We should compile a hit list of spammers and their addresses, the top 42 of them, on a hit list website. Tell them to stop sending spam or we'll shoot them with a high powered water gun through their kitchen window. Then, as we shoot them, and CNN (conservative news network) reports it nationally, cross each one off the list.
The reason why setup.exe (= installers in general) work is that they are foolproof and brainless and work the first time.
[cough cough. Well, not always. yes sometimes they don't work, sometimes they crash, sometimes you have to coax them along, sometimes they break stuff that used to work, sometimes you gotta reformat and reinstall the whole os...]
./configure; make; make install
is not the same:
1) it takes more typing
2) the chances of brainless success are lower, and you, the human installer, are assumed to have the skill to fix what's broken and coax it along.
This is why and when software is distributed as source: it's less likely to work as-is, therefore, for a programmer who can hack, the means are there to fight toward success.
A canned binary does not have this flexibility. A canned binary for x86 will not work on PowerPC and vice versa. A buildable source distro will.
I got a 4 CD set of software, bsd ports, for Mac OS X. Sources. Scanning through the whole thing one evening, I loaded up everything I thought would be useful. I mostly wanted bash. Most of it just didn't build as it was, including bash, and I didn't take the time to think and fight - if it didn't build, I went on. At the end of the evening, I'd gained one piddly little program that I don't even remember. Nothing else worked - not brainlessly at least. Some day I'll attack the bash sources again.
BY comparison, most Mac production software from apple or commercial publishers, installs a binary and works beautifully the first time. We're talking, Click. We're talking, the manual sez it won't work on a non-Apple USB, but it does anyway. But of course only for selected OS versions on PPC, and the installer took more work on the part of the publisher. There's no free lunch.
PS: I'm not worried about computers taking oover the world. They need us humans to fix the bugs and run the installers and build scripts.
> This biggest problem is normally the Mac to be honest, it doesn't read others, and others can't read it, if you get what I mean.
Not true, at least not completely true. I've gotten my (dorky old) firewire drive to move between a Mac (G3 gray with addon orangmicro firewire, running MacOS 9) and a PC (vaio laptop, win 98). This was in mid 2000 so stuff has upgraded obviously.
If partitioned and formatted on Win 98, the disk has a DOS partition table (primary and logical partitions) and I put both FAT 16 and FAT 32 partitions on it. Came up just fine in MacOS, just like floppys and zips that are dos formatted. Although MacOS can't format one of the partitions to be HFS without...
If partitioned under MacOS, with Mac partition tables and HFS and/or HFS+ partitions, Windows treated it like it was from mars. unformatted.
Haven't tried under Mac OS X yet, could be busted. I got more gigs than I can use right now these days. Yuck, just stepped on another disk drive!
actually, 1394 started out with four speeds:
S100, S200 and S400. By now most everybody is at S400 = 400mBITS/sec as you all say. My older vaio has some ports at S200 and some at S400.
Apple and others are pounding away trying to get S800 and S1600 going; I'm hoping they're shipping by the time USB 2.0 is shipping. And of course the SCSI people aren't asleep, developing YetMoreStupendouslyWide variations. I expect FireWire and SCSI to play leapfrog, and USB to stay behind.
But yes, as everyone says, the bus and disk and kernel/DMA (or whatever) are also bottlenecks in the way. YMMV. Solaris x86 ferinstance ships with DMA turned OFF for some reason but they tell how to turn it on (i dunno if it does 1394). Bleah. Trust benchmarks and your own personal experience.
The 6-pin 1394 has the power pins in the USB-sized plug - think disks. The tiny 4-pin 1394 doesn't - think vid cameras. Cables of 6-6, 6-4 and 4-4 kinds are available often at the usual inflated prices.
>Want to kill Microsoft? Sap it's growth, which is in server OS's and embedded systems (XBox, Pocket PC, etc.)
One of the things I've learned about business is, in general, it's impossible to "drive" anybody out of business from the outside. Companies that get derailed do so by their own mismanagement (often under stress, sometimes from a competitor). Being well-run means responding to a changing world and finding new revenue sources if old ones dry up.
In Microsoft's case, they have revenue from many, many different sources. Smothering one of their major revenue sources is possible. Smothering all of them at the same time, and squelching any new initiatives they have, is highly unlikely. Meanwhile, they have cash in the bank, too.
