"Microsoft believes only in the positive feedback of a strong market position - we're popular because we're popular.... This is backed up by the history where inferior MS products beat out supperior competition just because they got on more desktops."
Recently, I read a book about the Pacific war in World War II. Your comment about Microsoft reminds me of the concept of Japanese "Fighting Spirit" or "Divine Wind" somehow guaranteeing victory, in spite of the whole country being bombed into a stone age. Perhaps there are parallels between Microsoft management and Japanese military leadership at that time.
(I read only one book on this, so flame gently if I'm completely wrong)
it is interesting just how involved the contestants are. This contest is their life. They mentioned several times in the show how many months of long workdays they spent to build and program these cars. And, then, who owns the work? Do they at least get patent recognition on some of the innovations? Some of the software they talked about was truly seriously cool stuff.
Sidenote: One hour of Nova or Frontline is like watching 5 days worth of "learning" and "discovery" shows elsewhere. It's amazing how good some of these shows are.
"The computer industry seems to have this idea that we want to combine all our gadgetry into a single box."
It's good marketing. We all have this dream of the perfectly reliable infinitely durable all-media-playing box that is the size of a box of matches and costs under $200. Of course, reality is that these devices have mechanical components, capacitors that have limited lifetimes, fans that wear out, etc.
But the dream remains alive in the minds of customers. That's probably enough to keep the manufacturers going. Ten years ago, they were trying to sell 486/Pentium PCs as "multimedia centers", complete with cheap speakers, a microphone that clamps to the monitor, a fax/modem for "VOIP", and a cute encyclopedia CDROM. They completely sucked, and the encyclopedia was a novelty at best--yet every new PC was like this.
It's a complex problem for developers. They may have been given access to only Windows 95 and 98 for testing, and there might be other things about Windows NT/2000 that could break the game in subtle ways. This would result in lots of preventable support calls/e-mails. Even if DirectX worked, there are lots of other libraries whose behavior could change, and those changes could be undocumented.
Of course, they should have commited resources for testing under Windows NT, but it's possible they didn't have those resources or didn't think there was a significant gaming market there at the time.
Re:You want faster Windows?
on
Why Windows is Slow
·
· Score: 2, Interesting
"remove unwanted components/drivers, preset Windows settings, slipstream hotfixes and service packs..."
People have to be careful with this. I've been burned a couple times using the "custom install" options on software, where there is a list of 3000 packages to pick from, without any solid documentation of each package. I think I'm going to get this super-thin installation with only the software I need, but that lasts about two days before an application requires something that is missing.
We are so overloaded with information that we don't see everything that is presented to us. Content providers have determined the obvious solution is to present everything twice, hoping that people will get caught the second time around.
We are so overloaded with information that we don't see everything that is presented to us. Content providers have determined the obvious solution is to present everything twice, hoping that people will get caught the second time around.
Everytime BluRay vs HDDVD comes up, someone mentions that good players will probably support both formats, anyway, so it'll end up like CDDVD+-=!%*RWRAMRWORMROMWhatever.
I go to a store and price out a $40 DVD burner, and it just has a bazillion acronyms in the "features" section. Good enough, I didn't feel like reading up on 40 different standards to see what I want. In ten years, the $40 BluRay/HDDVD burners will probably be the same (I hope).
"Why go for just 90% of the software market when with additional effort and degraded performance you can approach 100% without recompiling."
That 90% is not a stable 90%! Any smart businessperson leaves a way out, if things go sour. Also, the performance hit is not that big, anymore. Several Java applications I use have very acceptable performance, which is what applications programmers should prioritize over blistering hand-optimized assembly performance. We're not talking kernel networking routines, here.
"Whenever I come up with an idea I ask the question: Is this old-school? If the answer is yes, I forget all about it no matter how promising it might be."
So, you would say platform-specific development which is tied to one vendor's bottom line is good? That is where software development is trending away from, as consumer-level prices continue to drop. From mainframes to UNIX to Windows to Linux and Open Source UNIX and tools like Java, Python, etc. Applications which are tied to Win32 or DOS or SunOS 4 or System 3 or whatever just aren't viable long term products!
"The whole.net issue was as typical marketing over technical sense, to push away people from Java (which did not work anyway)"
Many people have realized that developing with Java can leave the door open for relatively simple ports to Mac OS X, Solaris, and Linux, even if the original development was done on Windows. I'd think that any new software project would be silly to tie itself to one operating system at the programming language level (even if not in Java: Python, Ruby, etc.). For public companies, shareholders might even be smart to demand this as a way to protect the business model. Single-platform applications programming is pretty old-school, IMO.
"Can you find a 400 mile high hill, somewhere where there's no air resistance?"
