Re:What's wrong with 802.11?
on
Bluetooth Bombs
·
· Score: 2
802.11 will never approach che low price of bluetooth; also, bluetooth uses far less power than does 802.11. It is great for wireless networking, but would be way overkill for just connecting your printer and computer, your phone and handset, for wireless mice and keyboards and so on. You would in effect have a $60 cable, and the mouse/keyboard/whatever would have to be connected to a power outlet or be recharged every two or three hours.
My current bank gives away nice crypto tokens ("calculators") with every electronic banking account.
Cost? about $150/account.
But with the 'calculator' thingie (my bank does not use this kind of authentication, BTW), people still need a way to get online, and this terminal is clearly intended to lure over all those people that won't get a computer for themselves.
Actually, for a bank, the cost of giving away terminals like that should be a fairly good deal. The cost of staffing the banks with tellers, security and other costs make it a huge win for every customer that chooses to do their routine businness via the web rather than in person. An estimate I've seen is that every bill paid in person will cost the bank about $2. With five bills a month ($10), the perminal will probably have paid for itself after 2-3 years. Considering that somebody who has switched to internetbanking probably will continue to use it, a one time cost per customer of $300 or so is probably a good bargain.
The kind of programmer you want is probably already fairly interested in science in general, but (as you say) would choose CS or mathematics as their field of study. I believe that what you need to do is to use your own field (whatever it is) as a strength, and offer the applicants the possibility to learn your field from a computational perspective even as they are working for you as programmers. Depending on your field and organization (corporation, university), this might mean things like assigning a mentor or adviser to the applicant, allowing the applicant to co-write papers, having a bit of free time for projects of their own, etc.
The additional upside to this approach (apart from actually finding a qualified employee) is that after a couple of years you will have a highly competent person, conversant both in your own field as well as in all aspects of computation surrounding it -- in house and already familiar with your work. That is the kind of competence that can't be bought.
But none of this matters. Certificates aren't meant to be valuable or signify elite status. They're meant as identification. "Just anybody" should have a certificate.
And I agree. The problem lies in that the certificate has to be trusted. As a number of posters have stated, nothing prevents you from giving out your own certificates. Nobody would (rightly) trust it, but at least you have one. And that impression of trust is influenced by the points I wrote about above.
Sure, use a certificate from 'Johnnys tattoo, piercings and certificate authority', but wouldn't you agree that it would be overly trusting to take that certificate at face value? People that impression of value and solidity.
As a prevoius poster stated, if you or me gave out certificates for free, the value of the certificate fould be seen as low. This is in part because I'm not perceived as well known, stable, trustworthy entity (just ask my friends:) ). The other reason for distrust would be for the very fact that it is given for free; in other words, it's the same kind of effect taking place when people perceive linux of xBSD as being worse than Winxx because it doesn't cost anything.
Let's say a large, trusted entity (like IBM, or a large bank) would start giving out certificates for little or no money. Many would be concerned that they won't do a good background check, or be vigilant for abuses of certificates when the customers aren't paying for it. Also, the customers would get the sense that 'just anybody' could get hold of a certificate, which would seem to cheapen its value in itself. Basic psychology, really (and this is a large part of the reason why a Weight Watcher program works - at least temporarily - whereas magazine diets don't).
I just checked, and rambus is down to $26, with a fall of 26% as of this writing. Seems that investors saw those patent claims as a major part of the value of the company. Or else it's just the stock market being its normal, irrational self...
I appreciate the fact that one is supposed to work out their own solution to problem posed in class or on assignments. However, there is a form of academic cheating that is rampant throughout the university world.
A little more than a year ago, the pages of Science where alive with the debate over whether a certain professor had published erroneous results or not. The case was as follows: his graduate student had been doing a series of experiments, and published a couple of papers on this -- with the professor as the first author. Some of you may not know this, but the tradition is that the person responsible for most of the work is cited first, followed by others, and, sometimes, followed by the professor (or other 'grey eminence') last.
In this case, it turned out that the graduate student had faked his data to fit with his theory. The professor -- who had been the first name on the paper -- excused himself by saying that he had not even read the paper in question.
So, either he did read the paper, examined the results, and published anyway, and is guilty of fraud; or he had his name as the first author of a paper he didn't even read until it blew up in his face.
This _really_ makes you wonder about scientinsts that have hundreds of papers to their name -- papers that are really the basis of their careers...
