On 1/6/2012 5:13 PM, HardySpicer wrote:> On Jan 7, 11:14 am, Rob Gaddi<rga...@technologyhighland.invalid> > wrote: >> On Fri, 06 Jan 2012 04:05:37 -0800 >> >> Rick Lyons<R.Lyons@_BOGUS_ieee.org> wrote: >>> [snip] >>> So that's it. Here's our chance to let the IEEE Sig. >>> Processing Society officers know what we think. >>> If you have any suggestions, please reply to this post. >> >> I'm with the masses. Pricing, pricing, pricing. The IEEE paper that I'm looking for is absolutely worth a $35 download to me. The right to read enough of a paper to determine whether it really is the paper I'm looking for, or if it's mathematical gibberish with no practical application, isn't even with 35 cents. >> >> Yes, they need to be written with the practicing engineer in mind; as someone already said that means getting to the point. I don't care THAT you've created a novel means of transforming this matrix into that one, I care why. And if why doesn't have an answer, then I most certainly don't care. But none of that will mean anything if I don't even glance at the papers to see whether they're any better, and I'm absolutely not going to do so at $35/glance. > > Engineering has become Science. Often there doesn't need to be an > immediate reason or application, just a theory. If we went on > practicing engineers as reviewers only practical working projects > would be published. What is wrong is that there needs to be other > journals for practice just as the Transaction is for mainly theory. > > HardyIf "theory" is writing some terse, complex and not-very-useful paper then I object. Take a look at the classical papers on theory. They're generally understandable. Take a look at most current "theory" papers and wonder WTF?? are they talking about and who cares? Fred
Lyons needs your advice regarding the IEEE
Started by ●January 6, 2012
Reply by ●January 7, 20122012-01-07
Reply by ●January 7, 20122012-01-07
Fred Marshall <fmarshallxremove_the_x@acm.org> wrote: (snip)> If "theory" is writing some terse, complex and not-very-useful paper > then I object. Take a look at the classical papers on theory. They're > generally understandable. Take a look at most current "theory" papers > and wonder WTF?? are they talking about and who cares?The more general theory was done longer ago, so now the papers are on more and more specific things. Many years ago, when deciding which seminars I wanted to see, I decided that the less interesting the title sounded the more interesting the seminar. What that really means is the more general the subject, and so less interesting title, the more interesting the talk would be. I still find stories about how some things that we all know about were discovered many years ago, and the thought process that went into them. There is a book called "Feynman's Lost Lecture" about a lecture that Feynman gave on Newton's calculations showing planetary orbits were elliptical. While we now do it fairly easily with Calculus, Newton did it without Calculus. Either it hadn't been invented yet or no-one else would understand and believe the explanation. It seems that it was at the time when conic sections were a popular math subject, and Newton knew many details that we don't normally know today. Feynman gave a lecture on his understanding of Newton's derivation, but, unlike most of the lectures there was no transcript. There were some very brief notes, though. Then, while on vacation, David Goodstein rederived Feynman's rederivation of Newton's explanation and wrote a book about it. David Goodstein also did the "Mechanical Universe" series of videos on physics, which are unusually understandable even by people not so used to physics lectures, and are now available for web viewing. (And also interesting in following how many discoveries were made years ago.) -- glen
Reply by ●January 7, 20122012-01-07
On Jan 7, 6:06�pm, Fred Marshall <fmarshallxremove_th...@acm.org> wrote:> On 1/6/2012 5:13 PM, HardySpicer wrote: > > > > > On Jan 7, 11:14 am, Rob Gaddi<rga...@technologyhighland.invalid> > > wrote: > >> On Fri, 06 Jan 2012 04:05:37 -0800 > > >> Rick Lyons<R.Lyons@_BOGUS_ieee.org> �wrote: > >>> [snip] > >>> So that's it. �Here's our chance to let the IEEE Sig. > >>> Processing Society officers know what we think. > >>> If you have any suggestions, please reply to this post. > > >> I'm with the masses. �Pricing, pricing, pricing. �The IEEE paper that I'm looking for is absolutely worth a $35 download to me. �The right to read enough of a paper to determine whether it really is the paper I'm looking for, or if it's mathematical gibberish with no practical application, isn't even with 35 cents. > > >> Yes, they need to be written with the practicing engineer in mind; as someone already said that means getting to the point. �I don't care THAT you've created a novel means of transforming this matrix into that one, I care why. �And if why doesn't have an answer, then I most certainly don't care. �But none of that will mean anything if I don't even glance at the papers to see whether