these controls render only HTML/JavaScript, but I can't see any technical reason that these controls couldn't use browser-sniffing and some SOAP to drive pre-canned.NET 'applets' for a "richer" Microsoft experience. Anyway, if I was BillG, that's what I'd be thinking.
Furthermore, most of those INITs/CDEVs from back in the golden days of Mac UI hacking (as Wired tells it), were using undocumented and unsupported functionality, and were certainly in no way encouraged by Apple.
Forget breaking going from System 7 to MacOS 9, many of those hacks broke on minor revisions of the OS or even on new hardware. "Conflict Catcher" was pretty much a standard piece of the troubleshooting toolkit (manually finding INIT conflicts being a bitch).
And to counter happystink's point that it was mainly power users that used these things, there was quite a bit of commercial software that relied on dirty hacks that broke all the time. Now Contact Manager being the classic one that forced users to pay for upgrades every new OS release.
The built-in POSIX subsystem is pretty much worthless, but it's also been obsoleted by a $100 Microsoft product called "Services For UNIX" (formerly Interix), which provides a UNIX personality subsystem on top of NT's kernel, and ships with GCC (+source), korn shell etc.
I can't say much about this thing, but before MS bought it up, it was UNIX(tm) certified by The Open Group.
SFU/Interix is being sold as a migration solution, not a development platform, but in theory there's no reason one couldn't use it that way. According to MS, SFU was used in the Hotmail migration from FreeBSD.
Rapping MS's knuckles for something that's been rectified for a few years is unfair. I would expect the built-in POSIX to go away with the next server release, because according to MS it's a security best practice to remove it.
This was exactly the situation ten years ago when the first wave of personal computer purchases was dying down and everybody had the DOS machines they needed
Yeah, and PC sales were in the toliet for 5 years or more.
Then Windows came along and shortly thereafter the net entered the public conscioussness and there was a real reason to upgrade
The number one reason for upgrading from the i386 era to the Penitum III 500Mhz era was Windows itself, with MS Office coming in at #2. It looks like that demon has actually been defeated. I really doubt we will ever again see software that universally makes almost everyone want to continually upgrade.
Instead, I see things going the other way -- cheaper, quieter, cooler, and smaller. Imagine a little $100 box that you just toss into a cubical, connect a monitor to and go. You don't even know or care what the Ghz value is - it's in fine print in the back of the manual.
I remember LOTS of yuppie boomers who learned old wordperfect just fine.. and that's certainly not wysiwig.
Many did, however most management staff did not.
Obscure keypresses, hidden markup codes, they understood it all.. and some were really good at it.
True, WordPerfect was designed from the ground-up to be an "Expert" interface. It was specifically designed for professional typists. Even so, 90% of the users never figured out the more complex stuff like font formatting and tables -- stuff that virtually everyone can do in (say) Word.
What's important to understand is that the move from WordPerfect to Word lead to a fundemental restructuring of who types stuff in American business. WordPerfect was still grounded in the days of personal secretaries and typing pools. With Word, these 'yuppie' managers now have to type their own memos (although you could argue that e-mail was the final catalyst in this trend).
This lead to the immedate drop in admin salaries and training costs, which from a business standpoint was a good thing.
I don't buy this stuff about how unix is hard and other stuff is easy...
Just because people HAD to learn WordPerfect, doesn't mean they liked it. I certainly didn't -- the program basically sucked and relied on the user memorizing a bunch of pointess crap. Even in the context of DOS console apps, WordStar was much more user-friendly.
So, yeah, it's possible for people to learn (say) Emacs. But what's the point in doing so? Defeating the computer? Lowering productivity? Impressing the zitfaced IT geeks?
If your CPQ is like my CPQ keyboard, it isn't SCSI, it just uses a similar connector to a Mac PowerBook SCSI plug. It's plain ol PS/2 with the audio wires running out to a speaker and some plugs on the keyboard.
(Apologies if you do actually have a SCSI keyboard!)
Let's keep in mind that Linux users who find bugs or issues are far more likely to report them, document them, publicize them, and share them.