I gotthe stat from Browser News, but that stat is based on the TheCounter stat. Nevertheless, there's a good deal of other interesting statistics here, like support for various plugins and such.
Indeed, newer HTML versions are better for accessibility purposes than older versions.
Well, yes and no... People with various disabilities often have to have special versions of their software to function. This is (regrettably) often far down on the to-do list for manufacturers, and far down on the budget list for caregivers.
On a side note, for people with poor eyesight, new standards too often mean they can't see the site at all, as the designers are determined to showcase their talent, rather than making a usable site. If a user has chosen a particular typeface and size, it is often for a good reason... Too often you see sites overriding those choices, rendering (sic) the site unusable.
Some users (like vision impaired -- or people using small handhelds) need to be able to get text only content. There are really no excuses not to provide data in a text-only format.
Second, JavaScript is a _bad_ idea. A quick check reveals that the percentage of users not using Javascript at all was 20% in 2000, up from 14% in 1999. This is of course due to pop-ups and to the irritating habit of overriding user preferences that we all know and love, but also because it is more and more common for companies to filter out javascript at their firewalls.
I understand the reasoning behind their concerns, but as a practical matter, many web sites do _not_ wish to alienate more users than they have to (though some obviously does not understand this).
The thing is that free/open software plays by the rules. There is no stealing of intellectual property and nothing else illegal going on at all. open/free software is even copyrighted and licenced to its users, just like MS' products are. The only difference (from a legal standpoint) is the distribution method (source) and the terms of the licenses (you don't need to pay, but you have to share).
As far as I am aware, there is only one way that giving stuff away could pose problems, and that is if a market player dumps stuff at below cost in order to squeeze out their competitors. This should not apply in our case, as 'free' is the normal price, not a limited offer, and the cost is (close to) zero, as the vast majority writes the stuff as a hobby, on their free time. Even those companies employing programmers to work on open source should be off the hook as they legitimately can say that they are improving the software for internal use, then sharing the improvements as per the license; or that they are in the service business, not the software business.
I get the impression that MS has current, accurate figures about linux adoption and are getting scared. It'll be interesting to see when usage figures are next published...
Disclaim: I'm a CompSci grad, but I haven't done a real analysis on this or anything. Be warned: unsubstantiated opinions ahead...
Gnutella cna never scale in the sense that napster does. If you search for a given file, you won't find every occurence of it on the network (in fact, you probably won't find most occurrences). If too many clients try to connect, the network will break down, at random points, and split into subnetworks. This is, as I understand it, not as big a problem as it might seem at first, though.
The thing is, you don't want to find every instance of a given file, you just want to find one, and it doesn't matter which one. Say that the gnutella network has split into ten major subnetworks, where communication between them has broken down (the networks generate so much traffic that any server attempting to connect to both will be swamped, foir instance). If a given file exists in only one copy, you have only one chance in ten to find it. If, also, you are the only one to want it, you are (almost) out of luck.
Now, the gnutella network isn't static. People tend to start their clients, then stop them again at semirandom intervals (they turn off the computer for the night, software crashes, they want to limit bandwidth (for a game of Quake or whatever) and so on). When they reconnect, there is a nonzero chance that they will end up on a different subnet (and I believe that the chance increases if they use an automagic host catcher thingy). So if you want that special file, you might want to try now and again for some days, and get lucky. Also, if people tend to share the stuff they themselves download (and this chance probably increases as bandwidth and storage capacity increases), popular files will get spread over the network, making it steadily more probable that you will find the file in your neighbourhood.
The thing is, gnutella _will_ scale, it just won't scale linearly over the number of hosts. Now, this is just my conclusion; any other takers on this?
I (largely) agree with you; I merely responded to the prevous posters suggestion that the lack of friction would prevent a wing from generating lift at all.
As for the turbulent boundary, I seem to remember that this effect is independent of the friction against the surface, and that the best way to combat this would be managing the microstructure of the surface; making it bumby (in a controlled way), rather than getting it as smooth as possible. Now, IANAE, so please take this with a sack or two of salt.
It's not really due to friction; it's rather that the air on the upper side has longer to travel over the wing than the lower side. OTOH, this material won't help with vortex formation, which is (if I remember correctly) the largest part of the drag. OTTH, you need drag - not for generating lift, but for the craft to be stable.
I'm confused. I'm not christian, and celebrate the midwinter solstice instead -- should I leave my computer or stay online?
Please, I'm really happy for you, as this is an important holiday for you, but why not accept that in an international forum there will be plenty of people not sharing your set of beliefs?
