When I rewrote a some image effects in KDE (blending, convert to B&W etc) using MMX/SSE a while back, the SIMD version was around 4x faster on an Athlon and 10x on a P4. Of course, that's only for one very small part of KDE and the overall performance increase was probably less than 1%.
Come on, I'm a mac user and I wouldn't compare the Mini's price to a Dell. Add a mouse, keyboard and monitor and you're up to $750 or so - for that money, you could get two entry-level Dells. Personally, I'd rather have the Mini, but that was the OP's point.
Doesn't that somewhat negate the point of buying a pre-built machine? Installing the software, tracking down drivers, running windows update etc take much longer than building the machine itself.
The "programmers are replacable" idea is absolutely crazy. Even ignoring the creativity argument, it certainly explains why most games are buggy pieces of crap. Experience in software is hugely valuable: a kid straight out of college may be able to write an algorithm, but far more important is the ability to write maintainable, testable, readable code. The idea of a whole development team made up of people who write code like I did when I started out would scare the hell out of me.
Given those conditions, I'm amazed that there are any programmers willing to work in gaming at all even for a few years. I work in business software and, while we do have crunch weeks from time to time they are the exception and we always try to figure out afterwards what went wrong and how we can make sure it doesn't happen again.
Of course megapixels don't mean much without the optics to match. Compare these two images, both taken at 640x480. Spotting the photo taken on the Treo isn't hard...
The last conference I submitted to (CAPE2004) made it very clear that the use of Word was greatly preferred to LaTeX. They only provided a Word template and made lots of fuss about the PDF I submitted. Very annoying, but I think that might be the way things are headed.
OO.org2 has got a lot better, but I still found it easier to fire up VMWare and use Office 2003, mostly for Excel and some Word. Office 2003 is quicker, has an easier interface, fewer bugs and of course better Office file format compatibility. As a word processor, I actually prefer Pages overall, although Office 12 looks pretty good.
The existence of OO is important, of course, as competition for MS, and it has some nice features (print to PDF etc). In my experience, OO2 is somewhere around "Office 96.5" - pretty good, but clunky and quirky with some odd bugs.
While that is possible, I haven't seen anything like it in the file format so far. You can see samples and the draft spec on the OpenXML website. From what I've seen, it should at least be easier to interoperate with Office12 XML than with the old binary formats. Not that that is saying all that much, of course.
Are there even many mobile devices without Arm cores of some kind these days? Almost every PDA and mobile phone seems to have an ARM7 or 9 core, which probably means that ARM processors outnumber x86 processors several times over.
That's inevitable, though. A computer scientist simply can't add enough value to make it worthwhile paying them so much money without moving into senior management.
Quite the reverse - they've had to fire incompetent engineers and a lot of their staff is there based on recommendations from senior engineers. They do employ a lot of people from abroad (Europe, Africa even) at the same salaries because they have to.
I had lunch with a CEO of a US software company the other day. His biggest complaint? Finding competent engineers. He has the projects, the money, he's even got the empty desks. The hard bit is finding capable, disciplined engineers who can develop the products.
It is an extremely badly written article. Another example:
"Since tier one executives are "required" to be in touch due to the criticality of their professions, no one needs to be on 24/7. It's humanly impossible"
What does that mean? How is "no one needs to be on 24/7 a consequence of executives being required to do something?
I have a similar thing with my D-Link DSL-300 modem - the connection gradually slows down and then dies completely every three days or so and I have to restart it. The no-name modem it replaced (which wouldn't forward GRE) kept the same connection up for months. Anybody know any better ADSL modems?
There must be a template in Word for these kind of articles by now:
"$OLD_SKILL is not being used so much any more because of $NEW_TECHNOLOGY. Kids these days won't learn $PURPORTED_ADVANTAGE of the old ways. $RANT_ABOUT_WALKING_UPHILL_TO_SCHOOL_BOTH_WAYS_WHE N_A_KID."
Education time is finite, adding new skills necessarily means some old skills will be pushed out.
Yes, because Apple is going to throw away their main selling point, a decade of software development and a vastly superior OS in order to have to pay a licence fee to MS for every computer they sell.
Indeed, but I don't have to care about those people, I just buy the music. The idea of finding 6,000 people and getting them to agree up front what they want to pay for is laughable. The way it works at the moment is the standard business model: estimate what the customer wants, produce a product and try to sell it and make some profit on your investment.
The main point is, there is nothing stopping people implementing your system *right now*. Yet it hasn't happened. Instead, the market has spoken and I can walk into a music shop and buy any of thousands of high-quality recordings of good music for very litttle money.
OK, say in your system I want a recording of a piece of music on the scale of a Beethoven Symphony. So, I need (at a wild guess):
- a composer for (say) a year: $50,000, min - an orchestra, ~60 people for a couple of days: $30,000 - recording techs and equipment, studio hire: $10,000
So, now I have no ability to resell that recording, I need $90,000. Given the current system, I can get that recording for $15 or so, so now I need to find 6,000 other people with the same musical taste as me in order to match that price. Or I can just make do with whatever I already have, thereby removing the incentive for people to produce new music.
You are joking/trolling, right? Ignoring the freeloader problem for now, who the hell is going to bother doing that? That option is available right now, there is nothing stopping anyone doing it.
If you instigate that kind of system, you would end up in a few years with agencies that arrange new music for you at a price on the condition that you don't pass out copies to any one - so we'd be right back where we started.
Slashdot Mirror
When I rewrote a some image effects in KDE (blending, convert to B&W etc) using MMX/SSE a while back, the SIMD version was around 4x faster on an Athlon and 10x on a P4. Of course, that's only for one very small part of KDE and the overall performance increase was probably less than 1%.