In fact, when I was a young'n, the school was on top of that hill! We had to walk out of the atmosphere in freeze-drying temperatures every day! Kids, these days, get heated pressurized busses! They're spoiled, I say! There's nothing like a walk through the stratosphere to put the world in perspective!
It just occurred to me that the domain system has similar flaws as the DOS/Windows drive letter system. The top-level.com,.net, C:, D:, etc. are so separated that moving between them is inconvenient at best, and once committed to one choice it pretty much is permanent. E.g., don't try to move MS Office from C: to D: or vice versa...the registry is *not* your friend. There are just too many adult sites (spread internationally) committed to.com,.net,.biz, etc. that filters working on.xxx will accomplish nothing.
"With that as a starting point it should have been an order of magnitude faster than the previous step."
In theory, but rarely does a software project work out like that. As software modules get inter-connected the total number of possible interactions does increase exponentially, and thorough software testing can become practically impossible. Part of this is that experts in the higher-level modules tend to become experts in the underlying modules, because of discovering annoying bugs and lacking functionality in the underlying modules. Then, it can become a finger-pointing game of who is responsible for that functionality or if the bugs will simply be ignored and worked around. It can really suck for anyone who signed on to be "just an engineer" or "just a programmer".
Linux/BSD/OpenSolaris/KDE/GNOME/whatever development is spread globally and asynchronously. If GWidgetXYZ ships 6 months after KWidgetABC no one cares. The distribution maintainers glean what they can at the time they are doing their integration and ship FooBar Linux 4.7.9b with the latest GWidgetXYZ but defer KWidgetABC to a later update. Everyone basically goes with this flow and accepts where the latest-n-greatest vs. stability tradeoffs are.
This is why commercial UNIX and Linux continue to make money while there is an immense base of users who get it for free. The base of free users build up a huge experience pool that feeds into the commercial business. It seems to be working, because companies like Red Hat are still hanging on, in spite of being dwarfed by Microsoft, HP, IBM, etc.
The "many eyes" concept also seems to work to a degree. Even though I'm not a super-geek, I've reported a bug or two over the years that made a difference. Sure that's just one or two bugs, but that's more than I ever reported about Windows or Office! I wouldn't even know how to being reporting a bug in Windows. There's no transparency.
From what i understand they tried to rewrite the dungpile of spaghetticode in.Net technologies but failed to get any descent performance and stability, Late into the process they decided to rip the new code out and start over with the old code again.
Wow, if that's true, it seems they are managing their projects like an inexperienced stock trader. Should I sell?...ooh, the stock price went up...should I hold on?...wait is that a bearish pattern forming...uh, the analyst says to hold on...oh crap I should have sold when I had the chance...
Starting out with good fundamentals (design, staffing, etc.) would have saved a lot of trouble. However, I guess they were so in need of replacing Java in midst of all the lawsuits that they were in a real pinch.
Actually, it was supposed to be images of the new "5 4 3 2 1...CRASH" feature, but the software couldn't count, stopped, and couldn't remember where it was!
I'm really looking forward to restaurants serving "Fusion Broiled" hamburgers. I've heard it really has an edge over flame broiling, because the outer layer of the grilled burger patty is literally a new form of matter.
In reality managers want people who are generalists so they can be poured around easier and not become important enough at a specialty to rise above competition and demand more wages/be harder to replace.
This is balanced against the manager's need to actually produce competitive products for the company's bottom line. I've seen what happens when someone or even a small team tries to digest something like J2EE whole, on top of a database and UNIX/Windows. It falls apart, because they don't even know how to debug it!
Dev 1: "Hmmm....there's an ORA-19394938483838 error in the log file but no stack trace and the exception type was lost in the container somewhere....(*thud*)"
Yes, because a good manager and a good secretary take a huge load off of the developers/engineers. I consider any office who lays off their secretary to be an office which has 'jumped the shark', because the developers will now be wasting time making travel arrangements, dealing with expense reports, calling human resources, etc.
In any given project there are just so many parallel tasks. The optimum number of developers is about the same as that level of parallelism (plus a secretary and a manager). It allows compartmentalizing things, so each developer has a chance to become an expert in that area and be productive. Adding more developers just increases communication overhead, training overhead, petty squabbles, micromanagement of the mess, etc. Taking away developers leaves holes that will require additional time to complete.
I hope the article summary is wrong and that Microsoft isn't so incompetent as to substantially re-write an operating system in the last year of its development! Talk about a death spiral.
"That's no moon, it's the accumulated mass of all our new bugs!"
Slashdot Mirror
I bet they can't wait for Toilet 1.0!