As great as it would be to incorporate great new techniques into linux and thereby spread the use of linux among people, that should not be the end-all of the interaction. Instead, that should be the result (or byproduct) of an increasing dialogue between linux developers and IBMers.
Linus et.al. has many times indicated an interest in expanding the playground for linux into 'big iron' as well as embedded systems. Why not listen to IBM, and aske them what their customers would find lacking in linux as it stands? By becoming more informed about a market of which most of us really know little about, better choices can be made, old mistakes can be avoided, reinventing of wheels will not take place, etc.
That is not to say that linux developers should just lie down and think of Finland when IBM comes knocking; the goals for linux are at least partially different than the goals for a commercial system, and the development process tends to be very different. That would be lessons IBM could do well in learning from a fruitful dialogue (and they seem to have picked it up fairly well already).
New filesystems, failover capabilities etc. would then fall out fairly naturally.
Ps. open sourcing the OS/2 drivers would make for a nice midwinter gift:) Ds.
It makes NO sense to use a keybaord that slows your typing down, regardless of ergonomics or the "coolness" factor
...until you actually get RSI. It is painful, it is awkward (try picking stuff up with your wrists immobilized) and, believe me, it slows your typing down much more than those keyboards will.
Actually, I find 'ergonomic' keyboards (those I've tried) to be no better in themselves at reducing RSI. The problem is really not the position, but the fact that you do the same thing over and over. What does help is varying your position; switching between two different keyboards, switching between sitting and standing (you need a way to raise the keyboard and monitor), taking frequent brakes, and avoiding the mouse.
Besides, another genneration or two of Dragon/Viavoice, and we won't need a keyboard
Hmmm... talking to your computer all day. That's a recipe for laryngitis if Iv'e ever seen one. And I can imagine the noise level at a bysy department with fifty people dictating to their WP. Or an airplane full with people working. I think voice control is not nearly as good an idea as it seems at first.
I've found that compiling mozilla yourself, omitting all the debugging stuff, the speed increases tremendously. As for stability, there are _huge_ differences depending on exactly what nightly build (or cvs timestamp) you use; it can go from unusable to "why don't they just release this version?" in a day (and back).
Your question is well placed; is computer geegdom really reliably connected to scifi/fantasy, and is there a connection between the particular flavour of Geek and of the flavour of fantasy?
For my part, yes, I was more than a little penalized in school (I ended up in the emergency ward more than once in 7:th to 9:th grade); at the same time I really loved the Tolkien stories (today I find them too one-dimensional for my taste).
I am a Unix person, finding Tolkien rather aged, and Pratchett as the Great New Thing in fantasy (with a heatlthy side order of Laurell Hamilton)... As my friends are a very eclectic bunch of people, I don't find any OS preference to correlate with litterary inclinations, rather it is between people continuing to their doctorates versus people choosing to go directly towards a career that is the defining item.
Well, look at the 3d hardware business from the manufacturing angle: a state-of-the-art 3d chip is as complicated (if not more complicated) as a general-purpose CPU, with similar die-sizes and transistor counts. Unlike Intel and AMD, however, they can not sell the newest stuff at a hefty premium, instead they sell the chip complete with memory, bus logic and ports -- cheaper than a comparative general processor. Then add the requirement to get something entirely new (not just a clock speed tweak) out the door every six to nine months... No wonder that companies are falling faster than the Microsoft stock price.
On the other hand, the graphics companies are not encumbered with keeping up with legacy systems -- their stuff isn't really programmable from the user level, so all they need is a new set of drivers to support their new hardware. This should make for much speedier development.
It would be interesting -- but not entierly surprising -- if graphics hardware makers would start driving the IC business the same way that games are driving software today.
The question is of course if Nvidia will take a page from 3dfx:s book and further open their own drivers, or if this is the end of the (relative) openness from 3dfx. Then again, it's quite possible that they will continue to support the open effoert of current 3dfx cards, while continuing to produce closed drivers for their core Nvidia line.
As Nvidia has claimed that a big reason not to open their drivers is that they are forbidden to do so because of NDA:s with technology partners, one possibility is that as they now own 3dfx technology can use that in place of (probably quite expensive) 3rd party stuff. That would mean more the possibility of increasing the openness towards developers.