they're any better, and I'm absolutely not going to do so at $35/glance. > > > Engineering has become �Science. Often there doesn't need to be an > > immediate reason or application, just a theory. If we went on > > practicing engineers as reviewers only practical working projects > > would be published. What is wrong is that there needs to be other > > journals for practice just as the Transaction is for mainly theory. > > > Hardy > > If "theory" is writing some terse, complex and not-very-useful paper > then I object. �Take a look at the classical papers on theory. �They're > generally understandable. �Take a look at most current "theory" papers > and wonder WTF?? are they talking about and who cares? > > FredNot so sure I agree. Just because you don't understand a paper doesn't mean it's always going to be rubbish! The whole point of discovery is to push boundaries. You may not care and I may not care but that's not the point. Hardy
Reply by ●January 7, 20122012-01-07
On 7 Jan, 14:31, HardySpicer <gyansor...@gmail.com> wrote:> On Jan 7, 6:06�pm, Fred Marshall <fmarshallxremove_th...@acm.org> > wrote:> > If "theory" is writing some terse, complex and not-very-useful paper > > then I object. �Take a look at the classical papers on theory. �They're > > generally understandable. �Take a look at most current "theory" papers > > and wonder WTF?? are they talking about and who cares? > > > Fred > > Not so sure I agree. Just because you don't understand a paper doesn't > mean it's always going to be rubbish! The whole point of discovery is > to push boundaries. You may not care and I may not care but that's not > the point.In the late '90s, DoA estimators like MUSIC and ESPRIT was hyped up. 'Smart antennas' was the buzzword of the day, the question how to serve as many mobile phines as possible with inexpensive base stations. I happened to be one of very few who atually attempted to *use* those kinds of methods in a practical application, seismic data processing. The methods has an achilles heel which is obvious from the maths, but not awfully important my particular application (the seismic data qqcuisition sytems are 10-50-fold redundant): One needs to tune the reciever array to the signal one processes. The mathematics of parametric methods links the number of antenna elements N to the order P of the signal autocorrelation matrix with the number D of signals present, as D < P <= c*N. where c is a constant (either 1 or 1/2) that depends on certain details of the signal processing chain. So the end effect is that if P = c*N and D >= P, then the signal no longer complies with the assumptions behind the signal processing methods, and nothing works. All this is obvious from merely reading the papers. Again, in my particular application which was interactive, a human operator could play with N and P more or less at will, and so find a set of processing parameters which worked. Not so in an unsupervised system like a phone base station. Nonetheless, a colleague of mine was approached by some person who happened to have got a job with these smart antennas, with some questions about MUSIC and ESPRIT. I wrote a quick email much like some of the more elaborate posts one occationally finds here, with some rudimentary explanations and pointers to further references. The next thing I know, this person is quoted in periodicals as the expert on such methods, within the organization. He didn't know a lot more than the occasional posters who ask stray homework questiones here, but still he predicted what methods would be used and what lines of inquiry to pursue. Anyone with the ability to read and contemplate maths will spot the problem in half a second. The point is that IEEE and similar organizations have gone totally astray. They publish papers that are poorly written, written by people who don't understand the subject and whose sole interest in their work is to tick some boxes in order to get a diploma to hang on the office wall. Reviewers are not better - they are among the would-be authors for such journals, and are as such unable to spot the problems in the manuscripts. At least the phony degrees are recognizable as such. These people are a lot harder to spot, since their publications and diplomas are with used-once-upon-a- time-to-be-credible institutions. Rune
Reply by ●January 7, 20122012-01-07