If you want a goal, just stick to your own business. Make sure your own revenue keeps coming. That's something you do have control over.
>Go shoot your precious linux server with a.44, and see if it's still up. I'd guess the answer is no. Doing the same with a properly configured SunFire 15k would result in a high replacement cost, but an up and running system nonetheless.
That would make a cool TV ad!!! "Hi, My name is Scott McNealy, president of Sun Microsystems. We're here today in Taloqan, Afghanistan, to demonstrate the bullet-proof reliability of Sun servers. We brought along a SunFire 15000, serving web pages at this very moment, as George, our test engineer, demonstrates." [george clicks on a hotlink on his pc, another page comes up.] "Speaking of bullet-proof, our marksman here, Hafez Mohammed, from the Northern Alliance, is armed with a genuine Soviet-made Kalishnikov rifle..."
right. To clarify, I took that quote directly from the article. That's what the Intel guy said. An anectdote, not a data point. Actually measuring such a thing is really hard and subjective and depends on so many particulars, maybe it's unmeasurable.
For instance, maybe the windows sysadmins cost less per hour. So the beancounters have a number. But how much "management" do you get per sysadmin per hour? How much "management" does each system need? What's the extra cost of cleaning up from Nimda? Will Linux ever have a worm as bad as Nimda? (unlikely but that's what MS thought too.) What's the extra cost of reinstalling NT six times to get it "right"? Or Linux for that matter? Well if everybody had been using XYZ motherboards instead of PQR motherboards, you wouldn't have run into that hassle. It's the STU raid system! With the ACLs. If you hadn't messed with the ACLs, or had gotten a different RAID system...
I don't know how to measure this, but I know that measuring it is highly vulnerable to bias. EG it's easy for MS to come up with data proving whatever.
I've been reading/. for a year or so. I just kept on forgetting my account name, and never bothered to read how the whole point system works and the popup menus and which reply button replies to what... as usual, in this culture, if you don't RTFM, and play around with it a bit to figure out what TFM doesn't tell you, you're lost and intimidated.
> a name like "OsamaBinLogin"
Now THAT's a user name I can remember!!! Cute, huh? thought it up myself, and very proud of it thanyouverymuch. I just hope the FBI doesn't read something into it and haul my ass downtown. (Dear FBI: really, I'm just kidding around. I'm an agnostic cleanshaven honkie who's never used a firearm.)
>and the typical troll "normal people can't understand Linux"
It's got to be said. And I think more than a few people on this board agree with it. And if you get away from/. to talk to some non-technical people, you'll find more people who agree. This is what happens to the bearer of bad news.
> and not one moderatoor has modded this down? I'm disgusted.
are you trying to be INTIMIDATING?!?!?! Trying to make people ASHAMED for disagreeing with you?
>Macintosh may be more soothing -- appealing shapes and colors, friendlier text and error messages (bombs and frowns anyone?)...
I don't think "soothing" has anything to do with it. Amateur UI designers keep thinking they are making software "easy" when they stick nice but incomprehensible icons on the same unexplained concepts. That's why Microsoft invented tooltips - so you can figure out what the icon does.
Besides, I think MacOS and Windows keep leapfrogging each other on the "graphically cool" aspect. If it's a frozen photo in a magazine, that sells.
The frowning Mac face originally came about because the original Macintosh hardware was language-independent. you boot up with a French system, everything is in French. If there's an error booting up, you ain't got no text to spit out an error message. So you either put up an icon that makes it obvious what's happening, or you say in english, "this computer made by ugly americans who can't comprehend a foreigner without a CNN voiceover". "cuteness" has nothing to do with it.
>but it is not easier for a newbie to use.
I don't agree.
"Ease of Use" is a very tricky concept. In general, the easiest system to use is the one you already know how to use. And if you turn around and try to teach a newbie, your knowlege and prejudices and assumptions copy to their brain.
In general, my experience has been that unaided users still get a little farther on a Mac than on Windows.
Here's an example. Have your newbie user try to install an external Zip disk. Often I look at the installation instructions for new hw or sw, and I compare the bulk of text under "for windows" versus "for mac". It's a crude measure of ease of use, but often illuminating.