Slashdot Mirror
802.11 will never approach che low price of bluetooth; also, bluetooth uses far less power than does 802.11. It is great for wireless networking, but would be way overkill for just connecting your printer and computer, your phone and handset, for wireless mice and keyboards and so on. You would in effect have a $60 cable, and the mouse/keyboard/whatever would have to be connected to a power outlet or be recharged every two or three hours.
/Janne
My current bank gives away nice crypto tokens ("calculators") with every electronic banking account.
Cost? about $150/account.
But with the 'calculator' thingie (my bank does not use this kind of authentication, BTW), people still need a way to get online, and this terminal is clearly intended to lure over all those people that won't get a computer for themselves.
/Janne
Actually, for a bank, the cost of giving away terminals like that should be a fairly good deal. The cost of staffing the banks with tellers, security and other costs make it a huge win for every customer that chooses to do their routine businness via the web rather than in person. An estimate I've seen is that every bill paid in person will cost the bank about $2. With five bills a month ($10), the perminal will probably have paid for itself after 2-3 years. Considering that somebody who has switched to internetbanking probably will continue to use it, a one time cost per customer of $300 or so is probably a good bargain.
/Janne
The kind of programmer you want is probably already fairly interested in science in general, but (as you say) would choose CS or mathematics as their field of study. I believe that what you need to do is to use your own field (whatever it is) as a strength, and offer the applicants the possibility to learn your field from a computational perspective even as they are working for you as programmers. Depending on your field and organization (corporation, university), this might mean things like assigning a mentor or adviser to the applicant, allowing the applicant to co-write papers, having a bit of free time for projects of their own, etc.
The additional upside to this approach (apart from actually finding a qualified employee) is that after a couple of years you will have a highly competent person, conversant both in your own field as well as in all aspects of computation surrounding it -- in house and already familiar with your work. That is the kind of competence that can't be bought.
/Janne
But none of this matters. Certificates aren't meant to be valuable or signify elite status. They're meant as identification. "Just anybody" should have a certificate.
And I agree. The problem lies in that the certificate has to be trusted. As a number of posters have stated, nothing prevents you from giving out your own certificates. Nobody would (rightly) trust it, but at least you have one. And that impression of trust is influenced by the points I wrote about above.
Sure, use a certificate from 'Johnnys tattoo, piercings and certificate authority', but wouldn't you agree that it would be overly trusting to take that certificate at face value? People that impression of value and solidity.
As a prevoius poster stated, if you or me gave out certificates for free, the value of the certificate fould be seen as low. This is in part because I'm not perceived as well known, stable, trustworthy entity (just ask my friends:) ). The other reason for distrust would be for the very fact that it is given for free; in other words, it's the same kind of effect taking place when people perceive linux of xBSD as being worse than Winxx because it doesn't cost anything.
Let's say a large, trusted entity (like IBM, or a large bank) would start giving out certificates for little or no money. Many would be concerned that they won't do a good background check, or be vigilant for abuses of certificates when the customers aren't paying for it. Also, the customers would get the sense that 'just anybody' could get hold of a certificate, which would seem to cheapen its value in itself. Basic psychology, really (and this is a large part of the reason why a Weight Watcher program works - at least temporarily - whereas magazine diets don't).
I just checked, and rambus is down to $26, with a fall of 26% as of this writing. Seems that investors saw those patent claims as a major part of the value of the company. Or else it's just the stock market being its normal, irrational self...
I appreciate the fact that one is supposed to work out their own solution to problem posed in class or on assignments. However, there is a form of academic cheating that is rampant throughout the university world.
A little more than a year ago, the pages of Science where alive with the debate over whether a certain professor had published erroneous results or not. The case was as follows: his graduate student had been doing a series of experiments, and published a couple of papers on this -- with the professor as the first author. Some of you may not know this, but the tradition is that the person responsible for most of the work is cited first, followed by others, and, sometimes, followed by the professor (or other 'grey eminence') last.
In this case, it turned out that the graduate student had faked his data to fit with his theory. The professor -- who had been the first name on the paper -- excused himself by saying that he had not even read the paper in question.
So, either he did read the paper, examined the results, and published anyway, and is guilty of fraud; or he had his name as the first author of a paper he didn't even read until it blew up in his face.
This _really_ makes you wonder about scientinsts that have hundreds of papers to their name -- papers that are really the basis of their careers...
I gotthe stat from Browser News, but that stat is based on the TheCounter stat. Nevertheless, there's a good deal of other interesting statistics here, like support for various plugins and such.
Indeed, newer HTML versions are better for accessibility purposes than older versions.
