Define "successful". Windows is popular and a big money maker for Microsoft, but I'd argue that it isn't a good GUI (too many idiosyncrasies and non-obvious features).
How do I best seperate the UI code from the code that actually talks to the database.
Actually, in this question lies the answer. On a sheet of paper, draw a line separating the paper into two halves. In one half write "UI Module" and in the other half write "Database Interface Module".
These two modules can be written, so that the UI module uses only the public interfaces of the database module. This is the high level design of the application.
Designing each module well is an entirely different challenge, since the quality of the database module's interfaces determine how resilient the application is to changing requirements.
I guess the lesson here is that the really high-level design is trivial, but the design of each module is very hard and should be done by someone experienced in such things. Reading good books on the subject is only the beginning.
My hope is that the deal with Gateway is worth more than $135 million to Corel. My hope is that $135 million is more like Microsoft farting in Corel's house rather than Microsoft moving in and telling Corel how to live. If it is the latter, then this deal with Gateway will not outlast this week.
Let's say that operating systems was a truly "competitive" market with 1000 really world class, interoperable operating systems out there.
You're example is pretty rediculous. No market would support that many operating systems. However, a healthy market would support several operating systems (take commercial UNIX, for example). That's where true innovation can take place.
So yes, it is nice to see somebody lighting a fire under MS's butt and that's exactly what Corel, with an objectively inferior product will do--it will force MS to innovate and perhaps complete a little more on price.
No, what Microsoft needs is an inferno to shock them out of their lock-'em-all-in fantasy. You're looking for MS to innovate; however, what many of us really want is for MS to become humble, for cripes sake, shrink to a realistic size, and quit being such royal assholes.
...many people would say has tendencies towards natural (and in practice sometimes not so natural) monopolies.
No. The only reason Microsoft has a monopoly is through truly brilliant (and abusive) marketing of their truly shitty products. Microsoft is like the cult leader and the con artist in their methods. They are not naturally produced by any means.
If you want a real competitor for Word, take a look at Ami Pro.
Yes, Ami Pro was quite good. I used it early on in college and remember that its interface was much more intuitive than Word. It really focused on ease of use (but was quite useful) when Word was steadily ramping up its pile of useless features.
Any more, whenever I need to do real word processing, I use LaTeX. I also use OpenOffice.org for basic interoperability, but LaTeX offers more flexibility.
Live animals are being tortured for their entire lives just to bring you those pretty pictures, and you don't even care. Their microscopic howls of anguish leave you utterly unmoved.
I like flicking boogers onto a sun-baked sidewalk just to hear the little guys scream. Does that make me a bad person?
Not all religions. I believe that some Taoist/Buddhist sects regard sex as a spiritual exchange of Yin and Yang between male and female partners. In their philosophy, it actually makes very good sense.
It is really a Christian/Islamic oddity in the world that nudity and sex are so taboo.
Another thought: the universal identifier would have to be globally standardized. One way to accomplish this would be to set up a web site that simply doles out a sequence of 64 or 128-bit integers for each request. A person who writes a new utility would simply obtain a new number, which is guaranteed unique, from the website. When this utility is installed on another computer, this number is registered in the installed location (directory inode) and can be found using the same $PATH mechanism used today. A script or program dependent on this utility would simply make a request for the number instead of the name. It would be up to the utility author to use this number consistently across versions (also no different than the current practice of keeping the naming consistent).
A centralized website, however, is prone to scriptable request loops, which abuse the system. If a website is employed, some rules would need to be in place to prevent exhausting the 64 or 128 bit number space (for example, not allowing repeated requests from the same IP address in the same day).
Very interesting, but it seems there will frequently be some english or unixese visible on international desktops. This is mainly due to the tools, such as sh and grep, which were developed originally around ASCII at US universities and corporate research labs. I don't see a practical way around this using current tools.
What would be very interesting is a way of naming commands, so that their identity is independent of their names. Right now, grep is grep because it is called 'grep'. Why not give grep an identifier, such as '42', and have a name-identity mapping in the directory inode. This way, the identity never changes, but whatever name is displayed can be fully localized. An unfortunate side effect of this is that shell scripts would still contain the internationalized text but not the identifier, which still leads to non-portable scripts. This would probably be best solved with a simple post-processor that translates the script into a simple, yet portable, format where localized text is replaced by the universal identifiers.