On 1/7/2012 8:31 AM, glen herrmannsfeldt wrote: ...> There is a book called "Feynman's Lost Lecture" about a lecture > that Feynman gave on Newton's calculations showing planetary orbits > were elliptical. While we now do it fairly easily with Calculus, > Newton did it without Calculus. Either it hadn't been invented yet > or no-one else would understand and believe the explanation. > > It seems that it was at the time when conic sections were a popular > math subject, and Newton knew many details that we don't normally > know today. Feynman gave a lecture on his understanding of Newton's > derivation, but, unlike most of the lectures there was no transcript. > There were some very brief notes, though.... I find that hard to understand, given earlier history. Kepler had, by observation, calculated planetary orbits to be elliptical, with the sun at one focus, and a line from the sun to the planet (the "radius vector") sweeping out equal areas in equal times. Since Pythagoras, a circular orbit was considered perfect -- "ellipse" implies something missing -- and there was a philosophical conflict between going with the calculable and adhering to heavenly perfection. The issue was a major preoccupation of natural philosophers of the day. As I heard it, Newton was as aware of the question as anyone, but what set him apart was the notion of universal gravitation that had been germinating since an exchange he had had with Robert Hooke. Edmund Haley asked Newton for his opinion on the subject, and Newton answered bluntly, "They are ellipses." When Halley asked how he knew, Newton answered with what I think are the most astonishing words in the history of science: "I have calculated it." To do that calculation, he had to lay out the laws of motion and gravity, and invent a rudimentary calculus, "fluxions" as he called it. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
Reply by ●January 7, 20122012-01-07
On 1/7/2012 5:31 AM, HardySpicer wrote:> On Jan 7, 6:06 pm, Fred Marshall<fmarshallxremove_th...@acm.org> > wrote: >> On 1/6/2012 5:13 PM, HardySpicer wrote: >> >> >> >>> On Jan 7, 11:14 am, Rob Gaddi<rga...@technologyhighland.invalid> >>> wrote: >>>> On Fri, 06 Jan 2012 04:05:37 -0800 >> >>>> Rick Lyons<R.Lyons@_BOGUS_ieee.org> wrote: >>>>> [snip] >>>>> So that's it. Here's our chance to let the IEEE Sig. >>>>> Processing Society officers know what we think. >>>>> If you have any suggestions, please reply to this post. >> >>>> I'm with the masses. Pricing, pricing, pricing. The IEEE paper that I'm looking for is absolutely worth a $35 download to me. The right to read enough of a paper to determine whether it really is the paper I'm looking for, or if it's mathematical gibberish with no practical application, isn't even with 35 cents. >> >>>> Yes, they need to be written with the practicing engineer in mind; as someone already said that means getting to the point. I don't care THAT you've created a novel means of transforming this matrix into that one, I care why. And if why doesn't have an answer, then I most certainly don't care. But none of that will mean anything if I don't even glance at the papers to see whether they're any better, and I'm absolutely not going to do so at $35/glance. >> >>> Engineering has become Science. Often there doesn't need to be an >>> immediate reason or application, just a theory. If we went on >>> practicing engineers as reviewers only practical working projects >>> would be published. What is wrong is that there needs to be other >>> journals for practice just as the Transaction is for mainly theory. >> >>> Hardy >> >> If "theory" is writing some terse, complex and not-very-useful paper >> then I object. Take a look at the classical papers on theory. They're >> generally understandable. Take a look at most current "theory" papers >> and wonder WTF?? are they talking about and who cares? >> >> Fred > > Not so sure I agree. Just because you don't understand a paper doesn't > mean it's always going to be rubbish! The whole point of discovery is > to push boundaries. You may not care and I may not care but that's not > the point. > > HardyHardy, I didn't say "hard to digest". I agree that's part of research or study or ... whatever you want to call it. So, I'll try to express the thought more carefully: Take a look at the classical papers. Shannon, etc. Why are they "classical" i.e. having stood the test of time? My conjecture is that they are 1) important and 2) often easy to read ("easy" as compared to "obscure") so "easy" here is a subjective thing and surely what's easy for some is not so easy for another. Maybe I should have said "readable". I didn't say that *I* don't understand these papers ... heck there are lots of good ones I don't understand .. but I do believe that I can scan a paper and get a pretty good idea if they are going to be understandable to a broader audience. You can too I'll bet. In contrast, having paid some attention to the current output, I find papers that are 1) not important .. well by itself that may be OK if they are at least useful ... and are 2) tersely written for "show" and not for exposition really. Try this: Someone writes a paper that is very obscure but actually makes a good point. Later, someone else presents the ideas in an accessible way. Who gets read and who gets cited? (No, I'm not suggesting plagiarism or that the first author never gets cited). Unless you're an historian, whose paper do you value? I go with the latter and I think that's a key point of this thread. Back to your last post... well, *somebody* better care *sometime*! We agree it doesn't have to be you or me. It's not a perfect world and some things will always be irritating. But Rick's question has some important answers. Fred
