The question I want answered - so how is Transmeta doing these days? When the performance thing looked like it wasn't going to happen, they banked on the low power consumption -- it looks like they got hammered in that niche as well.
As a sidenote - everyone who thinks that Google has taken the world by storm because of their extraordinary search technology is naive. Google's strength was in their revenue/business model, not their search technology. What Google innovated is the idea of a clean, simple interface and simple, low bandwidth ads paying for a massive cluster of systems and databases. They did this at the same time that other organizations were trying to make companies pay to get their content spidered, and who put all of their emphasis on advertisement technology.
Here's the facts - Lookout is a trivial shell extension on Outlook (a rather half-assed one at that, missing countless trivial features while at the same time dramatically reducing the stability of Outlook. Indeed this was likely why Microsoft sought to supress the product after "acquiring" Lookout), and the "search technology" in it is actually a open source.NET port of Jakarta Lucene (http://sourceforge.net/projects/lucenedotnet/). Lookout is nothing more than a file/email iterator with a simple GUI.
It is absolutely astounding that people keep reiterating this myth that Microsoft, the people who've had full-text search as a feature of Office (findfast), and then a part of the OS (indexing service), not to mention in specific products (SQL Server full-text indexing), needed to "acquire" some trivial application, actually built atop an open source project, to show them the way in search. The idea that Lookout, which they "acquired" just a few months ago, is what gave them the start on desktop search is just awe inspiringly idiotic.
About the "acquired" - The acquisition of Lookout was almost certainly an HR headcount acquisition. They wanted to hire one of the people involved, but they wanted them to wrap up their current project, so this so-called acquisition occurred.
This is all so inane - Microsoft, and virtually every other company, have been talking about the all-encompassing database filesystem for over a decade. How unsurprizing that a Mac fan believes that Apple invented it.
If you are hostile towards your customers (like the ??AA) your revenues will shrink, and you will find yourself in a viscious cycle fighting with your customers and losing money in the process!
Several years back, when the Diamond was a big name in the video card market, and 3dfx was the king of the hill, I frequented comp.hardware.ibm.pc.video -- many of the forum regulars made no bones about the fact that they were devil customers: They purposefully would "buy" a video card, hold it for just under the return period, and then return it. They'd get a full refund of their original purchase price which they'd use to buy whatever was new (again starting the return clock anew). These customers are hugely costly for retailers -- it would be better not to have them as customers (in fact you wish them on your competitors). This sort of person will rationalize their behaviour (much like the cable modem user who rationalizes saturating their connection 24 hours a day) under the guise of "if they let me, let them suck it!", but the end result is naturally restrictive policies that hurt everyone because of the abuses of a few. Simliarly it isn't cost effective to have customers who'll bogart your salespeople's time for hours while they ruminate over a trivial decision -- one which they'll likely recant on, reappearing in your returns line. These people do exist.
I was going to ask the same thing - I'm relatively nearby in the Toronto area, and the coolest night has been around 4C. I've yet to awake to frost on cars, much less ice on trees.
Actually, I thought the whole point of an unlikely coincidence is that it doesn't happen all of the time.
As a singular event, sure. When you're talking about thousands, millions, billions, or trillions of events happening daily, the likelihood of something "unlikely" happening becomes, humorously, likely (in fact it'd be a statistical oddity if they didn't happen).
Millions of noteworthy events have happened during the course of the last dozen elections or so, and someone found one particular set of events (home games for one specific team in one specific sport in one specific timeframe) that correlated. It's neat, but it's not unfathomable that such a coincidental correlation could be found. If Bush wins and this run is squashed, I'm sure someone could do some digging and find something equally as amazing ("The number of lefty batter home runs in American League East baseball games for the 7 days prior to the inaugeration date accurately predict the winning candidate over the past X years...").
Sigh...drawing causation or correlation from 5 data points is just sad. Kerry supporters are grasping at straws.
