What does "great" mean? They were not high performance in the manner of Porsche. They were not high reliability like a Japanese car. They were not luxurious like a Rolls. That's not the point.
But they were "great" at their original design goal as stated: a good car in bad weather.
Horrifyingly poor management
on
A Requiem For Saab
·
· Score: 4, Insightful
Back before they developed the yuppie image and the high prices, they were just a nice solid car that was unstoppable in bad weather. Certainly they were more expensive than the typical car, but not so much so that they were unaffordable.
But GM really destroyed them by pushing them into a market that they were designed for.
We New Englanders still need a nice winter car, and Saab is not there for that purpose any more because they are just too darned expensive now. I only have one because I bought it used, there's no way I'm going to pay $40K for a car.
Saab was a modest company making a modest profit on a modest sales. GM came along and doubled their production and raised the prices. In the process they made the company much more fragile because now they had to maintain sales levels to pay down the expenses of expanding.
Really the story is not all that different from the typical failed high-tech company: crash and burn while attempting to grow out of the initial successful market. The projected sales increases don't happen. This failure pattern happens over and over again so many times, you'd think managers would learn.
A lesson to be learned and yet another reason for Europeans to be annoyed at Americans.
Your argument is over 20 years out of date
on
A Requiem For Saab
·
· Score: 3, Insightful
"Is that why they built a bunch of intensely front-heavy FWD vehicles with atrocious understeer?" "They also had reverse-mounted engines"
They stopped making these cars in the mid 80's.
Neither of those criticisms applies to the cars that they make today.
Well you are just trashing any concept of high performance as soon as you start dealing with SQL.
"Real" high-performance applications don't go anywhere near SQL, it's just a loss.
If your app is hitting an SQL database with any sort of frequency then the performance of your runtime is irrelevant.
I was thinking more along the lines of video games, number-crunching, physical simulations etc. where the heavy CPU utilization is happening in the app, not in the database.
I don't do.NET so I have no idea of its performance.
I do like the concept that Java is far removed from the physical hardware so the compiler can make all sorts of simplifying assumptions and nifty optimizations.
Modern Java compilers turn your java code into machine code that runs just as fast as languages like C and C++ which are pretty much just fancy macro assemblers.
Can someone with experience speak as to the relative performance of java and.NET?
My cousin was one of those guys with the keys and a gun and a buddy for many years. He's retired now and shares the stories at family reunions. He was a colonel so I'm sure he knows exactly what he can and cannot talk about. What's even better are his stories about winter life in rural North Dakota.
None of these products are perfect, they all have problems.
These are complex products with many many features and I don't imagine there are many of us who use all of them.
Stuff that seems like a minor issue to one user is a total show-stopper to another.
Your success with a product means nothing to another user if they are using it differently or are attempting to use a different set of features.
Really the only answer is you have to try them for yourself and see how they work for you.
Fortunately even the pay-for products have a free evaluation period so you can try them out without shelling out the cash.
If you work with the product all day then the cost of purchase is pretty irrelevant compared to what you are doing with it. How much is your time worth?
"$189 less than any other product. That makes it 189 times "BETTER""
I think you have some pretty serious issues with your math skills.
And you have no clue about costs.
The more expensive software product often makes for the lowest-cost system when you factor in the necessary hardware.
For example a $10,000 enterprise database is cheaper than a free one if it can do the same work on a $20,000 server that would require a $50,000 server for the free one.
I just flat out could not get VirtualBox to work correctly. I require a very complex network setup and their networking is not as robust as VMware.
My VMs are pushed out hard, running automated tests. I got occasional lock-ups in VirtualBox while VMware runs for days and days without a single problem.
"VMWare assumes the *entire* point of your system is to run VMWare"
Damned straight! Why else would I buy a machine with 8 cores and 32 Gb RAM?
"Try looking at the RPM"
What RPM? VMware Workstation 7 does not ship as an RPM any more. You are behind on the times.
"Contrast with VirtualBox"
Yes I did. They BOTH install lots of strange stuff on your machine. I did not see much difference.
The big difference I found is that VMware has sufficient quality for me to do my work. VirtualBox is so buggy that I cannot do my job with it. Believe me, I tried.
It's a drop in the bucket compared to the cost of the hardware needed to run any of these products well.