Good point, but it would be better if you took it out of the context of the "users" and put it in the context of the developers. It works out more like this:
Open Source Project X Developer (who may well be on someone's payroll) finds a previously unknown security bug. He patches the bug and informs RedHat and other distro vendors, who then issue a security bulletin. One strike against Linux in the security count.
Meanwhile Microsoft Product Y Developer finds 100 unknown security security bugs in his big Feburary cleanup period. They are all rolled in to service pack 3. Microsoft issues a bulletin recommending all customers upgrade immediately. Zero strikes against Microsoft.
So you are counting ALL security bugs on the Linux side verus only publically reported security bugs on the Windows/Solaris/whoever side.
(Furthermore, it seems nobody considers local root exploits on Windows to be that big of a deal. I remember when RedHat put out multiple advisories for vi, joe, ed, and a bunch of other editors for a temp file vulnerability. [You'd think that "ed" would be rock solid by now...] Would that sort of thing even be considered a bug on the Windows side?)
Slashdot Mirror
The whole point of a "Personal Computer" is that all of the computing is under control of the person using it.
In the early days, even the idea of a protected mode OS was abhorrent to PC users because of the loss of control. Now the industry is basically proposing a timesharing system where some foreign entity has root on your system and is using your cycles and your databus in a way that you have no control over. Good bye DOS, hello Multics.
They are doing this out of fear of closed-box media consumption terminals from AOL and Sony. It's a terrible day when longtime personal computing advocates like Bill Gates have decided they must destroy the PC in order to save it.
Yes, in a recessionary economy, management will cut all costs down to the bone. And frankly, after 7 years of bloated ever-expanding IT budgets, this makes perfect sense -- after Y2K, ERP, B2C, B2B, CRM, and so on, what's left to do?
The last recession was in the early-90s. Established vendors like IBM, Lotus, and Novell got shoved out the door in favor of Microsoft because they were the cheap solution. A similar process happened at the high-end where mainframes and VAXes were replaced by cheap UNIX systems.
When the economy picked up, the cheap-o vendors (MS, Sun) reaped most of the rewards.
Now we're entring another cost-cutting cycle. Linux will be adopted because it's cheap, but once adopted, it will be there to stay. Sure, the culture of "Free" is depressing if you are in the IT business, but that's just the logical extent of what's been happening all along.
You don't get it. The product being sold by spam isn't Herbal Viagra or College Diplomas -- it's the spam itself.
It's a pyramid scheme. It's not about selling the product. It's about convincing people to pay you to sell their product through spam, to buy your address lists, or buy your spam software.
It's not about the people stupid enough to buy, it's about the people stupid enough to think "With all this spam, someone out there must be buying."
A large percentage of spam doesn't even have a valid contact address/url/phone. It's purely about claiming to prospective clients that you can deliver X messages or have Y valid addresses.
So, go ahead and convince grandma not to buy any spam prodcuts. Great. Meanwhile these guys are on a sales arms-race that will eventually render standard netmail useless.
Apparently reading is all that hard.
The issue it that Mozilla has no easy way to build a whitelist of Javascript sites, unlike IE. You need to hack around in a prefs file.
Easier fix would be to configure Mozilla to disallow all Mouseover events
Since mouseover menus are arguably one of the two "good" uses of JavaScript (along with form validation), at that point you've gotten into a arms race and pretty much neutered the feature you are trying to save.
The real solution is to disable Javascript entriely and only enable it for a whitelist of sites where it's really needed (a few ecomm sites).
Oddly, it's much easier to do this in IE than Mozilla.
It most likely will not be less "open" than DOC.
However, as Office becomes more and more of a client-server^W^Wweb-services product, the file itself will become less and less accessible.
With a little setup magic, the usual shitty workflow of mailing each other DOC files could easily be be replaced by mailing each other URLs to a centralized content/licence mgmt server.
I got my mother one of these about 3-4 years ago (because I hate providing tech support), and guess what? I have NEVER had to provide tech support. Whoop! So in my book these things are great.
(Once the thing borked, but the WebTV people talked her through reseting it, which I gather involved redownloading the OS.)
So, in my book these devices are great. Sure Mom is "permanently challenged" (is that PC for "retarded"?), but she gets her e-mail when the little light blinks and that's all that really matters.