I think the sad thing is how seriously the Bush camp is taking this. While most everyone else is taking this as a "Ha ha! What a neat coincidence", repeating it tongue in cheek, the rabid frothy spittled Bush supporters have actually come out defensively regarding this (see your post, and many before it). Amazing.
There's been 17 elections since this 'pattern' supposedly emereged. At most, that means 9 times an incumbent was running. At most, that means 5 times the incumber lost when the Redskins did....sure helps to demystify curses with logic, but most people don't bother.
Huh? Firstly it's not a curse, it's a humorous coincidence. Ha ha. Secondly, RTFA - They're talking about the incumbent PARTY. Thirdly, your convoluted attempt at bringing "logic" to the table looks like it misfired - your logic is nonsensical.
It _is_ a pretty unlikely coincience....but unlikely coincidences happen all of the time.
I'd say more often the exact opposite is true. People choose Linux because of the general perception that it is the more stable, more secure choice. After a rooting the security admin can proclaim "All the press and the community said it was the greatest thing since sliced bread...I don't know what went wrong!"
Given all the bad publicity Microsoft has (deservedly) received, it is a huge risk for architects and security admins to choose Windows -- when things go wrong everyone can immediately claim "duh! You picked Windows you idiot!" (see the Navy fiasco with the dead-at-sea warship, or the recent LAX fiasco. Both were application layer faults but that didn't stop the routine presumption that the core fault was the idiots that chose to base them on Windows).
In other words reapply the old "no one got fired for choosing IBM" onto Microsoft, which is who I'm presuming you're implying, is a false comparison. People choose Windows at great peril, and when their line-level admin doesn't bother with patches or basic security practices, instead it's the guy who chose Windows that gets the blame.
The fact that a Canadian firm has a US parent doesn't excuse that company from giving away data entrusted to it by any Canadian government to any other firm or government not authorized to receive it by the same Canadian government
It's not just a US parent that's a risk, but even Canadian companies (such as the major banks) that own subsidiaries in the US - Royal Bank owns Liberty Insurance and several bank chains, TD owns an investment arm and some bank chains, BMO owns some bank chains. The US government can technically use the subsidiaries to weasle through to obtain information on the whole enterprise.
No you couldn't, that was my point. A non-standard addition to NT4 gave you that option, and with a bit of effort that add on could be grafted onto NT3.51 server. In NT5 it was included as standard.
Non-standard? Su came with the resouce kits going back to NT 3.1, allowing you to run a process under any account context. This is hardly some great chasm, and is no different than having a stock admin RPM installed.
And of course the API and security infrastructure was built since day one as a multi-user system -- services, for instance, take advantage of that functionality.
I find it absolutely astounding that you would consider NT not "multi-user" until it had a "RunAs" popup menu item in the context menu. Astounding.
How can this be to do with proprietry software and open source...
It has nothing whatsoever to do with proprietary Vs open source, and the addition of that incendiary flamebait in the submission was completely unnecessary trolling. Amazing how the majority of the comments thus far have been knee-jerk reactions with the chorus of the converted fervently preaching to their pewmates.
This isn't that hard to do. sender-id, spf, etc, does nothing.
These most certainly aren't total solutions, but they are gradual steps in the right direction (and really SMTP has always been the most absurdly abusable protocol. It's time to harden it a bit)....virus writers will just go to the extra effort of sending spam out the zombie PC through the owners' ISP mail server, and to your inbox...
When a company like AOL or GMail commits to schemes like SenderID, SPF, or DomainKeys, they are effectively declaring their total responsibility over that mail source -- no longer is there confusion or deniability over whether a piece of mail was just sent direct or actually went through the Gmail system, for instance. As such, you can be sure that they will ensure that minimal amounts of spam are sent from their system -- so when Joe Blow downloads MonkeyPunchTM and it starts spamming out of his gmail account, they'll just shut the account down (detecting spam being sent from a source is pretty easy). I doubt virus writers will find much value in sending a couple of emails from each owned PC before the accounts are locked out. On the flip side the big providers no longer would have to deal with billions of spam returns for messages that were never sent from their system in the first place. Win win win.