Before Plug & Play, Mac HW installation instructions were almost always shorter than for Windows. (Macs have pretty much always been out-of-the-box ready to go, by tradition. Windows went for many years with the IRQ conflict dust cloud in the user's face.)
If pcs had all implemented P&P in a fool proof way, they'd be up with Macs. But in fact P&P often fails. So now, the Windows hardware installation usually goes like this:
1) Try to get P&P to work.
2) sigh ok now do whatever you would have done before P&P
in other words, P&P has actually made the process harder, because it doesn't always work. In the Mac world, any product that didn't install instantly was, basically, broken, off the market, and out of business, cuz there was no alternative. They make that stuff work.
>Their focus groups need to spend less time asking people how they feel, and giving them real world tasks to complete.
I can't speak for apple, but they have a tradition of doing good usability testing. I know, however, that in the UI community, the consensus is:
- focus groups are lame, do usability testing
- real world tasks
- anybody in the room who knows how to do it keeps their mouth shut
- keep iterating the product until the humans get to the end
- if the user gets to the end, they'll feel good, but who cares. ship it.
>Are you honestly gonna tell me that only cost a company incurrs in running an OS is the administrator's paycheck?
That's an awful big part of it. People's time costs, what, $50/hour to $200/hour, for people who can grok being a sysadmin. (take your takehome pay, double it. this includes health insurance, electrical bills, rent on another 10x10 piece of floor, fuzzy cubicle walls, coffee machines, 10% of the cost of your manager, 1% of the cost of her manager, etc.) Therefore the two biggest costs are employees and consultants.
I figure I've personally spent thousands of dollars of my own time running Linux installers. This totally swamps whatever the "cost" of the distro was, whether $50 for a box or $0 for a CD that gets handed to me for free. It rivals and perhaps exceeds the cost of the hardware.
In fact, the $50 box is often "cheaper" cuz if the manuals that come inside can save me ONE hour of scratching my head and websurfing to find answers, it's paid for itself. And usually good manuals (or good knowledge already in my head) save lots and lots of hours.
SCSI disks are therefore "cheaper" than IDE cuz you plug it in and go. Plug N Play is not just convenient, but a money saver, when it works (yeah, yeah, don't get me started on that). And, bugs are very very expensive to work around. And, I'm typing this on a Mac, the cheapest machines on the market.
>When asked whether the company would ever consider replacing its
>Windows machines with Linux, Busch said absolutely not, noting
>the lack of "robust office packages" on that platform.
I often think that this excuse really is more like "we can't get naive users to use it without being crippled". Linux distros need to test their software on non-Unix people more. Humans. Typical office people who, if you ask them if they have a Mac or a Windows box, say, "Yeah, I think so".
>And Busch threw another wrench into any mass Linux migration by
>noting that the overall cost of Linux and Windows 2000 is almost
>identical after you factor in support and maintenance.
in other words, after you get done with the hassles of Linux, and the hassles of Win2k, the hassles of Linux are a little bit more. time=money, so the cost of that extra hassle is the same as the cost of Windows & its apps.
So much for free-as-in-beer.
This hassle is invisible to the Linux developers cuz they know how to fix or work around glitches when they arise. So it seems "easy to use" for them.
>Anyone know how many patents M$ has?
>Funny, don't hear much on that.
I saw a few just the other day. Patents on, for instance, using an HTML file to format a filemanager window display, and patents on making a file that's nothing but a URL shortcut (symlink). [Yes I know that konqueror and gnome do that too; not sure what's up with that.]
You can search yourself at
http://www.uspto.gov/patft/index.html
woo hoo!!! results, just the recent patents:
------------- snip -----------------
Searching 1996-2001...
Results of Search in 1996-2001 db for:
AN/Microsoft: 1748 patents.
Hits 1 through 50 out of 1748
PAT. NO. Title
1 6,311,323 Computer programming language statement building and information tool
2 6,311,228 Method and architecture for simplified communications with HID devices
3 6,311,216 Method, computer program product, and system for client-side deterministic routing and URL lookup into a distributed cache of URLS
4 6,311,209 Methods for performing client-hosted application sessions in distributed processing systems
5 6,311,142 Methods for designing pop-up cards, and cards produced thereby
6 6,311,058 System for delivering data content over a low bit rate transmission channel
7 6,308,274 Least privilege via restricted tokens
8 6,308,273 Method and system of security location discrimination
9 6,308,266 System and method for enabling different grades of cryptography strength in a product
10 6,308,222 Transcoding of audio data
- - - snip - - -
> 1.) dot net is still vaporware and has no installed base,...
no. They claim 160 million users. Izzat more than Linux? Maybe that's Passport. Trojan horsed by the fact that if you want to install XP, want a Hotmail account, want an MSN account, etc, you need your Passport.