Meanwhile her friends are scared shitless of their iMacs given to them by their 'helpful' sons and are always taking them in for "repair".
The one downside to these things is that the WebTV Internet is completely spamvertised. Everything about it is designed to encourage you to go to MS's marketing "partners" -- you can't set your home page, you can't set your search page, it's difficult to enter URLs, etc. But mom uses it mostly for e-mail, so it's OK.
The other downside is that MS takes a big loss on these things, they haven't proved very popular, dot-com stuff like this ain't cool anymore, and eventually the service will probably be disconnected (or there will be a XBox Forced Upgrade).
The thing to understand is that WebTV is a Dumb Terminal for End lUsers, which one would think fits the Unix jockey model of the world perfectly.
I agree that the "Personal Computer" is pretty much done.
However, I'm not sure if entertainment needs will drive the next wave in personal computing. It's just as likely that specialized PC-like Media Consumption Terminals (such as Tivo and XBox) will become the next adoption wave, and due to specialized software, the 'wait' that we've conqured in PCs won't be a problem. PCs then become an extention of the broader consumer electronics market, and that includes demands placed on CPUs, hard drives, and so on.
(Note that I say "PC-like", because depite the hw, these things will be DRMed to hell and (worse) will not be end-user programmable.)
On the other hand, for what the 80s lacked in Moore's Law, it made up for in creative hardware designs. Hopefully a hardware stagnation will inspire the range of stuff we saw back then, rather than just turning everything from your TV to your Oven into a casterated 80%-IBM-compatible PC.
I don't think this is true. this "commoditisation of the hardware market" will only make the percentage of pro-quality equipment drop but the raw number will probably increase.
That was true in the boom years of the 1990s, but it wasn't true in the stagnent 1980s. Then, the high-end (think UNIX Workstations, Macs, etc) got faster but also more and more expensive. Meanwhile, the low-end (think 8086 or 286-based PCs) never got any faster and only gradually got cheaper.
The fact that you could buy a new XT-type 8086-based PC from IBM in 1990, tells you that without compelling power-hungry software, there's no real nead for the hardware to progressively faster. I think this $200 Via chip thing is likely the start of a trend.
Looks like a slick setup, thanks for the tip.
I 'rolled my own' version of this many years ago. All you need is the "Network client for DOS" (see any NT/W2K Server CD), and the "NDIS2" driver disk for your card, and a little time on your hands.
"So my advice to you is give it a try. I gave up on the 9x series the day NT4 came out and never looked back."
Yeah, so did I, but there were several reasons that "not everyone" could do that.
Historically, NT's IDE drivers were shit (the printed docs even made references to "SCSI is the future"). It was also incompatible with a large number of Pentium-era chipsets, which meant you had to chose your hardware very carefully.
However, if you had 64MB and a supported SCSI system, NT4 just blew Win 98 out of the water.
Unfortunatly, MS torpedo'd plans for a NT4 "Home Edition" because they were waiting on ActiveDirectory and a bunch of other complex enterprise features in "NT5".
So they left NT DirectX, Plug-n-Play, decent IDE, and Power Mgmt on the shelf for 3 years. Windows 98 just should never have existed.
If it's the XBox, it ought to come with XCOPY, right?
Since you could never get it to network to anything
Hey, just yesterday I used an old DOS Netboot disk to copy some files over to a machine I was setting up.
Microsoft can obsolete DOS, but as of yet they haven't introduced a replacement that can get a machine on the network with a single floppy disk. I doubt they'll ever get a version of NT working from read-only media.
Actually, when ActiveX first came out (IE3), they didn't even have the "Do You Trust?" dialog. Led to some very nasty exploits because it ran whatever was signed by anyone.
.NET -- or to radically alter how memory-protection works on multi-user OSes (drumroll... Palladium).
The main issue is that it's impossible to sandbox C/C++ code that's running in the same process space as your browser. The ONLY solutions are something VM-based like Java or
This is equally true for Mozilla Plugins as it is for ActiveX controls, BTW, except it looks like that Mozilla is missing the signature layer that ActiveX has.
First of all the idea of using XML to describe formatted documents is not exactly a "Eureka!" idea -- it's one of the things XML is designed for.