We already know most semi-legitimate spammers are publishing SPF records on their throwaway domains which takes care of the other 10% of spam...
Obviously we're just getting started. Undoubtedly these systems, particularly DomainKeys, will develop into whitelist trust chains eventually, so it'll be rather easy to cut abusers out. It's also incredibly easy to build a "blacklist" of spamming domains, and again it's obvious that spammer will find little return in setting up domains for the sole purpose of spamming when it just gets cut out of the global loop in no time (not to mention that they're not stepping on legitimate email accounts in their from).
Air-conditioning, power-steering and even ABS still aren't standard despite costing next to nothing at build time
While it's incredibly dubious that they cost "next to nothing", they do cost something once the car drives off the lot - air conditioning, for instance, is probably the number 1 support issue in most vehicles, and under warranty those support issues add up to thousands of dollars for the manufacturer.
Manufacturers need to cripple cheaper cars to somehow...
So it's all a big collusion? Why doesn't a company like Kia stick all of this great stuff in their $9,000 car and competely own the market, instead of marketing an empty shell with minimal amenities. Right...because those things cost money.
Of course the margins increase on more expensive cars (and the margins on luxury SUVs are massive), but the idea that it's all just a big conspiracy is ludicrous - the car companies are locked in a very tight competition and want to kill each other, and it's not just coincidence that a Honda Civic is similarly equipped to a Hyundai Elantra and a Pontiac Sunfire.
Probably because the technology in supporting electronics and chipsets, not to mention programming knowledge, has been massively subsidized by the mainstream computing market. Going off and reinventing the wheel with their own transport system just isn't rational in such a case.
The best way to help the least fortunate is often by increasing the size of the economic pie - basically by making the middle class richer by encouraging economic commerce and innovation (cue naïve sarcastic remarks about Reagan's trickle down economics. Note, however, that I'm talking about the middle class rather than the rich. Furthermore, I am not speaking about tax cuts, but rather programs or initiatives that expand the economy, such as encouraging technological growth). Perhaps the premise behind free wireless is that it will lead to a slew of new programs and services in the San Francisco area that will lead to a lot of taxable commerce - tax revenue that can then be used to provide mental health support for some of the homeless. The most effective route to a goal isn't always the most direct.
In my neck of the woods there are few HDTV providers - Cogeco (a cable system), and Bell Canada ExpressVu (a satellite system). Both of them use entirely different hardware, and thus my choice of PVR box is dictated by my choice of provider (in my case it would have to be a Motorola DCT6208 or a Bell 6120..whatever manufacturer that is), and if I ponied up the $800CDN or so with one and became disatisfied with the service, I'd have to toss the box in the garbage.
Which is why I chose to rent the box for $20 a month - when the two-tuner version comes out (within the year) I'll do a swap. The point of all of this? I guess that ultimately I am choosing provider and the PVR comes along for the ride.
The NT series is NOW multiuser, since we now have full file permissions and can run different processes safely as different users - but it took many years to get to that point.
What are you talking about? NT has had full system wide ACLs (actually much more comprehensive and pervasive than Linux, or most other Unix variants) and process security context independence since the origin - don't think your Windows 95 knowledge has any relevance at all to the NT line. In other words your statement is just a load of ignorance.
Now the GUI has generally always been "single user" because the GUI doesn't (or rather didn't) have a "remote" connectivity option - with interactive logins it really is only possible for one person to be using the computer, so the GUI was designed around that. Of course you've always been able to launch processes under any security context.
So what you're saying is to get reliability out of a Windows server you need to be running a cluster.
I don't work for Microsoft, and I don't care if you buy or like Microsoft products, so I'm not speaking on behalf of Windows.