I dunno, I don't think I'd enter my real name if I was in that situation, but that's john Q public.
>2.) People are satisfied with your so called "second class
>standards based Internet" and frankly most don't have the...
Most people will go where Internet Explorer will take them. Now that MSN isn't accepting Netscape, soon IE won't be able to surf to Slashdot. Etc.
Yes. Get alarmed.
> 3.) There is plenty time to build our own superior open
>architecture for any 'enhanced web' that...
Won't happen. Guess what, Sun is trying this. But do you hear about it on SlashDot? No. Do you hear calls to support Sun's authentication system? No. That's because Unix people are too divisive.
Linux people say BSD sucks. BSD people say Linux sux. Both say Java sux.
Solaris and Linux are so "cozy" with each other that when I multiboot between Solaris and Linux, I have to move files to an MS DOS partition to exchange data between them - for this reason alone, Microsoft will never go out of business: it's the only way to get Linux and Solaris to talk.
Linux CAN mount Solaris UFS partitions - read only. At least they recognize that Solaris UFS is different from BSD UFS. Both Solaris and BSD considers the other one to be broken. Hey, instead of building a "superior enhanced web" how about a "superior file system" that isn't fucking broken all the time? Start with the basics.
I haven't managed to get FreeBSD on this system - just getting Linux and Solaris to coexist is like walking on eggshells.
>Finally, Mac OS X takes a different tack. From what I understand, all created accounts are user level accounts in the Unix sense. To access the admin-level account, you have to explicitly enable root. I don't know enough about OS X to comment, but on the face of things, this seems like a simple security policy that many users can actually understand if explained to them.
Yes. During install, you are asked to decide and enter a root password, as usual. After that, you CANNOT log in as "root" until a magic thing is enabled somewhere. I forget where, but it was not intuitive to find it, and I didn't get it enabled until I'd had it for a week. I think in NetInfo somewhere. Read it in a magazine. I'd guess most users won't bother or won't have the wherewithall.
Anyway, you don't really need the root account for most stuff. Yes this means software installed as your username will be owned by your username and will be vulnerable to whatever attacks, running as your username. Such as an executable downloaded and run automatically.
Apple has done a remarkable job of all of this./var,/etc, and some other traditional Unix directories are invisible to the traditional Mac applications - besides being protected by being owned by root. They are actually symlinks into the directory/private, and MacOS doesn't do symlinks, only aliases. And bsd doesn't do aliases, only simlinks. Crazy./usr and some others are real directories, not sure why mac apps can't see them, maybe it's hardwired in. I'm sure a Unix-side app can get at them in the Unix way, except for the permissions as usual.
It's ready for prime time and much better for the consumer market than any Unix ever has been. And the command line is always there if you want it.
it's the american public, they don't understand. Here's some arguments that can help them understand, in your coctail conversations:
Would you give the police department a key to your home, so that they can protect you from crime? No. Think of why not - several reasons, like an out of control cop could terrorize you, etc. Meanwhile, anybody who's a criminal will NOT give the police a key to their home, or will give the wrong key, or will put on an additional padlock.
Why not strip search, for drugs, all people crossing the Mexican/US border at Tiajuana? Because it's a pain for those being searched. And, the real people smuggling drugs will drive a truck along a back road into arizona or new mexico. The stripsearch will be totally ineffective.
Why not make backdoors for encryption? Because that jepordizes all law abiding encryption users. The crackers will figure it out before the law is even passed. Meanwhile, no criminal or terrorist in their right mind will use that encryption, they'll use their own. Even if they have to break the law... they're already breaking the law anyway. Computers don't change anything, especially not for technophobes living in tents in afghanistan.