Anyway, however it is done, really good localization will be one of the next major evolutions of software. Right now, it still appears that localization is superficial (like in the screenshots above).
Don't forget that Microsoft was first in brain-washing their customers to actually enjoy being raped daily.
Re:Blame college tuitions, not the dot-coms....
on
Generation Wrecked
·
· Score: 1
So why is it costing more? Huge state budget cuts for the state supported folks.
This is certainly true for state-supported schools (SC schools were hit hard recently), but private schools are also hiking tuition every year. I went to a private university who was increasing tuition at about 7%/year. Some other schools are even higher.
I think the larger problem is that all the universities are competing against eachother for student enrollment, research grants, sports contracts, and so on. They have this perception that they cannot fall behind no matter the cost. Literally, in the four years I was at school, nearly the whole campus was rennovated, a new visitor's center was built, four new dorms were built, the computer labs were kept state-of-the-art. No expense was spared. The campus was gorgeous after all this, but the amount of money spent was most likely obscene (there is no way alumni could have supported all of it).
8 CPUs stuck together, 1.33 times faster then a p4.
Actually, the IBM SpecFP numbers are from using only one core on the package giving the core all the cache that is normally shared. So, it is not a case of 8 Power 4 CPUs beating one P4 CPU; rather, it is a case of one Power 4 CPU with oodles of cache beating one P4 CPU.
Sun's 1.2GHz UltraSPARC III Cu runs at 50 watts. Their 650MHz UltraSPARC IIi runs at less than 18 watts.
For the same power consumption, a person can run two to six times as many non-Itanium 64-bit processors. Thus, any performance benefits of the Itanium for highly-parallel computation diminish in importance. It seems Intel can compete only on up-front cost, but the competition can surely respond to this, too.
Sun Microsystems will move into the desktop market, giving a familiar hardware name to Linux desktops, making it eaiser for IT staff to bring Linux PCs into their networks.
I absolutely love the idea that a company like Sun Microsystems can move into a market that Dell, Gateway, and HP have been too afraid to touch on a large scale. I read an insightful comment a while ago that Sun is one company that has absolutely nothing to lose (and potentially everything to gain) by shipping commodity PCs without Windows. Hopefully, Sun will market these PCs wisely (sticking to the traditional high quality of the Sun brand), so the lines distinguishing their SPARC offerings doesn't become blurred.
Microsoft will sit in the background, watching the TCO of Windows rise and the TCO of Linux drop, and the path for Linux domination will be ready for the world to walk down.
"Linux domination" might be a bit too strong, but one thing for certain is that Microsoft will definitely become less and less dominant as time passes. This will benefit us all as people's heads are removed from the proprietary arse of Microsoft and they can finally think and express themselves freely.
64-bit CPUs are neat, but almost no one will use anything other than 32-bit binarys. Even in UNIX systems that have been 64-bit for years, nearly all the userland is still 32-bit, because there are no advantages to having a 64-bit address space for those programs.
Unless Microsoft Office takes up 10GB of RAM one day to write a resume or I have to 'sed' many gigabytes of text in one file, 64-bit addressing isn't very useful.
I have a feeling that those people who have a real need for 64-bits already have 64-bit workstations from vendors other than Intel (Sun, SGI, IBM, HPaq). And, for serious work, workstations, such as the Sun Blade 2000, really aren't all that expensive, given their features (FibreChannel, 8MB CPU caches, beautiful case engineering, etc.).
Re:Blame college tuitions, not the dot-coms....
on
Generation Wrecked
·
· Score: 2
America needs to get college tuitions in check...
I agree. College tuition has been growing like the stock market should. It's a sin, really; do college administrators fail to care or understand the burden they place on students?
Another problem related to tuition is that young people just don't understand what $50,000 of debt really feels like. It is oppressive. I know a person who can't even make the minimum payments on his college-related debt. If he only had some foresight...
A complementary solution to getting colleges put in their place is to educate young people about the future value of a loan. Give them a theoretical salary and family expenses and, then, get them to find a way to pay everything off in less than ten years. Young people need to understand that college can be the second most expensive purchase (next to a home) in their lifetimes; yet, we are providing them almost no guidance in making that purchase responsibly.