Having said that, there are generally two types of environments
-Environments where planned downtime can be accommodated (for instance an insurance company where all the employees go home at night and the system is taken offline for batch processing).
-Environments where there is no leeway for downtime.
In the latter situation clusters are quite simply the norm - either physical clusters or virtual clusters. Why? No matter how confident you are in your OS, hardware fails -- even when you have RAID 5 arrays and dual backplanes and redundant power supplies, one day the RAID card dies or the UPS serving that server fries and ironically takes out the server. Given this suddenly it becomes somewhat irrelevant if you have to rotate patch deployments through your servers.
You perform semantic gymnastics to avoid understanding the obvious.
If "the obvious" is whatever serves some anti-Microsoft crusade, then I suppose you're right. However it is painfully obvious to anyone without the ideological blinders that a system catrostrophically going down and a system being selectively rebooted at an opportune time are two very, very, very different things.
"But they're vulnerable to haxxers if they don't install the patch immediate," you might say. Sure, except that most critical boxes have nothing but the essential services running, and are often firewalled. The reality is that most patches are entirely unnecessary on these boxes, and most admins accept the scheduled downtime simply because it's easier than determining if the patch has any applicability to their boxes.
That's sort of the point. You have to reboot a Windows server more often. If rebooting once a month or so is acceptable (see Murphy's Law for schedule), then that's fine.
But that's not the point - there is an implication that it is instability, i.e. uncontrolled downtime, when in reality it is controlled downtime (well accommodating the fact that sometimes security patches need to be installed relatively quickly). A controlled reboot of your server at 3 in the morning when all of your employees are at home is absolutely nothing like having your server crash at 10:00am. It is rhetorical hyperbole comparing them.
Of course for web applications this should be an entirely moot point - web apps with any requirement for reliability should be running in a cluster or network load balance arrangement (fully supported by.NET for shared session), both of which Windows 2003 fully supports out of the box. In that case, with multiple balanced servers, you can freely patch any of them (or deal with failed hardware) with minimal or no customer impact -- maybe slightly slower responses with a smaller cluster.
Undoubtedly there are a lot of incompetent sales folks out there (remember these people are making like $7-$10 an hour effectively -- you're not getting the pioneers in the field or the cream of the crop). However I worked in computer sales for a short while in the formative years, so I have perhaps a sore spot when it comes to salesperson complaints - occasionally I would get some weenie who has spent months meticulously reading about and absorbing every piece of information on one highly specialized area of the computing field, and they'd then march in and start quizzing me about my knowledge of the same (these were the types who never actually bought anything, but instead demanded my time on numerous days while they tried to solidify their decision to buy a $110 SCSI controller). Ultimately of course I was at a disadvantage (just as I was when the next weenie came in with highly specialized knowledge of some other area of the field) but this allowed the customer a temporary feeling of intellectual superiority. It was very annoying, and ultimately was a waste of time.
I should buy the Monster brand optical audio cable, because it has better shielding...He INSISTED that he had plugged the two in and 'head the difference'.
Remember that Monster cable is primarily sold through high end audio retailers, and those highly specialized salespeople will feed you the same BS.
Err...why would you ask them what it stands for? Ultimately all of the engineering buzzwords come down to a couple of simple results - price and performance. If MOSFET is dead gerbils minced and blended into a chip insulator, or an advanced technique of containing the power of fusion it doesn't matter - does it make the stereo louder? Does it make the stereo sound better? Does it make it cheaper?...oh I'm forgetting the guy who told me I'd need "Digital" wires for my DVD player...
Could he be referring to SPDIF cables, which indeed are different than normal audio cables?
Slashdot Mirror
The question I want answered - so how is Transmeta doing these days? When the performance thing looked like it wasn't going to happen, they banked on the low power consumption -- it looks like they got hammered in that niche as well.
