Well, let's wait a while... if it makes them unpopular, they'll stop it for sure. They can't afford losing their supporters. I guess they're just experimenting on a new idea, just like sending spam is. I personally never received their spam yet, but some people actually welcomed it. Unsure about the pop-up ads though.
Well, if this person decided to publish the record on the web and do Google bombing to crank up the search on certain keywords, it would come worse than your gossipy old cow....
For the uninformed like I was, here's X-Prize's webpage. The news is summed up nicely in the following paragraph:
Hegler said Cape Canaveral was the first choice, even though the Kennedy Space Center is not directly involved, and Cecil Field in Jacksonville is an alternative location.
But remember that Microsoft also takes an important role on enabling joe6pack to use computers. Also, Microsoft research has contributed quite a lot to scientific community...
Back then in June 1990 (as the date of the letter), Microsoft wasn't a monopoly yet, right? So, the anti-trust trial cannot use this as an evidence against them....
I would say that this may lead to anti-competitive lawsuit... (btw, is such lawsuit allowable in the USA?) And of course, as usual, IANAL...
I agree. Most well-respected meetings have an accept rate between 10-20%. Some actually have a bit lenient threshold (like GECCO), which is pretty close to 50% in the last 2 years. These people usually have 2-3 volumes of proceedings...
First, most industrial conference fees are not in the thousands of dollars.
Hmm, I think that's not the case. Most of industrial conference flyers that came to my door cost at least $1000. Compare that to academic ones, which are in $200-300 range.
Second, it is rather unusual for employees to pay their own way to conferences
Maybe. Many people I talk with usually don't have this luxury. They usually get paid day-offs and sometimes travel allowances. Some big companies are actually paying their employees to attend.
Third, research is by no means all or nothing. Most of it is incremental improvememnt on existing science, and gives a corresponding return on investment.
That's your point of view. In business point of view: If it doesn't generate stuffs useful to the corporations, it is useless. Will you guarantee that those incremental results will back up company's ROI? When? 25 years? 30 years?
Fourth, you accuse companies of "stealing" ideas at conferences. Well that's the whole idea, ding dong.
No, I'm not accusing them like that. What I said is that it's better for them to "steal" the ideas in the conferences rather than hiring the researchers in the first place.
In that community at least, the top researchers are about evenly divided between industry and academic, and no, the industrial researchers are NOT mostly ex-professors.
Really? I've attended various CS, AI, and Bioinf conferences. These areas are predominantly academias, esp in CS. In Bioinf, the ratio is somewhat lessened. Usually top researchers are either ex-profs or maybe plus their ex-students.
I suspect most academics are jealous of industrial researchers because they often have better financial backing and are involved in more "real world" problems. I also think industrial researchers are jealous of academics because they have more time and freedom to tackle basic, pure research.
This is bullsh!t. Many companies have elected to fund research projects in universities. They gave grants of millions of dollars and we solve "real world" problems too. This is reality. Sometimes academias are reluctant to accept such project because of the NDA stuff -- they can't take the great pride of publishing the results.
Also, in industry, people are actually have MORE time than the academia. You are so wrong! Why? It's because in academia, we still need to worry about courses and teaching duties. Most industrial people I've talked with usually do NOT concern about the time. In fact, their whole time can be dedicated to research, which is VERY nice. They are more concerned about bureaucracy -- the NDA and so forth.
So far as your assessment of the quality of conference papers from industry is concerned, it's just complete garbage.
No!!! I've been reviewing papers in top conferences as a committee member in the said areas. If you understand how the papers are shown in a particular journal or proceedings, they are peer-reviewed first by 3-4 people. Industrial papers are easy to spot: They're usually full of buzzwords and tied into a specific project. The content is very small: Most of which we already knew. It's like rehashing old ideas and apply it to some new problems. That's the issue. I usually rate such papers 1 or 2 out of 5 points possible and mostly recommended to reject such papers. The central committee will review my review and my peers' reviews and take the decision based on them. The central committee usually agreed to my decision, so I guess that other reviewers would have voiced the same concerns. Some industrial papers are indeed excellent, in which case I won't hesitate to give 5 points. People from Microsoft Research, IBM TJ Watson, and Intels are known of publishing good papers. However, lots of their papers got rejected too.
but single-digit percentage increases in performance isn't "dramatic". It's more like "scarcely noticeable".
In that case, maybe... But if you follow some compiler conference papers, single digit percentage of improvement *is* a dramatic improvement.
More than that single digit, we need to either change the underlying algorithm, or do a more dramatic overhaul, or correct a resource hogging mistakes. Well, we all know that Mozilla coders aren't that sloppy, so I guess that single digit improvements are really good because they usually involve quite a lot of cutting corners squeezing out more improvements over the already tight code.