Yet, financially speaking, you're smarter than 90% of the "adults" out there. However, don't save so aggressively that you are living like a hobo each day; save enough to cover your liabilities (insurance deductables, bills for a few months, watching out for the occasional car/house repair) and spend enough to live comfortably. Maintaining low or zero credit card debt is perhaps the smartest thing that you can do but do use your credit card regularly. Good credit is a lifesaver later on--you'll get awesome car loans and mortgages when those times come.
Also, don't fall for "extended warranties" (insurance on easily replaced things) and small-time financing (loans less than a few thousand dollars). They just suck money out of you needlessly and it's better just to own the small stuff outright. Basically, insurance covers catastrophes (not CD players & computers) and loans are for big things (again, not CD players and computers).
Keep managing you debt well, and you'll avoid a lot of stress later on. Good luck.
But don't tell me I'm doing something legally wrong.
Is it illegal to remove the catalytic converter? What about replacing the seat belts? Can you use tires not approved by the DOT?
Cars can be modded extensively, but these mods exist within a pretty specific legal framework. However, the main difference between car-related legislation and any legislation about software/DRM/copyright is that for cars the laws are for safety, emmissions, and consumer fraud protection. The laws for software/DRM/copyright are much less clear-cut (DRM for consumers? Yeah, that's it).
Support & maintenance contracts on Sun hardware are brutally expensive. This may account for the difference.
However, they are completely optional. Solaris and Sun hardware are very well documented, and there are a number of independent websites with FAQs and mailing lists. There is more than enough documentation to do without formal support from Sun, and regular maintenance items, such as patch clusters, are freely available. The only real benefits of Sun support are for people with truly critical applications, where the support is much cheaper than delays or downtime.
I think that most people who think that support contracts are too expensive are not in a position to really need them. Without support contracts, the Solaris sysadmin is essentially in the same boat as a Linux sysadmin, where the quality of the systems is somewhat independent of Solaris or Linux--the network architecture and policies are far more influential on TCO.
There are cases when Sun hardware can actually lower TCO, due to better OS-independent hardware diagnostic tools, which don't rely on a running OS kernel to work (they are accessible from a firmware command prompt). Although I haven't used them, Sun also offers remote management cards in some servers which allow OS-independent SNMP and telnet access to the card for monitoring, diagnosics, and remote reboot.
People who voice opinions about how much more Sun costs relative to System XYZ are either comparing apples to oranges, using pricing data from ten years ago, or overspecifying the Sun system to artificially inflate its price. Sun sells UltraSPARC-based systems costing $999 (one cpu) to millions of dollars (>100 CPUs). It is easy to pick a higher target when comparing Sun's systems to white-box x86 systems that really aren't fair comparisons. The "Linux TCO" article above is doing exactly this.
You forgot the little detail that Sun charges for Solaris--this might just account for the difference.
It doesn't account for the order-of-magnatude difference cited by the TCO "study". For small servers (one to four CPUs) the differences in software costs is really insignificant compared to other costs, such as the sysadmin's salary. As far as the hardware costs go, entry-level Sun boxes are damn cheap ($1000 to $2000), which also pales relative to other costs. As the servers get bigger (over two CPUs), Sun hardware really has many features that can be hard to find in any one x86-based package (ECC on all busses, FibreChannel, remote management support in hardware, hot swap CPUs/RAM/disks/power supplies, and so on). I feel very strongly that the TCO of Solaris is right on par with Linux (both of which have TCOs less than Windows).
I don't have to RTFA, because there is no way in hell Linux' TCO is 14% that of Solaris. Both are fundamentally the same type of system, with the same types of tools, requiring the same types of skills...their TCO numbers, in reality, are the same.
Whoever came up with those numbers, needs to be dragged out into the street and ridiculed. They are obviously biased, stupid, or both.
...why aren't people dropping their closed source software and downloading their open source counterparts in droves?
People are switching to Open Source in droves. What is Apache's share of public web servers? Why are entire governments seriously considering using Open Source as a matter of national security and enabling democracy? Why are the people around me increasingly becoming agitated at ass-hole companies like Microsoft and looking towards alternatives?