Slashdot Mirror
We should compile a hit list of spammers and their addresses, the top 42 of them, on a hit list website. Tell them to stop sending spam or we'll shoot them with a high powered water gun through their kitchen window. Then, as we shoot them, and CNN (conservative news network) reports it nationally, cross each one off the list.
They'll get the message.
The reason why setup.exe (= installers in general) work is that they are foolproof and brainless and work the first time.
[cough cough. Well, not always. yes sometimes they don't work, sometimes they crash, sometimes you have to coax them along, sometimes they break stuff that used to work, sometimes you gotta reformat and reinstall the whole os...]
./configure; make; make install
is not the same:
1) it takes more typing
2) the chances of brainless success are lower, and you, the human installer, are assumed to have the skill to fix what's broken and coax it along.
This is why and when software is distributed as source: it's less likely to work as-is, therefore, for a programmer who can hack, the means are there to fight toward success.
A canned binary does not have this flexibility. A canned binary for x86 will not work on PowerPC and vice versa. A buildable source distro will.
I got a 4 CD set of software, bsd ports, for Mac OS X. Sources. Scanning through the whole thing one evening, I loaded up everything I thought would be useful. I mostly wanted bash. Most of it just didn't build as it was, including bash, and I didn't take the time to think and fight - if it didn't build, I went on. At the end of the evening, I'd gained one piddly little program that I don't even remember. Nothing else worked - not brainlessly at least. Some day I'll attack the bash sources again.
BY comparison, most Mac production software from apple or commercial publishers, installs a binary and works beautifully the first time. We're talking, Click. We're talking, the manual sez it won't work on a non-Apple USB, but it does anyway. But of course only for selected OS versions on PPC, and the installer took more work on the part of the publisher. There's no free lunch.
PS: I'm not worried about computers taking oover the world. They need us humans to fix the bugs and run the installers and build scripts.
> This biggest problem is normally the Mac to be honest, it doesn't read others, and others can't read it, if you get what I mean.
Not true, at least not completely true. I've gotten my (dorky old) firewire drive to move between a Mac (G3 gray with addon orangmicro firewire, running MacOS 9) and a PC (vaio laptop, win 98). This was in mid 2000 so stuff has upgraded obviously.
If partitioned and formatted on Win 98, the disk has a DOS partition table (primary and logical partitions) and I put both FAT 16 and FAT 32 partitions on it. Came up just fine in MacOS, just like floppys and zips that are dos formatted. Although MacOS can't format one of the partitions to be HFS without...
If partitioned under MacOS, with Mac partition tables and HFS and/or HFS+ partitions, Windows treated it like it was from mars. unformatted.
Haven't tried under Mac OS X yet, could be busted. I got more gigs than I can use right now these days. Yuck, just stepped on another disk drive!
actually, 1394 started out with four speeds:
S100, S200 and S400. By now most everybody is at S400 = 400mBITS/sec as you all say. My older vaio has some ports at S200 and some at S400.
Apple and others are pounding away trying to get S800 and S1600 going; I'm hoping they're shipping by the time USB 2.0 is shipping. And of course the SCSI people aren't asleep, developing YetMoreStupendouslyWide variations. I expect FireWire and SCSI to play leapfrog, and USB to stay behind.
But yes, as everyone says, the bus and disk and kernel/DMA (or whatever) are also bottlenecks in the way. YMMV. Solaris x86 ferinstance ships with DMA turned OFF for some reason but they tell how to turn it on (i dunno if it does 1394). Bleah. Trust benchmarks and your own personal experience.
The 6-pin 1394 has the power pins in the USB-sized plug - think disks. The tiny 4-pin 1394 doesn't - think vid cameras. Cables of 6-6, 6-4 and 4-4 kinds are available often at the usual inflated prices.
>Want to kill Microsoft? Sap it's growth, which is in server OS's and embedded systems (XBox, Pocket PC, etc.)
One of the things I've learned about business is, in general, it's impossible to "drive" anybody out of business from the outside. Companies that get derailed do so by their own mismanagement (often under stress, sometimes from a competitor). Being well-run means responding to a changing world and finding new revenue sources if old ones dry up.
In Microsoft's case, they have revenue from many, many different sources. Smothering one of their major revenue sources is possible. Smothering all of them at the same time, and squelching any new initiatives they have, is highly unlikely. Meanwhile, they have cash in the bank, too.