You're saying... extra tits/second as an improvement unit? My mind is distorted....;)
Researchers are Paying Their Own Way
on
AT&T Labs' Brain Drain
·
· Score: 3, Informative
Well, researchers are often paying their own way except if they are one of the chairs, which they are offered some complimentary registration fee. Some of bonafide conferences actually pay their expenses if they are invited.
Honestly, researcher communities (especially the academic ones) are disdainful to the "achievements" of "industrial research". The reject rates on industrial papers have been pretty high (usually more than 50%). This is because that the "innovations" of industrial "research" are more or less either one or some of the following: rehashing old ideas, implementing old ideas with new looks / new aspects / into new problems which often not worth mentioning, combining several old ideas in some obvious ways.
Well, this is not to say that industrial papers are crap, but of course there are some excellent industry researchers, which are usually ex-professors which are already well known before they enter the industries. However, research is like a big gamble: either you win big or you lose big. Given the current situation of the economy, it's more likely you lose big because of "lack of genuinely new ideas" and you can never get a guarantee that your research group is actually producing the great useful results for your company. It's a whole lot better for the company to actually scour the conferences, spot the prominent person with the right ideas, and then "steal" them so that they can implement the said idea for your company. This is exactly what Microsoft has been doing in the past years.
Since I never attended trade/business oriented conferences, I can't comment on those. Moreover, these conferences are usually way more expensive than the academia ones (thousands of $$ vs hundreds of $$).
I remember speaking to one of the developers in the IRC channel specifically about this. Their response was "write a gnome-vfs module for it."
Well, that's typical OSS developers...:) I'll bet that the iPod support is because someone actually came to the IRC channel and flame "Rhythmbox sux because it doesn't support iPod".;)
Well, having the trend of people adding TLDs like nobody's business, we would come into the same problem albeit a lot slower. Then, someone has to invent some kind of smarter algorithm for this kind of problem. This is an inherent problem needs to be solved. Just watch for the next communication / network conferences...:)
If that's the case, then... either the strongest wins (mucho dinero), or the first come first served rules. As simple as that. This would happen exactly the same even with the TLD intact.
You can see their howto pages to see it in action. You probably want to check out some sample outputs. And this project also uses LilyPond. Check that out.
And you can take a conclusion: Aunt Tillie's intelligence >> PHB's intelligence. And... in an RPG world, the same would apply for strength, dexterity, wisdom, and charisma. Oh yes, it takes all these to get the CLI working...;)
Re:Open cool, but still keep distribution rights.
on
Beyond An Open Source Java
·
· Score: 2, Informative
X is designed around 1984 but available widely 1987 and NeWS is also around 1987. By then they are not well known and relatively had the same number of adherents. However, since X is open, X is embraced by many in a short time and thus X won. It's not just "open source and then win" idea like ESR suggested, but look at the situation closely.
Microsoft is doing good with C#, but they're open source because they're hoping that they get a quick momentum and get a new adherents shortly to contend Java, which has already been well-established, not just "weakening" per se.
Since Java has already had lots of adherents, opening up won't add some advantage. If later C# is gaining ground (which I highly doubt), then Sun may consider that (which, again, I highly doubt even such case happen except that C# starts overtaking Java).
No, my point is, to play catch-up, the acceleration must be boosted much more than the current level, which IMHO is not fast enoough to even play catch-up.
Swing and AWT has been planned way before and there are developments IIRC 3 years ago and only now it produces at least something real. Back then it's still Java 1.3 and they tried to get full Java 1.2 compatibility... until now. Of course there are parts of JDK 1.3 and 1.4 that have been added as well, but look at the status page and get a picture of how dire the situation is. I'm not saying that GCJ developers are not hardworking enough, though.
Of course I'd like to see many addition that GCJ offer such as native compilation and others, but until it's truly usable fast enough, the relative progress seems to be a downward spiral and thus making the great-grand parent post ("give GCJ hippie" thing) a moot.
If you have been following gcj's progress recently it is maturing rapidly, just give those dirty gnu hippies a few more years they'll be on par with Sun's own implimentation.
Two facts:
GCJ still far behind Sun's Javac. Lookie, they just "invent" ah so Java 1.2 Swing!
GCJ's progress is less than Sun's progress. See Java 1.5 maturing. I'm wondering when GCJ will catch up?
This lead me to conclude that GCJ would never catch Sun's level unless something drastic changes happen. Don't get me wrong, I like GCJ's idea and try to support it, but I firmly believe that they need to change to boost more progress.
Slashdot Mirror
Well, let's wait a while... if it makes them unpopular, they'll stop it for sure. They can't afford losing their supporters. I guess they're just experimenting on a new idea, just like sending spam is. I personally never received their spam yet, but some people actually welcomed it. Unsure about the pop-up ads though.