Calling MSO's format a "ripoff" without examining the implementation is retarded.
Second, "VFolders" is a ripoff of a old old old Lotus Notes feature called "Views", the lack of which has been a complaint about Outlook for years.
Claiming that this feature was "lifted directly" from some obscure Unix mail client when it's been in one of the most-widely used corporate mail client for a decade or more just shows your ignorance.
There's plenty of ways to farm karma around here without resorting to making shit up. NET Server's ability to boot in console mode and be admined from the command-line is a good place to start when looking for UI ideas borrowed from Unix.
Another example in this vein is Microsoft Word, which still has the *exact same* crappy modal dialog box layout that it had in 1994.
Not because MS doesn't know that it's crappy (Word for Mac has nice pallettes), but because if anyone had to spend a nickle on corporate training on top of the new licencing program, they'd shit bricks.
So all of the "freedom to innovate" is spent on crappy Clippy features that everyone disables and never invested into basic improvements in the program interface. (I haven't seen the O11 beta, but the last 4 versions are enough of a track-record.)
No, MS did examine the "Big Picture", and the Big Picture is that MIME types suck as a filetyping mechanism for web clients.
None of the major operating systems (Mac/Windows/Unix) use MIME natively, and the predominant web server platform (Unix) generally doesn't know a single thing about common Windows/Mac filetypes and relys on fragile manual mapping under the control of a fat bearded guy who couldn't give a fuck. Client-control is the correct answer (and the one that's used in virtually all mail-clients).
Not being able to download a Mozilla-based browser from a Mozilla-based browser should tell you something -- that Mozilla needs to start thinking 'big picture' themselves.
The one part you got right is that any given Microsoft implementation is going to make everyone else look "broken", but when it really is broken, that's a good thing. Netscape hasn't examined crap in this situation. It's the rattled M$-bashing tone that landed you on my foes list, but I browse at -1, so don't worry about it.
Ultimately MS made the right decision in doing client-side MIME-mappings -- Your average (Unix) webserver will never have a completely comprehensive set of MIME-mappings, and from the user's perspective it makes the browser appear "broken".
It took years before common filetypes like DOC were setup correctly. I've seen DOC mapped to application/x-lotus-manuscript for example -- apparently that MIME mapping file was like 20 years old. If they can't get DOC right, what are the chances that a Moz user will ever get a MPEG4 or an AutoCAD file off that server?
Even recently, I couldn't directly download early Chimera builds because Chimera's own host didn't recognize the Mac SMI archive format.
The problem with MS's implementation is that it doesn't follow a sane policy, and uses some obscure combination of MIME-types, extentions, and magicbits sniffing. Confused filetype rules contributed to Nimba, but the real problem was/is a Open() call that didn't distinguish between EXEs and JPGs.
If mozilla could implement this with sane rules and no stupidity, it would make their browser appear much less "broken" to certain users.
Do some research into what .NET offer web developers. I will give you a hint: its not embedded "applets".
.NET, but there's also very little client-deployment as of yet (did MS just start bundling it in OEM Windows?).
.NET 'applets' for a "richer" Microsoft experience. Anyway, if I was BillG, that's what I'd be thinking.
First of all, "applets" are completely possible through IE's normal ActiveX download mechanism. Unlike with Java, there's not much hype associated with it in
Second -- System.Web.UI.WebControls -- Note the non-HTMLish (instead VB-ish) object model used to render HTML. Note that many of the controls duplicate System.Web.UI.HTMLControls. Right now these controls render only HTML/JavaScript, but I can't see any technical reason that these controls couldn't use browser-sniffing and some SOAP to drive pre-canned
Furthermore, most of those INITs/CDEVs from back in the golden days of Mac UI hacking (as Wired tells it), were using undocumented and unsupported functionality, and were certainly in no way encouraged by Apple.
Forget breaking going from System 7 to MacOS 9, many of those hacks broke on minor revisions of the OS or even on new hardware. "Conflict Catcher" was pretty much a standard piece of the troubleshooting toolkit (manually finding INIT conflicts being a bitch).
And to counter happystink's point that it was mainly power users that used these things, there was quite a bit of commercial software that relied on dirty hacks that broke all the time. Now Contact Manager being the classic one that forced users to pay for upgrades every new OS release.