If you want a goal, just stick to your own business. Make sure your own revenue keeps coming. That's something you do have control over.
>Go shoot your precious linux server with a .44, and see if it's still up. I'd guess the answer is no. Doing the same with a properly configured SunFire 15k would result in a high replacement cost, but an up and running system nonetheless.
That would make a cool TV ad!!! "Hi, My name is Scott McNealy, president of Sun Microsystems. We're here today in Taloqan, Afghanistan, to demonstrate the bullet-proof reliability of Sun servers. We brought along a SunFire 15000, serving web pages at this very moment, as George, our test engineer, demonstrates." [george clicks on a hotlink on his pc, another page comes up.] "Speaking of bullet-proof, our marksman here, Hafez Mohammed, from the Northern Alliance, is armed with a genuine Soviet-made Kalishnikov rifle..."
right. To clarify, I took that quote directly from the article. That's what the Intel guy said. An anectdote, not a data point. Actually measuring such a thing is really hard and subjective and depends on so many particulars, maybe it's unmeasurable.
For instance, maybe the windows sysadmins cost less per hour. So the beancounters have a number. But how much "management" do you get per sysadmin per hour? How much "management" does each system need? What's the extra cost of cleaning up from Nimda? Will Linux ever have a worm as bad as Nimda? (unlikely but that's what MS thought too.) What's the extra cost of reinstalling NT six times to get it "right"? Or Linux for that matter? Well if everybody had been using XYZ motherboards instead of PQR motherboards, you wouldn't have run into that hassle. It's the STU raid system! With the ACLs. If you hadn't messed with the ACLs, or had gotten a different RAID system...
I don't know how to measure this, but I know that measuring it is highly vulnerable to bias. EG it's easy for MS to come up with data proving whatever.
> An extremely new account number
/. for a year or so. I just kept on forgetting my account name, and never bothered to read how the whole point system works and the popup menus and which reply button replies to what... as usual, in this culture, if you don't RTFM, and play around with it a bit to figure out what TFM doesn't tell you, you're lost and intimidated.
/. to talk to some non-technical people, you'll find more people who agree. This is what happens to the bearer of bad news.
I've been reading
> a name like "OsamaBinLogin"
Now THAT's a user name I can remember!!! Cute, huh? thought it up myself, and very proud of it thanyouverymuch. I just hope the FBI doesn't read something into it and haul my ass downtown. (Dear FBI: really, I'm just kidding around. I'm an agnostic cleanshaven honkie who's never used a firearm.)
>and the typical troll "normal people can't understand Linux"
It's got to be said. And I think more than a few people on this board agree with it. And if you get away from
> and not one moderatoor has modded this down? I'm disgusted.
are you trying to be INTIMIDATING?!?!?! Trying to make people ASHAMED for disagreeing with you?
ok, now i'm being a troll. mod me down.
>Macintosh may be more soothing -- appealing shapes and colors, friendlier text and error messages (bombs and frowns anyone?)...
I don't think "soothing" has anything to do with it. Amateur UI designers keep thinking they are making software "easy" when they stick nice but incomprehensible icons on the same unexplained concepts. That's why Microsoft invented tooltips - so you can figure out what the icon does.
Besides, I think MacOS and Windows keep leapfrogging each other on the "graphically cool" aspect. If it's a frozen photo in a magazine, that sells.
The frowning Mac face originally came about because the original Macintosh hardware was language-independent. you boot up with a French system, everything is in French. If there's an error booting up, you ain't got no text to spit out an error message. So you either put up an icon that makes it obvious what's happening, or you say in english, "this computer made by ugly americans who can't comprehend a foreigner without a CNN voiceover". "cuteness" has nothing to do with it.
>but it is not easier for a newbie to use.
I don't agree.
"Ease of Use" is a very tricky concept. In general, the easiest system to use is the one you already know how to use. And if you turn around and try to teach a newbie, your knowlege and prejudices and assumptions copy to their brain.
In general, my experience has been that unaided users still get a little farther on a Mac than on Windows.
Here's an example. Have your newbie user try to install an external Zip disk. Often I look at the installation instructions for new hw or sw, and I compare the bulk of text under "for windows" versus "for mac". It's a crude measure of ease of use, but often illuminating.