As a sidenote - everyone who thinks that Google has taken the world by storm because of their extraordinary search technology is naive. Google's strength was in their revenue/business model, not their search technology. What Google innovated is the idea of a clean, simple interface and simple, low bandwidth ads paying for a massive cluster of systems and databases. They did this at the same time that other organizations were trying to make companies pay to get their content spidered, and who put all of their emphasis on advertisement technology.
By buying a company. How like them.
.NET port of Jakarta Lucene (http://sourceforge.net/projects/lucenedotnet/). Lookout is nothing more than a file/email iterator with a simple GUI.
Wow, the myth of Lookout is extraordinary.
Here's the facts - Lookout is a trivial shell extension on Outlook (a rather half-assed one at that, missing countless trivial features while at the same time dramatically reducing the stability of Outlook. Indeed this was likely why Microsoft sought to supress the product after "acquiring" Lookout), and the "search technology" in it is actually a open source
It is absolutely astounding that people keep reiterating this myth that Microsoft, the people who've had full-text search as a feature of Office (findfast), and then a part of the OS (indexing service), not to mention in specific products (SQL Server full-text indexing), needed to "acquire" some trivial application, actually built atop an open source project, to show them the way in search. The idea that Lookout, which they "acquired" just a few months ago, is what gave them the start on desktop search is just awe inspiringly idiotic.
About the "acquired" - The acquisition of Lookout was almost certainly an HR headcount acquisition. They wanted to hire one of the people involved, but they wanted them to wrap up their current project, so this so-called acquisition occurred.
This is all so inane - Microsoft, and virtually every other company, have been talking about the all-encompassing database filesystem for over a decade. How unsurprizing that a Mac fan believes that Apple invented it.
If you are hostile towards your customers (like the ??AA) your revenues will shrink, and you will find yourself in a viscious cycle fighting with your customers and losing money in the process!
Several years back, when the Diamond was a big name in the video card market, and 3dfx was the king of the hill, I frequented comp.hardware.ibm.pc.video -- many of the forum regulars made no bones about the fact that they were devil customers: They purposefully would "buy" a video card, hold it for just under the return period, and then return it. They'd get a full refund of their original purchase price which they'd use to buy whatever was new (again starting the return clock anew). These customers are hugely costly for retailers -- it would be better not to have them as customers (in fact you wish them on your competitors). This sort of person will rationalize their behaviour (much like the cable modem user who rationalizes saturating their connection 24 hours a day) under the guise of "if they let me, let them suck it!", but the end result is naturally restrictive policies that hurt everyone because of the abuses of a few. Simliarly it isn't cost effective to have customers who'll bogart your salespeople's time for hours while they ruminate over a trivial decision -- one which they'll likely recant on, reappearing in your returns line. These people do exist.
http://archives.cnn.com/2002/WORLD/asiapcf/central /04/18/afghanistan.canada/
It's fast! How fast? Super fast!
I was going to ask the same thing - I'm relatively nearby in the Toronto area, and the coolest night has been around 4C. I've yet to awake to frost on cars, much less ice on trees.
Kudos on some very nice photos.
Actually, I thought the whole point of an unlikely coincidence is that it doesn't happen all of the time.
As a singular event, sure. When you're talking about thousands, millions, billions, or trillions of events happening daily, the likelihood of something "unlikely" happening becomes, humorously, likely (in fact it'd be a statistical oddity if they didn't happen).
Millions of noteworthy events have happened during the course of the last dozen elections or so, and someone found one particular set of events (home games for one specific team in one specific sport in one specific timeframe) that correlated. It's neat, but it's not unfathomable that such a coincidental correlation could be found. If Bush wins and this run is squashed, I'm sure someone could do some digging and find something equally as amazing ("The number of lefty batter home runs in American League East baseball games for the 7 days prior to the inaugeration date accurately predict the winning candidate over the past X years...").
Sigh...drawing causation or correlation from 5 data points is just sad. Kerry supporters are grasping at straws.