Come on, I'm a mac user and I wouldn't compare the Mini's price to a Dell. Add a mouse, keyboard and monitor and you're up to $750 or so - for that money, you could get two entry-level Dells. Personally, I'd rather have the Mini, but that was the OP's point.
Doesn't that somewhat negate the point of buying a pre-built machine? Installing the software, tracking down drivers, running windows update etc take much longer than building the machine itself.
The "programmers are replacable" idea is absolutely crazy. Even ignoring the creativity argument, it certainly explains why most games are buggy pieces of crap. Experience in software is hugely valuable: a kid straight out of college may be able to write an algorithm, but far more important is the ability to write maintainable, testable, readable code.
The idea of a whole development team made up of people who write code like I did when I started out would scare the hell out of me.
Given those conditions, I'm amazed that there are any programmers willing to work in gaming at all even for a few years. I work in business software and, while we do have crunch weeks from time to time they are the exception and we always try to figure out afterwards what went wrong and how we can make sure it doesn't happen again.
Of course megapixels don't mean much without the optics to match. Compare these two images, both taken at 640x480. Spotting the photo taken on the Treo isn't hard...
There are lots of phones that will do that - just put a 1Gb MMC/memory stick in it.
The last conference I submitted to (CAPE2004) made it very clear that the use of Word was greatly preferred to LaTeX. They only provided a Word template and made lots of fuss about the PDF I submitted. Very annoying, but I think that might be the way things are headed.
I agree, it's a terrible website layout. I've emailed them a couple of times very politely pointing it out, but never even got an automated reply.
OO.org2 has got a lot better, but I still found it easier to fire up VMWare and use Office 2003, mostly for Excel and some Word. Office 2003 is quicker, has an easier interface, fewer bugs and of course better Office file format compatibility. As a word processor, I actually prefer Pages overall, although Office 12 looks pretty good.
The existence of OO is important, of course, as competition for MS, and it has some nice features (print to PDF etc). In my experience, OO2 is somewhere around "Office 96.5" - pretty good, but clunky and quirky with some odd bugs.
A decent office suite would be a start.
Can you point out some of the binary-only elements in the file format? I've yet to see any myself.
While that is possible, I haven't seen anything like it in the file format so far. You can see samples and the draft spec on the OpenXML website. From what I've seen, it should at least be easier to interoperate with Office12 XML than with the old binary formats. Not that that is saying all that much, of course.
Are there even many mobile devices without Arm cores of some kind these days? Almost every PDA and mobile phone seems to have an ARM7 or 9 core, which probably means that ARM processors outnumber x86 processors several times over.
That's inevitable, though. A computer scientist simply can't add enough value to make it worthwhile paying them so much money without moving into senior management.
Quite the reverse - they've had to fire incompetent engineers and a lot of their staff is there based on recommendations from senior engineers. They do employ a lot of people from abroad (Europe, Africa even) at the same salaries because they have to.
I had lunch with a CEO of a US software company the other day. His biggest complaint? Finding competent engineers. He has the projects, the money, he's even got the empty desks. The hard bit is finding capable, disciplined engineers who can develop the products.
It is an extremely badly written article. Another example:
"Since tier one executives are "required" to be in touch due to the criticality of their professions, no one needs to be on 24/7. It's humanly impossible"
What does that mean? How is "no one needs to be on 24/7 a consequence of executives being required to do something?
I guess he'll be suing Google video next - They have quite a few versions of the video on there...
I have a similar thing with my D-Link DSL-300 modem - the connection gradually slows down and then dies completely every three days or so and I have to restart it. The no-name modem it replaced (which wouldn't forward GRE) kept the same connection up for months. Anybody know any better ADSL modems?
There must be a template in Word for these kind of articles by now:
E N_A_KID."
"$OLD_SKILL is not being used so much any more because of $NEW_TECHNOLOGY. Kids these days won't learn $PURPORTED_ADVANTAGE of the old ways. $RANT_ABOUT_WALKING_UPHILL_TO_SCHOOL_BOTH_WAYS_WH
Education time is finite, adding new skills necessarily means some old skills will be pushed out.
Yes, because Apple is going to throw away their main selling point, a decade of software development and a vastly superior OS in order to have to pay a licence fee to MS for every computer they sell.
Indeed, but I don't have to care about those people, I just buy the music. The idea of finding 6,000 people and getting them to agree up front what they want to pay for is laughable. The way it works at the moment is the standard business model: estimate what the customer wants, produce a product and try to sell it and make some profit on your investment.
The main point is, there is nothing stopping people implementing your system *right now*. Yet it hasn't happened. Instead, the market has spoken and I can walk into a music shop and buy any of thousands of high-quality recordings of good music for very litttle money.
OK, say in your system I want a recording of a piece of music on the scale of a Beethoven Symphony. So, I need (at a wild guess):
- a composer for (say) a year: $50,000, min
- an orchestra, ~60 people for a couple of days: $30,000
- recording techs and equipment, studio hire: $10,000
So, now I have no ability to resell that recording, I need $90,000. Given the current system, I can get that recording for $15 or so, so now I need to find 6,000 other people with the same musical taste as me in order to match that price. Or I can just make do with whatever I already have, thereby removing the incentive for people to produce new music.
You are joking/trolling, right? Ignoring the freeloader problem for now, who the hell is going to bother doing that? That option is available right now, there is nothing stopping anyone doing it.
If you instigate that kind of system, you would end up in a few years with agencies that arrange new music for you at a price on the condition that you don't pass out copies to any one - so we'd be right back where we started.