Reply by ●January 7, 20122012-01-07
>On Jan 7, 6:06=A0pm, Fred Marshall <fmarshallxremove_th...@acm.org> >wrote: >> On 1/6/2012 5:13 PM, HardySpicer wrote: >> >> >> >> > On Jan 7, 11:14 am, Rob Gaddi<rga...@technologyhighland.invalid> >> > wrote: >> >> On Fri, 06 Jan 2012 04:05:37 -0800 >> >> >> Rick Lyons<R.Lyons@_BOGUS_ieee.org> =A0wrote: >> >>> [snip] >> >>> So that's it. =A0Here's our chance to let the IEEE Sig. >> >>> Processing Society officers know what we think. >> >>> If you have any suggestions, please reply to this post. >> >> >> I'm with the masses. =A0Pricing, pricing, pricing. =A0The IEEE papert=>hat I'm looking for is absolutely worth a $35 download to me. =A0The right=>to read enough of a paper to determine whether it really is the paper I'ml=>ooking for, or if it's mathematical gibberish with no practicalapplication=>, isn't even with 35 cents. >> >> >> Yes, they need to be written with the practicing engineer in mind; as=>someone already said that means getting to the point. =A0I don't care THAT=>you've created a novel means of transforming this matrix into that one, Ic=>are why. =A0And if why doesn't have an answer, then I most certainly don't=>care. =A0But none of that will mean anything if I don't even glance at the=>papers to see whether they're any better, and I'm absolutely not going tod=>o so at $35/glance. >> >> > Engineering has become =A0Science. Often there doesn't need to be an >> > immediate reason or application, just a theory. If we went on >> > practicing engineers as reviewers only practical working projects >> > would be published. What is wrong is that there needs to be other >> > journals for practice just as the Transaction is for mainly theory. >> >> > Hardy >> >> If "theory" is writing some terse, complex and not-very-useful paper >> then I object. =A0Take a look at the classical papers on theory.=A0They'=>re >> generally understandable. =A0Take a look at most current "theory"papers>> and wonder WTF?? are they talking about and who cares? >> >> Fred > >Not so sure I agree. Just because you don't understand a paper doesn't >mean it's always going to be rubbish! The whole point of discovery is >to push boundaries. You may not care and I may not care but that's not >the point. > >HardyI bet 7 out of 10 engineers would say PHDs in the field are hard to talk to. They even can be 'spacey' at times. Now try to read a paper of theirs. If you want to relate to the 'practicing engineer', you gotta be more down to earth then some people will read and care about your paper (not for 35 dollars) but maybe the google PDF version. (Hell... even a woman will possibly talk to you once or twice).
Reply by ●January 7, 20122012-01-07
On Sat, 07 Jan 2012 12:20:23 -0500, Jerry Avins <jya@ieee.org> wrote:>On 1/7/2012 8:31 AM, glen herrmannsfeldt wrote: > > ... > >> There is a book called "Feynman's Lost Lecture" about a lecture >> that Feynman gave on Newton's calculations showing planetary orbits >> were elliptical. While we now do it fairly easily with Calculus, >> Newton did it without Calculus. Either it hadn't been invented yet >> or no-one else would understand and believe the explanation. >> >> It seems that it was at the time when conic sections were a popular >> math subject, and Newton knew many details that we don't normally >> know today. Feynman gave a lecture on his understanding of Newton's >> derivation, but, unlike most of the lectures there was no transcript. >> There were some very brief notes, though. > > ... > >I find that hard to understand, given earlier history. Kepler had, by >observation, calculated planetary orbits to be elliptical, with the sun >at one focus, and a line from the sun to the planet (the "radius >vector") sweeping out equal areas in equal times. Since Pythagoras, a >circular orbit was considered perfect -- "ellipse" implies something >missing -- and there was a philosophical conflict between going with the >calculable and adhering to heavenly perfection. The issue was a major >preoccupation of natural philosophers of the day. > >As I heard it, Newton was as aware of the question as anyone, but what >set him apart was the notion of universal gravitation that had been >germinating since an exchange he had had with Robert Hooke. Edmund Haley >asked Newton for his opinion on the subject, and Newton answered >bluntly, "They are ellipses." When Halley asked how he knew, Newton >answered with what I think are the most astonishing words in the history >of science: "I have calculated it."Abraham de Moivre described it thusly: "In 1684 Dr Halley came to visit [Newton] at Cambridge. After they had been some time together, the Dr asked him what he thought the curve would be that would be described by the planets supposing the force of attraction towards the sun to be reciprocal to the square of their distance from it. Sir Isaac replied immediately that it would be an ellipse. The Doctor, struck with joy and amazement, asked him how he knew it. Why, saith he, I have calculated it. Whereupon Dr Halley asked him for his calculation without any farther delay. Sir Isaac looked among his papers but could not find it, but he promised him to renew it and then to send it him..." Which became the Principia.>To do that calculation, he had to lay out the laws of motion and >gravity, and invent a rudimentary calculus, "fluxions" as he called it.Nature and nature's laws lay hid in night; God said "Let Newton be" and all was light. -- Alexander Pope -- Rich Webb Norfolk, VA