...sure helps to demystify curses with logic, but most people don't bother.
I think the sad thing is how seriously the Bush camp is taking this. While most everyone else is taking this as a "Ha ha! What a neat coincidence", repeating it tongue in cheek, the rabid frothy spittled Bush supporters have actually come out defensively regarding this (see your post, and many before it). Amazing.
There's been 17 elections since this 'pattern' supposedly emereged. At most, that means 9 times an incumbent was running. At most, that means 5 times the incumber lost when the Redskins did.
Huh? Firstly it's not a curse, it's a humorous coincidence. Ha ha. Secondly, RTFA - They're talking about the incumbent PARTY. Thirdly, your convoluted attempt at bringing "logic" to the table looks like it misfired - your logic is nonsensical.
It _is_ a pretty unlikely coincience....but unlikely coincidences happen all of the time.
Interesting.
I'd say more often the exact opposite is true. People choose Linux because of the general perception that it is the more stable, more secure choice. After a rooting the security admin can proclaim "All the press and the community said it was the greatest thing since sliced bread...I don't know what went wrong!"
Given all the bad publicity Microsoft has (deservedly) received, it is a huge risk for architects and security admins to choose Windows -- when things go wrong everyone can immediately claim "duh! You picked Windows you idiot!" (see the Navy fiasco with the dead-at-sea warship, or the recent LAX fiasco. Both were application layer faults but that didn't stop the routine presumption that the core fault was the idiots that chose to base them on Windows).
In other words reapply the old "no one got fired for choosing IBM" onto Microsoft, which is who I'm presuming you're implying, is a false comparison. People choose Windows at great peril, and when their line-level admin doesn't bother with patches or basic security practices, instead it's the guy who chose Windows that gets the blame.
The fact that a Canadian firm has a US parent doesn't excuse that company from giving away data entrusted to it by any Canadian government to any other firm or government not authorized to receive it by the same Canadian government
It's not just a US parent that's a risk, but even Canadian companies (such as the major banks) that own subsidiaries in the US - Royal Bank owns Liberty Insurance and several bank chains, TD owns an investment arm and some bank chains, BMO owns some bank chains. The US government can technically use the subsidiaries to weasle through to obtain information on the whole enterprise.
No you couldn't, that was my point. A non-standard addition to NT4 gave you that option, and with a bit of effort that add on could be grafted onto NT3.51 server. In NT5 it was included as standard.
Non-standard? Su came with the resouce kits going back to NT 3.1, allowing you to run a process under any account context. This is hardly some great chasm, and is no different than having a stock admin RPM installed.
And of course the API and security infrastructure was built since day one as a multi-user system -- services, for instance, take advantage of that functionality.
I find it absolutely astounding that you would consider NT not "multi-user" until it had a "RunAs" popup menu item in the context menu. Astounding.
How can this be to do with proprietry software and open source...
It has nothing whatsoever to do with proprietary Vs open source, and the addition of that incendiary flamebait in the submission was completely unnecessary trolling. Amazing how the majority of the comments thus far have been knee-jerk reactions with the chorus of the converted fervently preaching to their pewmates.
This isn't that hard to do. sender-id, spf, etc, does nothing.
...virus writers will just go to the extra effort of sending spam out the zombie PC through the owners' ISP mail server, and to your inbox...
These most certainly aren't total solutions, but they are gradual steps in the right direction (and really SMTP has always been the most absurdly abusable protocol. It's time to harden it a bit).
When a company like AOL or GMail commits to schemes like SenderID, SPF, or DomainKeys, they are effectively declaring their total responsibility over that mail source -- no longer is there confusion or deniability over whether a piece of mail was just sent direct or actually went through the Gmail system, for instance. As such, you can be sure that they will ensure that minimal amounts of spam are sent from their system -- so when Joe Blow downloads MonkeyPunchTM and it starts spamming out of his gmail account, they'll just shut the account down (detecting spam being sent from a source is pretty easy). I doubt virus writers will find much value in sending a couple of emails from each owned PC before the accounts are locked out. On the flip side the big providers no longer would have to deal with billions of spam returns for messages that were never sent from their system in the first place. Win win win.