What about time to market?
How is time to market relevant for Open Source software? Getting software released on a deadline implies that the released software is, by definition, immature and buggy. Not having a deadline means whatever is released was ready to be released. What does an Open Source software project have to lose by taking the amount of time really needed to do something right?
What about lack of common features that customers want?
What features are you referring to? Would the "target audience" really be better served by hard-core MS Office lock-in or undocumented private kernel APIs? What about extended communications protocols that dictate what type of server clients can connect to? Are these good things?
The biggest problem is that the XBox runs an x86 and the PS2 an MIPS CPU.
Is it possible to translate already-compiled x86 code directly to MIPS code? This would allow parts of an XBox game to run natively on the PS2, and interfacing software could make the PS2 appear like an XBox to the translated code.
I know this is far from trivial, but it is interesting.
I've heard many claims that Solaris is very reliable - more reliable than Linux.
It's more accurate to say that Solaris is extremely reliable. There's oddball bugs here and there; however, I've witnessed a Solaris kernel panic once only after forgetting to upgrade a device driver after upgrading from Solaris 7 to Solaris 8. Otherwise, I work my workstation pretty hard months at a time (rebooting only when doing routine maintenance). The servers here are similarly reliable.
How much stability comes from Solaris itself, and how much comes from Sun's end-to-end control of the hardware?
The software itself is very robust. The hardware does help when the hardware has extra reliability features (RAID, ECC, hot-plugging), which the software leverages for better uptime. Random hardware failures are the most common cause of Solaris downtime.
When Solaris 9 is running on ferrel x86 hardware, will it display the same reliability as it's UltraSparc sibbling?
Generally, yes, but the device drivers are different and can have bugs unique to the x86 platform. The lower reliability x86 hardware would probably be more significant.
More importantly, will it even prove to be as reliable as Linux?
In the long term, Solaris should prove more reliable than Linux, because Sun has a more conservative approach to software upgrades and maintenance. There is generally less risk associated with updates to Solaris versus updates to Linux.
Slashdot Mirror
Learn from other successful software GUIs
Define "successful". Windows is popular and a big money maker for Microsoft, but I'd argue that it isn't a good GUI (too many idiosyncrasies and non-obvious features).
How do I best seperate the UI code from the code that actually talks to the database.
Actually, in this question lies the answer. On a sheet of paper, draw a line separating the paper into two halves. In one half write "UI Module" and in the other half write "Database Interface Module".
These two modules can be written, so that the UI module uses only the public interfaces of the database module. This is the high level design of the application.
Designing each module well is an entirely different challenge, since the quality of the database module's interfaces determine how resilient the application is to changing requirements.
I guess the lesson here is that the really high-level design is trivial, but the design of each module is very hard and should be done by someone experienced in such things. Reading good books on the subject is only the beginning.
My hope is that the deal with Gateway is worth more than $135 million to Corel. My hope is that $135 million is more like Microsoft farting in Corel's house rather than Microsoft moving in and telling Corel how to live. If it is the latter, then this deal with Gateway will not outlast this week.
Let's say that operating systems was a truly "competitive" market with 1000 really world class, interoperable operating systems out there.
...many people would say has tendencies towards natural (and in practice sometimes not so natural) monopolies.
You're example is pretty rediculous. No market would support that many operating systems. However, a healthy market would support several operating systems (take commercial UNIX, for example). That's where true innovation can take place.
So yes, it is nice to see somebody lighting a fire under MS's butt and that's exactly what Corel, with an objectively inferior product will do--it will force MS to innovate and perhaps complete a little more on price.
No, what Microsoft needs is an inferno to shock them out of their lock-'em-all-in fantasy. You're looking for MS to innovate; however, what many of us really want is for MS to become humble, for cripes sake, shrink to a realistic size, and quit being such royal assholes.
No. The only reason Microsoft has a monopoly is through truly brilliant (and abusive) marketing of their truly shitty products. Microsoft is like the cult leader and the con artist in their methods. They are not naturally produced by any means.
If you want a real competitor for Word, take a look at Ami Pro.
Yes, Ami Pro was quite good. I used it early on in college and remember that its interface was much more intuitive than Word. It really focused on ease of use (but was quite useful) when Word was steadily ramping up its pile of useless features.