They say a max of 4 virtual CPU in the review but they FAIL to mention that each of these CPU can have 4 cores, so really it supports 16 virtual processors.
These two inaccuracies also just happen to be the only two "Cons" that they assign to VMware.
If you actually bother to boot up and try VirtualBox you will find it very buggy compared to VMware, to the point of being not very usable. I spent several days trying to get VirtualBox to work for me but there were just too many problems.
All of the things you are talking about are "solved problems" in just about every modern Linux distribution.
Your rant sounds like it was written circa 2002.
Perhaps you are talking about Fedora? Fedora is the "home distribution" for many of the projects that you mention.
Maybe you forget that Ubuntu cribbed almost all of that stuff from RedHat/Fedora?
Video resolution? Huh? Really who runs their monitor at less than the max? I'm running at 1920x1200 and there is NO reason to use anything else.
"without special package installation requirements" What does this mean? Does that mean that the distribution is bloated out with extra stuff that only a few users need? Why is this an advantage?
"It's all the same under the hood" - that is just not correct. There are some pretty MAJOR differences between the distributions regarding kernel versions, supported kernel features, versions of libraries, etc. These can be real show-stoppers when it comes time to install third-party applications.
Slashdot Mirror
"My point was that it was NEVER great."
What does "great" mean? They were not high performance in the manner of Porsche. They were not high reliability like a Japanese car. They were not luxurious like a Rolls. That's not the point.
But they were "great" at their original design goal as stated: a good car in bad weather.
Back before they developed the yuppie image and the high prices, they were just a nice solid car that was unstoppable in bad weather. Certainly they were more expensive than the typical car, but not so much so that they were unaffordable.
But GM really destroyed them by pushing them into a market that they were designed for.
We New Englanders still need a nice winter car, and Saab is not there for that purpose any more because they are just too darned expensive now. I only have one because I bought it used, there's no way I'm going to pay $40K for a car.
Saab was a modest company making a modest profit on a modest sales. GM came along and doubled their production and raised the prices. In the process they made the company much more fragile because now they had to maintain sales levels to pay down the expenses of expanding.
Really the story is not all that different from the typical failed high-tech company: crash and burn while attempting to grow out of the initial successful market. The projected sales increases don't happen. This failure pattern happens over and over again so many times, you'd think managers would learn.
A lesson to be learned and yet another reason for Europeans to be annoyed at Americans.
"Is that why they built a bunch of intensely front-heavy FWD vehicles with atrocious understeer?"
"They also had reverse-mounted engines"
They stopped making these cars in the mid 80's.
Neither of those criticisms applies to the cars that they make today.
.NET is not a language.
" do not collection $200."
Your English fails, you don't collect $200 either
Well you are just trashing any concept of high performance as soon as you start dealing with SQL.
"Real" high-performance applications don't go anywhere near SQL, it's just a loss.
If your app is hitting an SQL database with any sort of frequency then the performance of your runtime is irrelevant.
I was thinking more along the lines of video games, number-crunching, physical simulations etc. where the heavy CPU utilization is happening in the app, not in the database.
Really you should compare JVM to .NET.
JVM and .NET are the implementation platforms.
Both platforms are host to different languages.
Compare languages to languages and platforms to platforms.
I don't see anyone talking about performance.
I don't do .NET so I have no idea of its performance.
I do like the concept that Java is far removed from the physical hardware so the compiler can make all sorts of simplifying assumptions and nifty optimizations.
Modern Java compilers turn your java code into machine code that runs just as fast as languages like C and C++ which are pretty much just fancy macro assemblers.
Can someone with experience speak as to the relative performance of java and .NET?
You don't know my cousin. He was the prototype when they invented the word "conservative". He also left out a lot of details.
My cousin was one of those guys with the keys and a gun and a buddy for many years. He's retired now and shares the stories at family reunions. He was a colonel so I'm sure he knows exactly what he can and cannot talk about. What's even better are his stories about winter life in rural North Dakota.
This stuff has been out in the open for years.
"Any decent boxed pre-loaded PC from a place like Best Buy or Costco is more than adequate for running this stuff."
I guess that would depend.
You're not going to be able to run an automated test of a eight-machine database cluster on any kind of hardware available at Costco.
"since all our desktop computers are very old and run XP "
"there's no money in the budget for that".