Reply by ●January 7, 20122012-01-07
On Jan 8, 4:06�am, Rune Allnor <all...@tele.ntnu.no> wrote:> On 7 Jan, 14:31, HardySpicer <gyansor...@gmail.com> wrote: > > > On Jan 7, 6:06�pm, Fred Marshall...@acm.org> > > wrote: > > > If "theory" is writing some terse, complex and not-very-useful paper > > > then I object. �Take a look at the classical papers on theory. �They're > > > generally understandable. �Take a look at most current "theory" papers > > > and wonder WTF?? are they talking about and who cares? > > > > Fred > > > Not so sure I agree. Just because you don't understand a paper doesn't > > mean it's always going to be rubbish! The whole point of discovery is > > to push boundaries. You may not care and I may not care but that's not > > the point. > > In the late '90s, DoA estimators like MUSIC and ESPRIT > was hyped up. 'Smart antennas' was the buzzword of the day, > the question how to serve as many mobile phines as possible > with inexpensive base stations. > > I happened to be one of very few who atually attempted to > *use* those kinds of methods in a practical application, > seismic data processing. The methods has an achilles heel > which is obvious from the maths, but not awfully important > my particular application (the seismic data qqcuisitionsytems are 10-50-fold redundant): > > One needs to tune the reciever array to the signal one > processes. > > The mathematics of parametric methods links the number > of antenna elements N to the order P of the signal > autocorrelation matrix with the number D of signals > present, as > > D < P <= c*N. > > where c is a constant (either 1 or 1/2) that depends on > certain details of the signal processing chain. > > So the end effect is that if P = c*N and D >= P, > then the signal no longer complies with the assumptions > behind the signal processing methods, and nothing works. > > All this is obvious from merely reading the papers. > > Again, in my particular application which was interactive, > a human operator could play with N and P more or less at > will, and so find a set of processing parameters which > worked. > > Not so in an unsupervised system like a phone base station. > > Nonetheless, a colleague of mine was approached by some > person who happened to have got a job with these smart > antennas, with some questions about MUSIC and ESPRIT. > I wrote a quick email much like some of the more elaborate > posts one occationally finds here, with some rudimentary > explanations and pointers to further references. > > The next thing I know, this person is quoted in periodicals > as the expert on such methods, within the organization. > > He didn't know a lot more than the occasional posters > who ask stray homework questiones here, but still he > predicted what methods would be used and what lines > of inquiry to pursue. > > Anyone with the ability to read and contemplate maths > will spot the problem in half a second. > > The point is that IEEE and similar organizations have > gone totally astray. They publish papers that are poorly > written, written by people who don't understand the > subject and whose sole interest in their work is to > tick some boxes in order to get a diploma to hang on > the office wall. Reviewers are not better - they are > among the would-be authors for such journals, and are > as such unable to spot the problems in the manuscripts. > > At least the phony degrees are recognizable as such. > These people are a lot harder to spot, since their > publications and diplomas are with used-once-upon-a- > time-to-be-credible institutions. > > RuneWell if it has gone astray it has done so for the past 30 years or more at least! That's the time period it has been producing such papers. I disagree however that all such papers are rubbish which is the argument being raised here. What I don't like are people publishing the same thing over and over again with a different title! When blind source separation first came out I thought it would be rubbish but it has turned out to be quite a tool and now even some are attempting real-time applications of the (convolutive) theory. In the US at least, most Ph.Ds at Unis who have tenure track have done their stint in industry so are well aware of the practicalities and limitations. You have to start somewhere. I remember when people scoffed at Kalman filters. I remember iit well. They also scoffed at H infinity controllers. Now H infinity has been used as a control method in the European fighter jet. Not saying there aren't alternatives to many or all such theories but if you don't try you will never know what you're missing. Cellular networks was a theory at one time. People scoffed at speech recognition and said it had no applications. Look at it now in mobile phones. Software radio - what a daft idea, why did we need that when we could do it all in hardware? and so on... Hardy