Sometimes being an alpha tester isn't worth it.
"Microsoft believes only in the positive feedback of a strong market position - we're popular because we're popular. ... This is backed up by the history where inferior MS products beat out supperior competition just because they got on more desktops."
Recently, I read a book about the Pacific war in World War II. Your comment about Microsoft reminds me of the concept of Japanese "Fighting Spirit" or "Divine Wind" somehow guaranteeing victory, in spite of the whole country being bombed into a stone age. Perhaps there are parallels between Microsoft management and Japanese military leadership at that time.
(I read only one book on this, so flame gently if I'm completely wrong)
it is interesting just how involved the contestants are. This contest is their life. They mentioned several times in the show how many months of long workdays they spent to build and program these cars. And, then, who owns the work? Do they at least get patent recognition on some of the innovations? Some of the software they talked about was truly seriously cool stuff.
Sidenote: One hour of Nova or Frontline is like watching 5 days worth of "learning" and "discovery" shows elsewhere. It's amazing how good some of these shows are.
Huh? Only 768? Bah!
I will be designing and selling an Earth Simulator on one chip. Only $50,000,000,000...paid in advance.
"...entire wall as a monitor."
Just don't click on that Goatse link!
"The computer industry seems to have this idea that we want to combine all our gadgetry into a single box."
It's good marketing. We all have this dream of the perfectly reliable infinitely durable all-media-playing box that is the size of a box of matches and costs under $200. Of course, reality is that these devices have mechanical components, capacitors that have limited lifetimes, fans that wear out, etc.
But the dream remains alive in the minds of customers. That's probably enough to keep the manufacturers going. Ten years ago, they were trying to sell 486/Pentium PCs as "multimedia centers", complete with cheap speakers, a microphone that clamps to the monitor, a fax/modem for "VOIP", and a cute encyclopedia CDROM.
They completely sucked, and the encyclopedia was a novelty at best--yet every new PC was like this.
It's a complex problem for developers. They may have been given access to only Windows 95 and 98 for testing, and there might be other things about Windows NT/2000 that could break the game in subtle ways. This would result in lots of preventable support calls/e-mails. Even if DirectX worked, there are lots of other libraries whose behavior could change, and those changes could be undocumented.
Of course, they should have commited resources for testing under Windows NT, but it's possible they didn't have those resources or didn't think there was a significant gaming market there at the time.
"remove unwanted components/drivers, preset Windows settings, slipstream hotfixes and service packs..."
People have to be careful with this. I've been burned a couple times using the "custom install" options on software, where there is a list of 3000 packages to pick from, without any solid documentation of each package. I think I'm going to get this super-thin installation with only the software I need, but that lasts about two days before an application requires something that is missing.
We are so overloaded with information that we don't see everything that is presented to us. Content providers have determined the obvious solution is to present everything twice, hoping that people will get caught the second time around.
We are so overloaded with information that we don't see everything that is presented to us. Content providers have determined the obvious solution is to present everything twice, hoping that people will get caught the second time around.
Everytime BluRay vs HDDVD comes up, someone mentions that good players will probably support both formats, anyway, so it'll end up like CDDVD+-=!%*RWRAMRWORMROMWhatever.
I go to a store and price out a $40 DVD burner, and it just has a bazillion acronyms in the "features" section. Good enough, I didn't feel like reading up on 40 different standards to see what I want. In ten years, the $40 BluRay/HDDVD burners will probably be the same (I hope).
"Why go for just 90% of the software market when with additional effort and degraded performance you can approach 100% without recompiling."
That 90% is not a stable 90%! Any smart businessperson leaves a way out, if things go sour. Also, the performance hit is not that big, anymore. Several Java applications I use have very acceptable performance, which is what applications programmers should prioritize over blistering hand-optimized assembly performance. We're not talking kernel networking routines, here.
"Whenever I come up with an idea I ask the question: Is this old-school? If the answer is yes, I forget all about it no matter how promising it might be."
So, you would say platform-specific development which is tied to one vendor's bottom line is good? That is where software development is trending away from, as consumer-level prices continue to drop. From mainframes to UNIX to Windows to Linux and Open Source UNIX and tools like Java, Python, etc. Applications which are tied to Win32 or DOS or SunOS 4 or System 3 or whatever just aren't viable long term products!
"The whole
Many people have realized that developing with Java can leave the door open for relatively simple ports to Mac OS X, Solaris, and Linux, even if the original development was done on Windows. I'd think that any new software project would be silly to tie itself to one operating system at the programming language level (even if not in Java: Python, Ruby, etc.). For public companies, shareholders might even be smart to demand this as a way to protect the business model. Single-platform applications programming is pretty old-school, IMO.