Well, yes and no... People with various disabilities often have to have special versions of their software to function. This is (regrettably) often far down on the to-do list for manufacturers, and far down on the budget list for caregivers.
On a side note, for people with poor eyesight, new standards too often mean they can't see the site at all, as the designers are determined to showcase their talent, rather than making a usable site. If a user has chosen a particular typeface and size, it is often for a good reason... Too often you see sites overriding those choices, rendering (sic) the site unusable.
/Janne
Some users (like vision impaired -- or people using small handhelds) need to be able to get text only content. There are really no excuses not to provide data in a text-only format.
Second, JavaScript is a _bad_ idea. A quick check reveals that the percentage of users not using Javascript at all was 20% in 2000, up from 14% in 1999. This is of course due to pop-ups and to the irritating habit of overriding user preferences that we all know and love, but also because it is more and more common for companies to filter out javascript at their firewalls.
I understand the reasoning behind their concerns, but as a practical matter, many web sites do _not_ wish to alienate more users than they have to (though some obviously does not understand this).
/Janne
2001-03-18 21:08:33
Why? Because I said so!
The thing is that free/open software plays by the rules. There is no stealing of intellectual property and nothing else illegal going on at all. open/free software is even copyrighted and licenced to its users, just like MS' products are. The only difference (from a legal standpoint) is the distribution method (source) and the terms of the licenses (you don't need to pay, but you have to share).
As far as I am aware, there is only one way that giving stuff away could pose problems, and that is if a market player dumps stuff at below cost in order to squeeze out their competitors. This should not apply in our case, as 'free' is the normal price, not a limited offer, and the cost is (close to) zero, as the vast majority writes the stuff as a hobby, on their free time. Even those companies employing programmers to work on open source should be off the hook as they legitimately can say that they are improving the software for internal use, then sharing the improvements as per the license; or that they are in the service business, not the software business.
I get the impression that MS has current, accurate figures about linux adoption and are getting scared. It'll be interesting to see when usage figures are next published...
Disclaim: I'm a CompSci grad, but I haven't done a real analysis on this or anything. Be warned: unsubstantiated opinions ahead...
Gnutella cna never scale in the sense that napster does. If you search for a given file, you won't find every occurence of it on the network (in fact, you probably won't find most occurrences). If too many clients try to connect, the network will break down, at random points, and split into subnetworks. This is, as I understand it, not as big a problem as it might seem at first, though.
The thing is, you don't want to find every instance of a given file, you just want to find one, and it doesn't matter which one. Say that the gnutella network has split into ten major subnetworks, where communication between them has broken down (the networks generate so much traffic that any server attempting to connect to both will be swamped, foir instance). If a given file exists in only one copy, you have only one chance in ten to find it. If, also, you are the only one to want it, you are (almost) out of luck.
Now, the gnutella network isn't static. People tend to start their clients, then stop them again at semirandom intervals (they turn off the computer for the night, software crashes, they want to limit bandwidth (for a game of Quake or whatever) and so on). When they reconnect, there is a nonzero chance that they will end up on a different subnet (and I believe that the chance increases if they use an automagic host catcher thingy). So if you want that special file, you might want to try now and again for some days, and get lucky. Also, if people tend to share the stuff they themselves download (and this chance probably increases as bandwidth and storage capacity increases), popular files will get spread over the network, making it steadily more probable that you will find the file in your neighbourhood.
The thing is, gnutella _will_ scale, it just won't scale linearly over the number of hosts. Now, this is just my conclusion; any other takers on this?
Some time ago, this text circulated arount the net. It was funny then - how times change...
I (largely) agree with you; I merely responded to the prevous posters suggestion that the lack of friction would prevent a wing from generating lift at all.
As for the turbulent boundary, I seem to remember that this effect is independent of the friction against the surface, and that the best way to combat this would be managing the microstructure of the surface; making it bumby (in a controlled way), rather than getting it as smooth as possible. Now, IANAE, so please take this with a sack or two of salt.
It's not really due to friction; it's rather that the air on the upper side has longer to travel over the wing than the lower side. OTOH, this material won't help with vortex formation, which is (if I remember correctly) the largest part of the drag. OTTH, you need drag - not for generating lift, but for the craft to be stable.
/Janne
If the CTO should report to the CIO, but instead needs to talk to the COO, that would have to be OK:d by the CEO, or somebody made a BOO-BOO...
Sorry.