"nor can we afford $2000-5000 for a decent computer"
"causing many days' worth of lost time."
You clearly have big problems with the whole concept of what is cheap and what is expensive.
None of these products are perfect, they all have problems.
These are complex products with many many features and I don't imagine there are many of us who use all of them.
Stuff that seems like a minor issue to one user is a total show-stopper to another.
Your success with a product means nothing to another user if they are using it differently or are attempting to use a different set of features.
Really the only answer is you have to try them for yourself and see how they work for you.
Fortunately even the pay-for products have a free evaluation period so you can try them out without shelling out the cash.
If you work with the product all day then the cost of purchase is pretty irrelevant compared to what you are doing with it. How much is your time worth?
Is "write to other parts of my filesystem" is the biggest complaint you can come up with?
Most of us go by criteria like "it does the job for me" or "I save two hours a week in downtime" when we evaluate software.
"$189 less than any other product. That makes it 189 times "BETTER""
I think you have some pretty serious issues with your math skills.
And you have no clue about costs.
The more expensive software product often makes for the lowest-cost system when you factor in the necessary hardware.
For example a $10,000 enterprise database is cheaper than a free one if it can do the same work on a $20,000 server that would require a $50,000 server for the free one.
Yes I do work with systems like that.
You are saying it yourself:
"There are cases where VMWare may be preferable"
My system is one of those cases.
I just flat out could not get VirtualBox to work correctly. I require a very complex network setup and their networking is not as robust as VMware.
My VMs are pushed out hard, running automated tests. I got occasional lock-ups in VirtualBox while VMware runs for days and days without a single problem.
There's no 3-D support. On a modern machine that is not "just fine".
Or watching videos. I have a Windows XP VM on my Linux machine and Netflix movies play great on it.
Show your face if you want to engage in name calling. Otherwise shut up.
VirtualBox is fine for what it is, but I don't have the time to struggle with it.
"some really far out edge cases" - welcome to my life.
But VMware is BETTER. And $189 approaches zero compared to the cost of the time I spend with it.
"VMWare assumes the *entire* point of your system is to run VMWare"
Damned straight! Why else would I buy a machine with 8 cores and 32 Gb RAM?
"Try looking at the RPM"
What RPM? VMware Workstation 7 does not ship as an RPM any more. You are behind on the times.
"Contrast with VirtualBox"
Yes I did. They BOTH install lots of strange stuff on your machine. I did not see much difference.
The big difference I found is that VMware has sufficient quality for me to do my work. VirtualBox is so buggy that I cannot do my job with it. Believe me, I tried.
Oops I was wrong about the max number of processors, it really is 4, I just tried it.
They call VMware Workstation at $189 "expensive".
It's a drop in the bucket compared to the cost of the hardware needed to run any of these products well.
They say a max of 4 virtual CPU in the review but they FAIL to mention that each of these CPU can have 4 cores, so really it supports 16 virtual processors.
These two inaccuracies also just happen to be the only two "Cons" that they assign to VMware.
If you actually bother to boot up and try VirtualBox you will find it very buggy compared to VMware, to the point of being not very usable. I spent several days trying to get VirtualBox to work for me but there were just too many problems.
"You're a classic example of why Linux has problems,"
Bullshit. You can't find a single problem in Linux that can be blamed on me, cuz I didn't write any of it.
"Ever tried using a 15" laptop with 1600x1200 resolution? The text is impossible to read"
Make the text bigger. If you can't then there is a problem with the software.
Huh?
All of the things you are talking about are "solved problems" in just about every modern Linux distribution.
Your rant sounds like it was written circa 2002.
Perhaps you are talking about Fedora? Fedora is the "home distribution" for many of the projects that you mention.
Maybe you forget that Ubuntu cribbed almost all of that stuff from RedHat/Fedora?
Video resolution? Huh? Really who runs their monitor at less than the max? I'm running at 1920x1200 and there is NO reason to use anything else.
"without special package installation requirements" What does this mean? Does that mean that the distribution is bloated out with extra stuff that only a few users need? Why is this an advantage?
"It's all the same under the hood" - that is just not correct. There are some pretty MAJOR differences between the distributions regarding kernel versions, supported kernel features, versions of libraries, etc. These can be real show-stoppers when it comes time to install third-party applications.