Well, it is more affordable...isn't that what Windows originally was? A more affordable version of UNIX?
"Can you find a 400 mile high hill, somewhere where there's no air resistance?"
In fact, when I was a young'n, the school was on top of that hill! We had to walk out of the atmosphere in freeze-drying temperatures every day! Kids, these days, get heated pressurized busses! They're spoiled, I say! There's nothing like a walk through the stratosphere to put the world in perspective!
It just occurred to me that the domain system has similar flaws as the DOS/Windows drive letter system. The top-level
"With that as a starting point it should have been an order of magnitude faster than the previous step."
In theory, but rarely does a software project work out like that. As software modules get inter-connected the total number of possible interactions does increase exponentially, and thorough software testing can become practically impossible. Part of this is that experts in the higher-level modules tend to become experts in the underlying modules, because of discovering annoying bugs and lacking functionality in the underlying modules. Then, it can become a finger-pointing game of who is responsible for that functionality or if the bugs will simply be ignored and worked around. It can really suck for anyone who signed on to be "just an engineer" or "just a programmer".
Linux/BSD/OpenSolaris/KDE/GNOME/whatever development is spread globally and asynchronously. If GWidgetXYZ ships 6 months after KWidgetABC no one cares. The distribution maintainers glean what they can at the time they are doing their integration and ship FooBar Linux 4.7.9b with the latest GWidgetXYZ but defer KWidgetABC to a later update. Everyone basically goes with this flow and accepts where the latest-n-greatest vs. stability tradeoffs are.
This is why commercial UNIX and Linux continue to make money while there is an immense base of users who get it for free. The base of free users build up a huge experience pool that feeds into the commercial business. It seems to be working, because companies like Red Hat are still hanging on, in spite of being dwarfed by Microsoft, HP, IBM, etc.
The "many eyes" concept also seems to work to a degree. Even though I'm not a super-geek, I've reported a bug or two over the years that made a difference. Sure that's just one or two bugs, but that's more than I ever reported about Windows or Office! I wouldn't even know how to being reporting a bug in Windows. There's no transparency.
From what i understand they tried to rewrite the dungpile of spaghetticode in .Net technologies but failed to get any descent performance and stability, Late into the process they decided to rip the new code out and start over with the old code again.
Wow, if that's true, it seems they are managing their projects like an inexperienced stock trader. Should I sell?...ooh, the stock price went up...should I hold on?...wait is that a bearish pattern forming...uh, the analyst says to hold on...oh crap I should have sold when I had the chance...
Starting out with good fundamentals (design, staffing, etc.) would have saved a lot of trouble. However, I guess they were so in need of replacing Java in midst of all the lawsuits that they were in a real pinch.
Yes, you now owe me $0.39 for every person who reads your post!
Actually, it was supposed to be images of the new "5 4 3 2 1...CRASH" feature, but the software couldn't count, stopped, and couldn't remember where it was!
I'm really looking forward to restaurants serving "Fusion Broiled" hamburgers. I've heard it really has an edge over flame broiling, because the outer layer of the grilled burger patty is literally a new form of matter.
Isn't the 'g' necessary only to replace multiple occurences?
$ echo "This election sponsored by Diebold" | sed -e "s/sponsored/decided/"
This election decided by Diebold
In reality managers want people who are generalists so they can be poured around easier and not become important enough at a specialty to rise above competition and demand more wages/be harder to replace.
This is balanced against the manager's need to actually produce competitive products for the company's bottom line. I've seen what happens when someone or even a small team tries to digest something like J2EE whole, on top of a database and UNIX/Windows. It falls apart, because they don't even know how to debug it!
Dev 1: "Hmmm....there's an ORA-19394938483838 error in the log file but no stack trace and the exception type was lost in the container somewhere....(*thud*)"
Dev 2" "Hey, is that guy still breathing?!?"
Yes, because a good manager and a good secretary take a huge load off of the developers/engineers. I consider any office who lays off their secretary to be an office which has 'jumped the shark', because the developers will now be wasting time making travel arrangements, dealing with expense reports, calling human resources, etc.
In any given project there are just so many parallel tasks. The optimum number of developers is about the same as that level of parallelism (plus a secretary and a manager). It allows compartmentalizing things, so each developer has a chance to become an expert in that area and be productive. Adding more developers just increases communication overhead, training overhead, petty squabbles, micromanagement of the mess, etc. Taking away developers leaves holes that will require additional time to complete.
I hope the article summary is wrong and that Microsoft isn't so incompetent as to substantially re-write an operating system in the last year of its development! Talk about a death spiral.
"That's no moon, it's the accumulated mass of all our new bugs!"