Any more, whenever I need to do real word processing, I use LaTeX. I also use OpenOffice.org for basic interoperability, but LaTeX offers more flexibility.
Live animals are being tortured for their entire lives just to bring you those pretty pictures, and you don't even care. Their microscopic howls of anguish leave you utterly unmoved.
I like flicking boogers onto a sun-baked sidewalk just to hear the little guys scream. Does that make me a bad person?
Sex is really the cardinal sin in religion.
Not all religions. I believe that some Taoist/Buddhist sects regard sex as a spiritual exchange of Yin and Yang between male and female partners. In their philosophy, it actually makes very good sense.
It is really a Christian/Islamic oddity in the world that nudity and sex are so taboo.
Another thought: the universal identifier would have to be globally standardized. One way to accomplish this would be to set up a web site that simply doles out a sequence of 64 or 128-bit integers for each request. A person who writes a new utility would simply obtain a new number, which is guaranteed unique, from the website. When this utility is installed on another computer, this number is registered in the installed location (directory inode) and can be found using the same $PATH mechanism used today. A script or program dependent on this utility would simply make a request for the number instead of the name. It would be up to the utility author to use this number consistently across versions (also no different than the current practice of keeping the naming consistent).
A centralized website, however, is prone to scriptable request loops, which abuse the system. If a website is employed, some rules would need to be in place to prevent exhausting the 64 or 128 bit number space (for example, not allowing repeated requests from the same IP address in the same day).
Very interesting, but it seems there will frequently be some english or unixese visible on international desktops. This is mainly due to the tools, such as sh and grep, which were developed originally around ASCII at US universities and corporate research labs. I don't see a practical way around this using current tools.
What would be very interesting is a way of naming commands, so that their identity is independent of their names. Right now, grep is grep because it is called 'grep'. Why not give grep an identifier, such as '42', and have a name-identity mapping in the directory inode. This way, the identity never changes, but whatever name is displayed can be fully localized. An unfortunate side effect of this is that shell scripts would still contain the internationalized text but not the identifier, which still leads to non-portable scripts. This would probably be best solved with a simple post-processor that translates the script into a simple, yet portable, format where localized text is replaced by the universal identifiers.
Anyway, however it is done, really good localization will be one of the next major evolutions of software. Right now, it still appears that localization is superficial (like in the screenshots above).
Don't forget that Microsoft was first in brain-washing their customers to actually enjoy being raped daily.
So why is it costing more? Huge state budget cuts for the state supported folks.
This is certainly true for state-supported schools (SC schools were hit hard recently), but private schools are also hiking tuition every year. I went to a private university who was increasing tuition at about 7%/year. Some other schools are even higher.
I think the larger problem is that all the universities are competing against eachother for student enrollment, research grants, sports contracts, and so on. They have this perception that they cannot fall behind no matter the cost. Literally, in the four years I was at school, nearly the whole campus was rennovated, a new visitor's center was built, four new dorms were built, the computer labs were kept state-of-the-art. No expense was spared. The campus was gorgeous after all this, but the amount of money spent was most likely obscene (there is no way alumni could have supported all of it).
8 CPUs stuck together, 1.33 times faster then a p4.
Actually, the IBM SpecFP numbers are from using only one core on the package giving the core all the cache that is normally shared. So, it is not a case of 8 Power 4 CPUs beating one P4 CPU; rather, it is a case of one Power 4 CPU with oodles of cache beating one P4 CPU.
(compare with 135 Watts for Intel's Itanium 2)
Sun's 1.2GHz UltraSPARC III Cu runs at 50 watts. Their 650MHz UltraSPARC IIi runs at less than 18 watts.
For the same power consumption, a person can run two to six times as many non-Itanium 64-bit processors. Thus, any performance benefits of the Itanium for highly-parallel computation diminish in importance. It seems Intel can compete only on up-front cost, but the competition can surely respond to this, too.
Cell Phone "camming" will become a rage among teenagers...
How long until cell phones are hidden or left accidentally in the high school locker rooms in "streaming" mode?
Sun Microsystems will move into the desktop market, giving a familiar hardware name to Linux desktops, making it eaiser for IT staff to bring Linux PCs into their networks.