We already know most semi-legitimate spammers are publishing SPF records on their throwaway domains which takes care of the other 10% of spam...
Obviously we're just getting started. Undoubtedly these systems, particularly DomainKeys, will develop into whitelist trust chains eventually, so it'll be rather easy to cut abusers out. It's also incredibly easy to build a "blacklist" of spamming domains, and again it's obvious that spammer will find little return in setting up domains for the sole purpose of spamming when it just gets cut out of the global loop in no time (not to mention that they're not stepping on legitimate email accounts in their from).
Air-conditioning, power-steering and even ABS still aren't standard despite costing next to nothing at build time
While it's incredibly dubious that they cost "next to nothing", they do cost something once the car drives off the lot - air conditioning, for instance, is probably the number 1 support issue in most vehicles, and under warranty those support issues add up to thousands of dollars for the manufacturer.
Manufacturers need to cripple cheaper cars to somehow...
So it's all a big collusion? Why doesn't a company like Kia stick all of this great stuff in their $9,000 car and competely own the market, instead of marketing an empty shell with minimal amenities. Right...because those things cost money.
Of course the margins increase on more expensive cars (and the margins on luxury SUVs are massive), but the idea that it's all just a big conspiracy is ludicrous - the car companies are locked in a very tight competition and want to kill each other, and it's not just coincidence that a Honda Civic is similarly equipped to a Hyundai Elantra and a Pontiac Sunfire.
1394... WHY?
Probably because the technology in supporting electronics and chipsets, not to mention programming knowledge, has been massively subsidized by the mainstream computing market. Going off and reinventing the wheel with their own transport system just isn't rational in such a case.
The best way to help the least fortunate is often by increasing the size of the economic pie - basically by making the middle class richer by encouraging economic commerce and innovation (cue naïve sarcastic remarks about Reagan's trickle down economics. Note, however, that I'm talking about the middle class rather than the rich. Furthermore, I am not speaking about tax cuts, but rather programs or initiatives that expand the economy, such as encouraging technological growth). Perhaps the premise behind free wireless is that it will lead to a slew of new programs and services in the San Francisco area that will lead to a lot of taxable commerce - tax revenue that can then be used to provide mental health support for some of the homeless. The most effective route to a goal isn't always the most direct.
BTW: An enjoyable read for the armchair economist is the very enlightening The Birth of Plenty : How the Prosperity of the Modern World was Created. It basically covers why some countries achieved such prosperity (hint - it isn't that they stole it from the poor countries).
In my neck of the woods there are few HDTV providers - Cogeco (a cable system), and Bell Canada ExpressVu (a satellite system). Both of them use entirely different hardware, and thus my choice of PVR box is dictated by my choice of provider (in my case it would have to be a Motorola DCT6208 or a Bell 6120..whatever manufacturer that is), and if I ponied up the $800CDN or so with one and became disatisfied with the service, I'd have to toss the box in the garbage.
Which is why I chose to rent the box for $20 a month - when the two-tuner version comes out (within the year) I'll do a swap. The point of all of this? I guess that ultimately I am choosing provider and the PVR comes along for the ride.
The NT series is NOW multiuser, since we now have full file permissions and can run different processes safely as different users - but it took many years to get to that point.
What are you talking about? NT has had full system wide ACLs (actually much more comprehensive and pervasive than Linux, or most other Unix variants) and process security context independence since the origin - don't think your Windows 95 knowledge has any relevance at all to the NT line. In other words your statement is just a load of ignorance.
Now the GUI has generally always been "single user" because the GUI doesn't (or rather didn't) have a "remote" connectivity option - with interactive logins it really is only possible for one person to be using the computer, so the GUI was designed around that. Of course you've always been able to launch processes under any security context.