Reply by ●January 7, 20122012-01-07
On 7 Jan, 20:37, HardySpicer <gyansor...@gmail.com> wrote:> On Jan 8, 4:06�am, Rune Allnor <all...@tele.ntnu.no> wrote: > > > > > > > On 7 Jan, 14:31, HardySpicer <gyansor...@gmail.com> wrote: > > > > On Jan 7, 6:06�pm, Fred Marshall...@acm.org> > > > wrote: > > > > If "theory" is writing some terse, complex and not-very-useful paper > > > > then I object. �Take a look at the classical papers on theory. �They're > > > > generally understandable. �Take a look at most current "theory" papers > > > > and wonder WTF?? are they talking about and who cares? > > > > > Fred > > > > Not so sure I agree. Just because you don't understand a paper doesn't > > > mean it's always going to be rubbish! The whole point of discovery is > > > to push boundaries. You may not care and I may not care but that's not > > > the point. > > > In the late '90s, DoA estimators like MUSIC and ESPRIT > > was hyped up. 'Smart antennas' was the buzzword of the day, > > the question how to serve as many mobile phines as possible > > with inexpensive base stations. > > > I happened to be one of very few who atually attempted to > > *use* those kinds of methods in a practical application, > > seismic data processing. The methods has an achilles heel > > which is obvious from the maths, but not awfully important > > my particular application (the seismic data qqcuisitionsytems are 10-50-fold redundant): > > > One needs to tune the reciever array to the signal one > > processes. > > > The mathematics of parametric methods links the number > > of antenna elements N to the order P of the signal > > autocorrelation matrix with the number D of signals > > present, as > > > D < P <= c*N. > > > where c is a constant (either 1 or 1/2) that depends on > > certain details of the signal processing chain. > > > So the end effect is that if P = c*N and D >= P, > > then the signal no longer complies with the assumptions > > behind the signal processing methods, and nothing works. > > > All this is obvious from merely reading the papers. > > > Again, in my particular application which was interactive, > > a human operator could play with N and P more or less at > > will, and so find a set of processing parameters which > > worked. > > > Not so in an unsupervised system like a phone base station. > > > Nonetheless, a colleague of mine was approached by some > > person who happened to have got a job with these smart > > antennas, with some questions about MUSIC and ESPRIT. > > I wrote a quick email much like some of the more elaborate > > posts one occationally finds here, with some rudimentary > > explanations and pointers to further references. > > > The next thing I know, this person is quoted in periodicals > > as the expert on such methods, within the organization. > > > He didn't know a lot more than the occasional posters > > who ask stray homework questiones here, but still he > > predicted what methods would be used and what lines > > of inquiry to pursue. > > > Anyone with the ability to read and contemplate maths > > will spot the problem in half a second. > > > The point is that IEEE and similar organizations have > > gone totally astray. They publish papers that are poorly > > written, written by people who don't understand the > > subject and whose sole interest in their work is to > > tick some boxes in order to get a diploma to hang on > > the office wall. Reviewers are not better - they are > > among the would-be authors for such journals, and are > > as such unable to spot the problems in the manuscripts. > > > At least the phony degrees are recognizable as such. > > These people are a lot harder to spot, since their > > publications and diplomas are with used-once-upon-a- > > time-to-be-credible institutions. > > > Rune > > Well if it has gone astray it has done so for the past 30 years or > more �at least! That's the time period it has been producing such > papers. > I disagree however that all such papers are rubbish which is the > argument being raised here. What I don't like are people publishing > the same thing over and over again with a different title!Agreed. That was the direct reason I stopped publishing on my PhD thesis. The same stuff over and over.> When blind source separation first came out I thought it would be > rubbish but it has turned out to be quite a tool and now even some are > attempting real-time applications of the (convolutive) �theory. In the > US at least, most Ph.Ds at Unis who have tenure track have done their > stint in industry so are well aware of the practicalities and > limitations. You have to start somewhere.That's not the point. The point is that there is a vast difference between 'does not obviously work' and 'does obviously not work'. Rune