That must be a powerful ball! Whose ball is it? ;)
Well, if this person decided to publish the record on the web and do Google bombing to crank up the search on certain keywords, it would come worse than your gossipy old cow....
Check this thread for an interesting discussion. This is just posted around today...
For the uninformed like I was, here's X-Prize's webpage. The news is summed up nicely in the following paragraph:
Hegler said Cape Canaveral was the first choice, even though the Kennedy Space Center is not directly involved, and Cecil Field in Jacksonville is an alternative location.
Actually, Ballmer regretted that because if MS were the king of search engine, he would've been able to filter a search like this.
Ah, yes...
But remember that Microsoft also takes an important role on enabling joe6pack to use computers. Also, Microsoft research has contributed quite a lot to scientific community...
Back then in June 1990 (as the date of the letter), Microsoft wasn't a monopoly yet, right? So, the anti-trust trial cannot use this as an evidence against them....
I would say that this may lead to anti-competitive lawsuit... (btw, is such lawsuit allowable in the USA?) And of course, as usual, IANAL...
I agree. Most well-respected meetings have an accept rate between 10-20%. Some actually have a bit lenient threshold (like GECCO), which is pretty close to 50% in the last 2 years. These people usually have 2-3 volumes of proceedings...
First, most industrial conference fees are not in the thousands of dollars.
Hmm, I think that's not the case. Most of industrial conference flyers that came to my door cost at least $1000. Compare that to academic ones, which are in $200-300 range.
Second, it is rather unusual for employees to pay their own way to conferences
Maybe. Many people I talk with usually don't have this luxury. They usually get paid day-offs and sometimes travel allowances. Some big companies are actually paying their employees to attend.
Third, research is by no means all or nothing. Most of it is incremental improvememnt on existing science, and gives a corresponding return on investment.
That's your point of view. In business point of view: If it doesn't generate stuffs useful to the corporations, it is useless. Will you guarantee that those incremental results will back up company's ROI? When? 25 years? 30 years?
Fourth, you accuse companies of "stealing" ideas at conferences. Well that's the whole idea, ding dong.
No, I'm not accusing them like that. What I said is that it's better for them to "steal" the ideas in the conferences rather than hiring the researchers in the first place.
In that community at least, the top researchers are about evenly divided between industry and academic, and no, the industrial researchers are NOT mostly ex-professors.
Really? I've attended various CS, AI, and Bioinf conferences. These areas are predominantly academias, esp in CS. In Bioinf, the ratio is somewhat lessened. Usually top researchers are either ex-profs or maybe plus their ex-students.
I suspect most academics are jealous of industrial researchers because they often have better financial backing and are involved in more "real world" problems. I also think industrial researchers are jealous of academics because they have more time and freedom to tackle basic, pure research.
This is bullsh!t. Many companies have elected to fund research projects in universities. They gave grants of millions of dollars and we solve "real world" problems too. This is reality. Sometimes academias are reluctant to accept such project because of the NDA stuff -- they can't take the great pride of publishing the results.
Also, in industry, people are actually have MORE time than the academia. You are so wrong! Why? It's because in academia, we still need to worry about courses and teaching duties. Most industrial people I've talked with usually do NOT concern about the time. In fact, their whole time can be dedicated to research, which is VERY nice. They are more concerned about bureaucracy -- the NDA and so forth.
So far as your assessment of the quality of conference papers from industry is concerned, it's just complete garbage.
No!!! I've been reviewing papers in top conferences as a committee member in the said areas. If you understand how the papers are shown in a particular journal or proceedings, they are peer-reviewed first by 3-4 people. Industrial papers are easy to spot: They're usually full of buzzwords and tied into a specific project. The content is very small: Most of which we already knew. It's like rehashing old ideas and apply it to some new problems. That's the issue. I usually rate such papers 1 or 2 out of 5 points possible and mostly recommended to reject such papers. The central committee will review my review and my peers' reviews and take the decision based on them. The central committee usually agreed to my decision, so I guess that other reviewers would have voiced the same concerns. Some industrial papers are indeed excellent, in which case I won't hesitate to give 5 points. People from Microsoft Research, IBM TJ Watson, and Intels are known of publishing good papers. However, lots of their papers got rejected too.
Even worse. Take the following ex
Please...
This may be an old news, but the details of that machine is here. That's some stuff to drool over. Some excerpts:
And now this machine is up for a rent. Here's the company website.
but single-digit percentage increases in performance isn't "dramatic". It's more like "scarcely noticeable".
In that case, maybe... But if you follow some compiler conference papers, single digit percentage of improvement *is* a dramatic improvement.