Before Plug & Play, Mac HW installation instructions were almost always shorter than for Windows. (Macs have pretty much always been out-of-the-box ready to go, by tradition. Windows went for many years with the IRQ conflict dust cloud in the user's face.)
If pcs had all implemented P&P in a fool proof way, they'd be up with Macs. But in fact P&P often fails. So now, the Windows hardware installation usually goes like this:
1) Try to get P&P to work.
2) sigh ok now do whatever you would have done before P&P
in other words, P&P has actually made the process harder, because it doesn't always work. In the Mac world, any product that didn't install instantly was, basically, broken, off the market, and out of business, cuz there was no alternative. They make that stuff work.
>Their focus groups need to spend less time asking people how they feel, and giving them real world tasks to complete.
I can't speak for apple, but they have a tradition of doing good usability testing. I know, however, that in the UI community, the consensus is:
- focus groups are lame, do usability testing
- real world tasks
- anybody in the room who knows how to do it keeps their mouth shut
- keep iterating the product until the humans get to the end
- if the user gets to the end, they'll feel good, but who cares. ship it.
>Are you honestly gonna tell me that only cost a company incurrs in running an OS is the administrator's paycheck?
That's an awful big part of it. People's time costs, what, $50/hour to $200/hour, for people who can grok being a sysadmin. (take your takehome pay, double it. this includes health insurance, electrical bills, rent on another 10x10 piece of floor, fuzzy cubicle walls, coffee machines, 10% of the cost of your manager, 1% of the cost of her manager, etc.) Therefore the two biggest costs are employees and consultants.
I figure I've personally spent thousands of dollars of my own time running Linux installers. This totally swamps whatever the "cost" of the distro was, whether $50 for a box or $0 for a CD that gets handed to me for free. It rivals and perhaps exceeds the cost of the hardware.
In fact, the $50 box is often "cheaper" cuz if the manuals that come inside can save me ONE hour of scratching my head and websurfing to find answers, it's paid for itself. And usually good manuals (or good knowledge already in my head) save lots and lots of hours.
SCSI disks are therefore "cheaper" than IDE cuz you plug it in and go. Plug N Play is not just convenient, but a money saver, when it works (yeah, yeah, don't get me started on that). And, bugs are very very expensive to work around. And, I'm typing this on a Mac, the cheapest machines on the market.
>When asked whether the company would ever consider replacing its
>Windows machines with Linux, Busch said absolutely not, noting
>the lack of "robust office packages" on that platform.
I often think that this excuse really is more like "we can't get naive users to use it without being crippled". Linux distros need to test their software on non-Unix people more. Humans. Typical office people who, if you ask them if they have a Mac or a Windows box, say, "Yeah, I think so".
>And Busch threw another wrench into any mass Linux migration by
>noting that the overall cost of Linux and Windows 2000 is almost
>identical after you factor in support and maintenance.
in other words, after you get done with the hassles of Linux, and the hassles of Win2k, the hassles of Linux are a little bit more. time=money, so the cost of that extra hassle is the same as the cost of Windows & its apps.
So much for free-as-in-beer.
This hassle is invisible to the Linux developers cuz they know how to fix or work around glitches when they arise. So it seems "easy to use" for them.
Try it on grandma. then report back.
>Anyone know how many patents M$ has?
>Funny, don't hear much on that.
I saw a few just the other day. Patents on, for instance, using an HTML file to format a filemanager window display, and patents on making a file that's nothing but a URL shortcut (symlink). [Yes I know that konqueror and gnome do that too; not sure what's up with that.]
You can search yourself at
http://www.uspto.gov/patft/index.html
woo hoo!!! results, just the recent patents:
------------- snip -----------------
Searching 1996-2001...
Results of Search in 1996-2001 db for:
AN/Microsoft: 1748 patents.