So what you're saying is to get reliability out of a Windows server you need to be running a cluster.
I don't work for Microsoft, and I don't care if you buy or like Microsoft products, so I'm not speaking on behalf of Windows.
Having said that, there are generally two types of environments
-Environments where planned downtime can be accommodated (for instance an insurance company where all the employees go home at night and the system is taken offline for batch processing).
-Environments where there is no leeway for downtime.
In the latter situation clusters are quite simply the norm - either physical clusters or virtual clusters. Why? No matter how confident you are in your OS, hardware fails -- even when you have RAID 5 arrays and dual backplanes and redundant power supplies, one day the RAID card dies or the UPS serving that server fries and ironically takes out the server. Given this suddenly it becomes somewhat irrelevant if you have to rotate patch deployments through your servers.
You perform semantic gymnastics to avoid understanding the obvious.
If "the obvious" is whatever serves some anti-Microsoft crusade, then I suppose you're right. However it is painfully obvious to anyone without the ideological blinders that a system catrostrophically going down and a system being selectively rebooted at an opportune time are two very, very, very different things.
"But they're vulnerable to haxxers if they don't install the patch immediate," you might say. Sure, except that most critical boxes have nothing but the essential services running, and are often firewalled. The reality is that most patches are entirely unnecessary on these boxes, and most admins accept the scheduled downtime simply because it's easier than determining if the patch has any applicability to their boxes.
That's sort of the point. You have to reboot a Windows server more often. If rebooting once a month or so is acceptable (see Murphy's Law for schedule), then that's fine.
.NET for shared session), both of which Windows 2003 fully supports out of the box. In that case, with multiple balanced servers, you can freely patch any of them (or deal with failed hardware) with minimal or no customer impact -- maybe slightly slower responses with a smaller cluster.
But that's not the point - there is an implication that it is instability, i.e. uncontrolled downtime, when in reality it is controlled downtime (well accommodating the fact that sometimes security patches need to be installed relatively quickly). A controlled reboot of your server at 3 in the morning when all of your employees are at home is absolutely nothing like having your server crash at 10:00am. It is rhetorical hyperbole comparing them.
Of course for web applications this should be an entirely moot point - web apps with any requirement for reliability should be running in a cluster or network load balance arrangement (fully supported by
Undoubtedly there are a lot of incompetent sales folks out there (remember these people are making like $7-$10 an hour effectively -- you're not getting the pioneers in the field or the cream of the crop). However I worked in computer sales for a short while in the formative years, so I have perhaps a sore spot when it comes to salesperson complaints - occasionally I would get some weenie who has spent months meticulously reading about and absorbing every piece of information on one highly specialized area of the computing field, and they'd then march in and start quizzing me about my knowledge of the same (these were the types who never actually bought anything, but instead demanded my time on numerous days while they tried to solidify their decision to buy a $110 SCSI controller). Ultimately of course I was at a disadvantage (just as I was when the next weenie came in with highly specialized knowledge of some other area of the field) but this allowed the customer a temporary feeling of intellectual superiority. It was very annoying, and ultimately was a waste of time.
I should buy the Monster brand optical audio cable, because it has better shielding...He INSISTED that he had plugged the two in and 'head the difference'.
Remember that Monster cable is primarily sold through high end audio retailers, and those highly specialized salespeople will feed you the same BS.
...just ask them what MOSFET stands for...
...oh I'm forgetting the guy who told me I'd need "Digital" wires for my DVD player...
Err...why would you ask them what it stands for? Ultimately all of the engineering buzzwords come down to a couple of simple results - price and performance. If MOSFET is dead gerbils minced and blended into a chip insulator, or an advanced technique of containing the power of fusion it doesn't matter - does it make the stereo louder? Does it make the stereo sound better? Does it make it cheaper?
Could he be referring to SPDIF cables, which indeed are different than normal audio cables?