More than that single digit, we need to either change the underlying algorithm, or do a more dramatic overhaul, or correct a resource hogging mistakes. Well, we all know that Mozilla coders aren't that sloppy, so I guess that single digit improvements are really good because they usually involve quite a lot of cutting corners squeezing out more improvements over the already tight code.
That's about an extra 0.01 tits/second.
You're saying... extra tits/second as an improvement unit? My mind is distorted.... ;)
Well, researchers are often paying their own way except if they are one of the chairs, which they are offered some complimentary registration fee. Some of bonafide conferences actually pay their expenses if they are invited.
Honestly, researcher communities (especially the academic ones) are disdainful to the "achievements" of "industrial research". The reject rates on industrial papers have been pretty high (usually more than 50%). This is because that the "innovations" of industrial "research" are more or less either one or some of the following: rehashing old ideas, implementing old ideas with new looks / new aspects / into new problems which often not worth mentioning, combining several old ideas in some obvious ways.
Well, this is not to say that industrial papers are crap, but of course there are some excellent industry researchers, which are usually ex-professors which are already well known before they enter the industries. However, research is like a big gamble: either you win big or you lose big. Given the current situation of the economy, it's more likely you lose big because of "lack of genuinely new ideas" and you can never get a guarantee that your research group is actually producing the great useful results for your company. It's a whole lot better for the company to actually scour the conferences, spot the prominent person with the right ideas, and then "steal" them so that they can implement the said idea for your company. This is exactly what Microsoft has been doing in the past years.
Since I never attended trade/business oriented conferences, I can't comment on those. Moreover, these conferences are usually way more expensive than the academia ones (thousands of $$ vs hundreds of $$).
I remember speaking to one of the developers in the IRC channel specifically about this. Their response was "write a gnome-vfs module for it."
Well, that's typical OSS developers... :) I'll bet that the iPod support is because someone actually came to the IRC channel and flame "Rhythmbox sux because it doesn't support iPod". ;)
For Slashdot tradition completeness sake:
[] with Cowboyneal
or
[] I don't code in Linux, you insensitive clod!
Well, having the trend of people adding TLDs like nobody's business, we would come into the same problem albeit a lot slower. Then, someone has to invent some kind of smarter algorithm for this kind of problem. This is an inherent problem needs to be solved. Just watch for the next communication / network conferences... :)
If that's the case, then... either the strongest wins (mucho dinero), or the first come first served rules. As simple as that. This would happen exactly the same even with the TLD intact.
Not having TLD at all... Like http://slashdot
That would be cooler because most modern browser may omit the http:// part. Lots of business would covet those!
You can see their howto pages to see it in action. You probably want to check out some sample outputs. And this project also uses LilyPond. Check that out.
And you can take a conclusion: Aunt Tillie's intelligence >> PHB's intelligence. And... in an RPG world, the same would apply for strength, dexterity, wisdom, and charisma. Oh yes, it takes all these to get the CLI working... ;)
Well, IMHO, this is exactly what Sun did with Sun Community Source Licensing (SCSL). Check out their principles/a and see how their stand on Open Source. (Scroll down a little bit)
X is designed around 1984 but available widely 1987 and NeWS is also around 1987. By then they are not well known and relatively had the same number of adherents. However, since X is open, X is embraced by many in a short time and thus X won. It's not just "open source and then win" idea like ESR suggested, but look at the situation closely.
Microsoft is doing good with C#, but they're open source because they're hoping that they get a quick momentum and get a new adherents shortly to contend Java, which has already been well-established, not just "weakening" per se.
Since Java has already had lots of adherents, opening up won't add some advantage. If later C# is gaining ground (which I highly doubt), then Sun may consider that (which, again, I highly doubt even such case happen except that C# starts overtaking Java).
No, my point is, to play catch-up, the acceleration must be boosted much more than the current level, which IMHO is not fast enoough to even play catch-up.
Swing and AWT has been planned way before and there are developments IIRC 3 years ago and only now it produces at least something real. Back then it's still Java 1.3 and they tried to get full Java 1.2 compatibility... until now. Of course there are parts of JDK 1.3 and 1.4 that have been added as well, but look at the status page and get a picture of how dire the situation is. I'm not saying that GCJ developers are not hardworking enough, though.
Of course I'd like to see many addition that GCJ offer such as native compilation and others, but until it's truly usable fast enough, the relative progress seems to be a downward spiral and thus making the great-grand parent post ("give GCJ hippie" thing) a moot.
If you have been following gcj's progress recently it is maturing rapidly, just give those dirty gnu hippies a few more years they'll be on par with Sun's own implimentation.
Two facts:
This lead me to conclude that GCJ would never catch Sun's level unless something drastic changes happen. Don't get me wrong, I like GCJ's idea and try to support it, but I firmly believe that they need to change to boost more progress.