The built-in POSIX subsystem is pretty much worthless, but it's also been obsoleted by a $100 Microsoft product called "Services For UNIX" (formerly Interix), which provides a UNIX personality subsystem on top of NT's kernel, and ships with GCC (+source), korn shell etc.
I can't say much about this thing, but before MS bought it up, it was UNIX(tm) certified by The Open Group.
SFU/Interix is being sold as a migration solution, not a development platform, but in theory there's no reason one couldn't use it that way. According to MS, SFU was used in the Hotmail migration from FreeBSD.
Rapping MS's knuckles for something that's been rectified for a few years is unfair. I would expect the built-in POSIX to go away with the next server release, because according to MS it's a security best practice to remove it.
This was exactly the situation ten years ago when the first wave of personal computer purchases was dying down and everybody had the DOS machines they needed
Yeah, and PC sales were in the toliet for 5 years or more.
Then Windows came along and shortly thereafter the net entered the public conscioussness and there was a real reason to upgrade
The number one reason for upgrading from the i386 era to the Penitum III 500Mhz era was Windows itself, with MS Office coming in at #2. It looks like that demon has actually been defeated. I really doubt we will ever again see software that universally makes almost everyone want to continually upgrade.
Instead, I see things going the other way -- cheaper, quieter, cooler, and smaller. Imagine a little $100 box that you just toss into a cubical, connect a monitor to and go. You don't even know or care what the Ghz value is - it's in fine print in the back of the manual.
I remember LOTS of yuppie boomers who learned old wordperfect just fine.. and that's certainly not wysiwig.
Many did, however most management staff did not.
Obscure keypresses, hidden markup codes, they understood it all.. and some were really good at it.
True, WordPerfect was designed from the ground-up to be an "Expert" interface. It was specifically designed for professional typists. Even so, 90% of the users never figured out the more complex stuff like font formatting and tables -- stuff that virtually everyone can do in (say) Word.
What's important to understand is that the move from WordPerfect to Word lead to a fundemental restructuring of who types stuff in American business. WordPerfect was still grounded in the days of personal secretaries and typing pools. With Word, these 'yuppie' managers now have to type their own memos (although you could argue that e-mail was the final catalyst in this trend).
This lead to the immedate drop in admin salaries and training costs, which from a business standpoint was a good thing.
I don't buy this stuff about how unix is hard and other stuff is easy...
Just because people HAD to learn WordPerfect, doesn't mean they liked it. I certainly didn't -- the program basically sucked and relied on the user memorizing a bunch of pointess crap. Even in the context of DOS console apps, WordStar was much more user-friendly.
So, yeah, it's possible for people to learn (say) Emacs. But what's the point in doing so? Defeating the computer? Lowering productivity? Impressing the zitfaced IT geeks?
If your CPQ is like my CPQ keyboard, it isn't SCSI, it just uses a similar connector to a Mac PowerBook SCSI plug. It's plain ol PS/2 with the audio wires running out to a speaker and some plugs on the keyboard.
(Apologies if you do actually have a SCSI keyboard!)
Let's keep in mind that Linux users who find bugs or issues are far more likely to report them, document them, publicize them, and share them.
Good point, but it would be better if you took it out of the context of the "users" and put it in the context of the developers. It works out more like this:
Open Source Project X Developer (who may well be on someone's payroll) finds a previously unknown security bug. He patches the bug and informs RedHat and other distro vendors, who then issue a security bulletin. One strike against Linux in the security count.
Meanwhile Microsoft Product Y Developer finds 100 unknown security security bugs in his big Feburary cleanup period. They are all rolled in to service pack 3. Microsoft issues a bulletin recommending all customers upgrade immediately. Zero strikes against Microsoft.
So you are counting ALL security bugs on the Linux side verus only publically reported security bugs on the Windows/Solaris/whoever side.
(Furthermore, it seems nobody considers local root exploits on Windows to be that big of a deal. I remember when RedHat put out multiple advisories for vi, joe, ed, and a bunch of other editors for a temp file vulnerability. [You'd think that "ed" would be rock solid by now...] Would that sort of thing even be considered a bug on the Windows side?)