I absolutely love the idea that a company like Sun Microsystems can move into a market that Dell, Gateway, and HP have been too afraid to touch on a large scale. I read an insightful comment a while ago that Sun is one company that has absolutely nothing to lose (and potentially everything to gain) by shipping commodity PCs without Windows. Hopefully, Sun will market these PCs wisely (sticking to the traditional high quality of the Sun brand), so the lines distinguishing their SPARC offerings doesn't become blurred.
Microsoft will sit in the background, watching the TCO of Windows rise and the TCO of Linux drop, and the path for Linux domination will be ready for the world to walk down.
"Linux domination" might be a bit too strong, but one thing for certain is that Microsoft will definitely become less and less dominant as time passes. This will benefit us all as people's heads are removed from the proprietary arse of Microsoft and they can finally think and express themselves freely.
But I want 64 bit processing to take off.
64-bit CPUs are neat, but almost no one will use anything other than 32-bit binarys. Even in UNIX systems that have been 64-bit for years, nearly all the userland is still 32-bit, because there are no advantages to having a 64-bit address space for those programs.
Unless Microsoft Office takes up 10GB of RAM one day to write a resume or I have to 'sed' many gigabytes of text in one file, 64-bit addressing isn't very useful.
I have a feeling that those people who have a real need for 64-bits already have 64-bit workstations from vendors other than Intel (Sun, SGI, IBM, HPaq). And, for serious work, workstations, such as the Sun Blade 2000, really aren't all that expensive, given their features (FibreChannel, 8MB CPU caches, beautiful case engineering, etc.).
America needs to get college tuitions in check...
I agree. College tuition has been growing like the stock market should. It's a sin, really; do college administrators fail to care or understand the burden they place on students?
Another problem related to tuition is that young people just don't understand what $50,000 of debt really feels like. It is oppressive. I know a person who can't even make the minimum payments on his college-related debt. If he only had some foresight...
A complementary solution to getting colleges put in their place is to educate young people about the future value of a loan. Give them a theoretical salary and family expenses and, then, get them to find a way to pay everything off in less than ten years. Young people need to understand that college can be the second most expensive purchase (next to a home) in their lifetimes; yet, we are providing them almost no guidance in making that purchase responsibly.
I'm only 16.
Yet, financially speaking, you're smarter than 90% of the "adults" out there. However, don't save so aggressively that you are living like a hobo each day; save enough to cover your liabilities (insurance deductables, bills for a few months, watching out for the occasional car/house repair) and spend enough to live comfortably. Maintaining low or zero credit card debt is perhaps the smartest thing that you can do but do use your credit card regularly. Good credit is a lifesaver later on--you'll get awesome car loans and mortgages when those times come.
Also, don't fall for "extended warranties" (insurance on easily replaced things) and small-time financing (loans less than a few thousand dollars). They just suck money out of you needlessly and it's better just to own the small stuff outright. Basically, insurance covers catastrophes (not CD players & computers) and loans are for big things (again, not CD players and computers).
Keep managing you debt well, and you'll avoid a lot of stress later on. Good luck.
But don't tell me I'm doing something legally wrong.
Is it illegal to remove the catalytic converter? What about replacing the seat belts? Can you use tires not approved by the DOT?
Cars can be modded extensively, but these mods exist within a pretty specific legal framework. However, the main difference between car-related legislation and any legislation about software/DRM/copyright is that for cars the laws are for safety, emmissions, and consumer fraud protection. The laws for software/DRM/copyright are much less clear-cut (DRM for consumers? Yeah, that's it).
Support & maintenance contracts on Sun hardware are brutally expensive. This may account for the difference.
However, they are completely optional. Solaris and Sun hardware are very well documented, and there are a number of independent websites with FAQs and mailing lists. There is more than enough documentation to do without formal support from Sun, and regular maintenance items, such as patch clusters, are freely available. The only real benefits of Sun support are for people with truly critical applications, where the support is much cheaper than delays or downtime.
I think that most people who think that support contracts are too expensive are not in a position to really need them. Without support contracts, the Solaris sysadmin is essentially in the same boat as a Linux sysadmin, where the quality of the systems is somewhat independent of Solaris or Linux--the network architecture and policies are far more influential on TCO.