I'm confused. I'm not christian, and celebrate the midwinter solstice instead -- should I leave my computer or stay online?
Please, I'm really happy for you, as this is an important holiday for you, but why not accept that in an international forum there will be plenty of people not sharing your set of beliefs?
As great as it would be to incorporate great new techniques into linux and thereby spread the use of linux among people, that should not be the end-all of the interaction. Instead, that should be the result (or byproduct) of an increasing dialogue between linux developers and IBMers.
:) Ds.
Linus et.al. has many times indicated an interest in expanding the playground for linux into 'big iron' as well as embedded systems. Why not listen to IBM, and aske them what their customers would find lacking in linux as it stands? By becoming more informed about a market of which most of us really know little about, better choices can be made, old mistakes can be avoided, reinventing of wheels will not take place, etc.
That is not to say that linux developers should just lie down and think of Finland when IBM comes knocking; the goals for linux are at least partially different than the goals for a commercial system, and the development process tends to be very different. That would be lessons IBM could do well in learning from a fruitful dialogue (and they seem to have picked it up fairly well already).
New filesystems, failover capabilities etc. would then fall out fairly naturally.
Ps. open sourcing the OS/2 drivers would make for a nice midwinter gift
It makes NO sense to use a keybaord that slows your typing down, regardless of ergonomics or the "coolness" factor
...until you actually get RSI. It is painful, it is awkward (try picking stuff up with your wrists immobilized) and, believe me, it slows your typing down much more than those keyboards will.
Actually, I find 'ergonomic' keyboards (those I've tried) to be no better in themselves at reducing RSI. The problem is really not the position, but the fact that you do the same thing over and over. What does help is varying your position; switching between two different keyboards, switching between sitting and standing (you need a way to raise the keyboard and monitor), taking frequent brakes, and avoiding the mouse.
Besides, another genneration or two of Dragon/Viavoice, and we won't need a keyboard
Hmmm... talking to your computer all day. That's a recipe for laryngitis if Iv'e ever seen one. And I can imagine the noise level at a bysy department with fifty people dictating to their WP. Or an airplane full with people working. I think voice control is not nearly as good an idea as it seems at first.
I've found that compiling mozilla yourself, omitting all the debugging stuff, the speed increases tremendously. As for stability, there are _huge_ differences depending on exactly what nightly build (or cvs timestamp) you use; it can go from unusable to "why don't they just release this version?" in a day (and back).
Your question is well placed; is computer geegdom really reliably connected to scifi/fantasy, and is there a connection between the particular flavour of Geek and of the flavour of fantasy?
For my part, yes, I was more than a little penalized in school (I ended up in the emergency ward more than once in 7:th to 9:th grade); at the same time I really loved the Tolkien stories (today I find them too one-dimensional for my taste).
I am a Unix person, finding Tolkien rather aged, and Pratchett as the Great New Thing in fantasy (with a heatlthy side order of Laurell Hamilton)... As my friends are a very eclectic bunch of people, I don't find any OS preference to correlate with litterary inclinations, rather it is between people continuing to their doctorates versus people choosing to go directly towards a career that is the defining item.
Well, look at the 3d hardware business from the manufacturing angle: a state-of-the-art 3d chip is as complicated (if not more complicated) as a general-purpose CPU, with similar die-sizes and transistor counts. Unlike Intel and AMD, however, they can not sell the newest stuff at a hefty premium, instead they sell the chip complete with memory, bus logic and ports -- cheaper than a comparative general processor. Then add the requirement to get something entirely new (not just a clock speed tweak) out the door every six to nine months... No wonder that companies are falling faster than the Microsoft stock price.
On the other hand, the graphics companies are not encumbered with keeping up with legacy systems -- their stuff isn't really programmable from the user level, so all they need is a new set of drivers to support their new hardware. This should make for much speedier development.
It would be interesting -- but not entierly surprising -- if graphics hardware makers would start driving the IC business the same way that games are driving software today.
The question is of course if Nvidia will take a page from 3dfx:s book and further open their own drivers, or if this is the end of the (relative) openness from 3dfx. Then again, it's quite possible that they will continue to support the open effoert of current 3dfx cards, while continuing to produce closed drivers for their core Nvidia line.
As Nvidia has claimed that a big reason not to open their drivers is that they are forbidden to do so because of NDA:s with technology partners, one possibility is that as they now own 3dfx technology can use that in place of (probably quite expensive) 3rd party stuff. That would mean more the possibility of increasing the openness towards developers.