Hits 1 through 50 out of 1748
PAT. NO. Title
1 6,311,323 Computer programming language statement building and information tool
2 6,311,228 Method and architecture for simplified communications with HID devices
3 6,311,216 Method, computer program product, and system for client-side deterministic routing and URL lookup into a distributed cache of URLS
4 6,311,209 Methods for performing client-hosted application sessions in distributed processing systems
5 6,311,142 Methods for designing pop-up cards, and cards produced thereby
6 6,311,058 System for delivering data content over a low bit rate transmission channel
7 6,308,274 Least privilege via restricted tokens
8 6,308,273 Method and system of security location discrimination
9 6,308,266 System and method for enabling different grades of cryptography strength in a product
10 6,308,222 Transcoding of audio data
- - - snip - - -
> 1.) dot net is still vaporware and has no installed base,...
no. They claim 160 million users. Izzat more than Linux? Maybe that's Passport. Trojan horsed by the fact that if you want to install XP, want a Hotmail account, want an MSN account, etc, you need your Passport.
I dunno, I don't think I'd enter my real name if I was in that situation, but that's john Q public.
>2.) People are satisfied with your so called "second class
>standards based Internet" and frankly most don't have the...
Most people will go where Internet Explorer will take them. Now that MSN isn't accepting Netscape, soon IE won't be able to surf to Slashdot. Etc.
Yes. Get alarmed.
> 3.) There is plenty time to build our own superior open
>architecture for any 'enhanced web' that...
Won't happen. Guess what, Sun is trying this. But do you hear about it on SlashDot? No. Do you hear calls to support Sun's authentication system? No. That's because Unix people are too divisive.
Linux people say BSD sucks. BSD people say Linux sux. Both say Java sux.
Solaris and Linux are so "cozy" with each other that when I multiboot between Solaris and Linux, I have to move files to an MS DOS partition to exchange data between them - for this reason alone, Microsoft will never go out of business: it's the only way to get Linux and Solaris to talk.
Linux CAN mount Solaris UFS partitions - read only. At least they recognize that Solaris UFS is different from BSD UFS. Both Solaris and BSD considers the other one to be broken. Hey, instead of building a "superior enhanced web" how about a "superior file system" that isn't fucking broken all the time? Start with the basics.
I haven't managed to get FreeBSD on this system - just getting Linux and Solaris to coexist is like walking on eggshells.
And bill gates laughs all the way to the bank.
>Finally, Mac OS X takes a different tack. From what I understand, all created accounts are user level accounts in the Unix sense. To access the admin-level account, you have to explicitly enable root. I don't know enough about OS X to comment, but on the face of things, this seems like a simple security policy that many users can actually understand if explained to them.
/var, /etc, and some other traditional Unix directories are invisible to the traditional Mac applications - besides being protected by being owned by root. They are actually symlinks into the directory /private, and MacOS doesn't do symlinks, only aliases. And bsd doesn't do aliases, only simlinks. Crazy. /usr and some others are real directories, not sure why mac apps can't see them, maybe it's hardwired in. I'm sure a Unix-side app can get at them in the Unix way, except for the permissions as usual.
Yes. During install, you are asked to decide and enter a root password, as usual. After that, you CANNOT log in as "root" until a magic thing is enabled somewhere. I forget where, but it was not intuitive to find it, and I didn't get it enabled until I'd had it for a week. I think in NetInfo somewhere. Read it in a magazine. I'd guess most users won't bother or won't have the wherewithall.
Anyway, you don't really need the root account for most stuff. Yes this means software installed as your username will be owned by your username and will be vulnerable to whatever attacks, running as your username. Such as an executable downloaded and run automatically.
Apple has done a remarkable job of all of this.
It's ready for prime time and much better for the consumer market than any Unix ever has been. And the command line is always there if you want it.
it's the american public, they don't understand. Here's some arguments that can help them understand, in your coctail conversations:
... they're already breaking the law anyway. Computers don't change anything, especially not for technophobes living in tents in afghanistan.
Would you give the police department a key to your home, so that they can protect you from crime? No. Think of why not - several reasons, like an out of control cop could terrorize you, etc. Meanwhile, anybody who's a criminal will NOT give the police a key to their home, or will give the wrong key, or will put on an additional padlock.
Why not strip search, for drugs, all people crossing the Mexican/US border at Tiajuana? Because it's a pain for those being searched. And, the real people smuggling drugs will drive a truck along a back road into arizona or new mexico. The stripsearch will be totally ineffective.
Why not make backdoors for encryption? Because that jepordizes all law abiding encryption users. The crackers will figure it out before the law is even passed. Meanwhile, no criminal or terrorist in their right mind will use that encryption, they'll use their own. Even if they have to break the law