There are cases when Sun hardware can actually lower TCO, due to better OS-independent hardware diagnostic tools, which don't rely on a running OS kernel to work (they are accessible from a firmware command prompt). Although I haven't used them, Sun also offers remote management cards in some servers which allow OS-independent SNMP and telnet access to the card for monitoring, diagnosics, and remote reboot.
People who voice opinions about how much more Sun costs relative to System XYZ are either comparing apples to oranges, using pricing data from ten years ago, or overspecifying the Sun system to artificially inflate its price. Sun sells UltraSPARC-based systems costing $999 (one cpu) to millions of dollars (>100 CPUs). It is easy to pick a higher target when comparing Sun's systems to white-box x86 systems that really aren't fair comparisons. The "Linux TCO" article above is doing exactly this.
You forgot the little detail that Sun charges for Solaris--this might just account for the difference.
It doesn't account for the order-of-magnatude difference cited by the TCO "study". For small servers (one to four CPUs) the differences in software costs is really insignificant compared to other costs, such as the sysadmin's salary. As far as the hardware costs go, entry-level Sun boxes are damn cheap ($1000 to $2000), which also pales relative to other costs. As the servers get bigger (over two CPUs), Sun hardware really has many features that can be hard to find in any one x86-based package (ECC on all busses, FibreChannel, remote management support in hardware, hot swap CPUs/RAM/disks/power supplies, and so on). I feel very strongly that the TCO of Solaris is right on par with Linux (both of which have TCOs less than Windows).
I don't have to RTFA, because there is no way in hell Linux' TCO is 14% that of Solaris. Both are fundamentally the same type of system, with the same types of tools, requiring the same types of skills...their TCO numbers, in reality, are the same.
Whoever came up with those numbers, needs to be dragged out into the street and ridiculed. They are obviously biased, stupid, or both.
...why aren't people dropping their closed source software and downloading their open source counterparts in droves?
People are switching to Open Source in droves. What is Apache's share of public web servers? Why are entire governments seriously considering using Open Source as a matter of national security and enabling democracy? Why are the people around me increasingly becoming agitated at ass-hole companies like Microsoft and looking towards alternatives?
What about time to market?
How is time to market relevant for Open Source software? Getting software released on a deadline implies that the released software is, by definition, immature and buggy. Not having a deadline means whatever is released was ready to be released. What does an Open Source software project have to lose by taking the amount of time really needed to do something right?
What about lack of common features that customers want?
What features are you referring to? Would the "target audience" really be better served by hard-core MS Office lock-in or undocumented private kernel APIs? What about extended communications protocols that dictate what type of server clients can connect to? Are these good things?
The biggest problem is that the XBox runs an x86 and the PS2 an MIPS CPU.
Is it possible to translate already-compiled x86 code directly to MIPS code? This would allow parts of an XBox game to run natively on the PS2, and interfacing software could make the PS2 appear like an XBox to the translated code.
I know this is far from trivial, but it is interesting.
I've heard many claims that Solaris is very reliable - more reliable than Linux.
It's more accurate to say that Solaris is extremely reliable. There's oddball bugs here and there; however, I've witnessed a Solaris kernel panic once only after forgetting to upgrade a device driver after upgrading from Solaris 7 to Solaris 8. Otherwise, I work my workstation pretty hard months at a time (rebooting only when doing routine maintenance). The servers here are similarly reliable.
How much stability comes from Solaris itself, and how much comes from Sun's end-to-end control of the hardware?
The software itself is very robust. The hardware does help when the hardware has extra reliability features (RAID, ECC, hot-plugging), which the software leverages for better uptime. Random hardware failures are the most common cause of Solaris downtime.
When Solaris 9 is running on ferrel x86 hardware, will it display the same reliability as it's UltraSparc sibbling?
Generally, yes, but the device drivers are different and can have bugs unique to the x86 platform. The lower reliability x86 hardware would probably be more significant.
More importantly, will it even prove to be as reliable as Linux?
In the long term, Solaris should prove more reliable than Linux, because Sun has a more conservative approach to software upgrades and maintenance. There is generally less risk associated with updates to Solaris versus updates to Linux.