I don't really get this... it's only "suggesting" what you're probably trying to type. If you prefer to just type out the whole url yourself, it's not going to stop you or overwrite it with their suggestion. They are just there for convenience -- if the algorithm is good enough, it will actually save you typing at least some of the time, and will never require you to type more than if it weren't there.
The problem is that it already had this feature, implemented in a way that I, and many other people, liked. Now they've reimplemented it in a way that we hate. For instance, I used to be in the habit of typing "en", and having it complete that to "en.wikipedia.org", so I could go to wikipedia with just a few keystrokes. Now that no longer works, because it pops up a dozen possibilities, such as clusty.com because its title is "clusty, the clustering search engine."
I'd just like to point out that it adaptively learns how to sort the results, so you shouldn't discard it on first use. Give it some time to come up with the most relevant URLs (for you) on top.
I've been using it for months now, and it hasn't adapted to what I want. Anyway, I don't want to have to train software by painstaking repetition, and with uncertain prospects of success, like it's a puppy that needs to be housebroken. I want to be able to configure it to operate the way I want. Software that tries is guess what you want is a bad idea. Users don't want software that behaves unpredicatably, they want software that behaves predictably.
In the version of FF I'm running (3.0), I don't have a boolean browser.urlbar.richResults, I have a browser.urlbar.maxRichResults, which is an integer.
Here's a summary of what I've been able to figure out about how to get rid of the awesome bar:
To revert to the old-style graphics, use the oldbar addon. This has
no effect on the actual completion algorithm, just on the way the results are displayed.
Set browser.urlbar.matchOnlyTyped to true. This is supposed to make it only match results that you've actually typed before. However, it will
not stop it from matching titles of pages.
Setting browser.urlbar.maxRichResults to 0 does not change it back to the old behavior. It just prevents results from being shown in
a pop-up menu at all. Setting it to less than the default of 12 can, however, reduce the amount of screen space taken up by the menu.
Setting browser.urlbar.matchBehavior to 2 makes the matching algorithm slightly more like the old one. It will only match things that lie at the beginning of a word boundary. This cuts down on the number of stupid matches, e.g., it will no longer match "ebay" with "thepiratebay."
What I really want is a way to make it search only on urls, not titles. When I type a url in the url bar, I have a url in mind. I don't want it to match titles.
Not only is this an extra install for the end user, it also means not all platforms and browsers will be supported (A great example being no flash/silverlight on the iPhone...)
I'm with you all the way as far as preferring standards over proprietary stuff. However, the iPhone seems like a bad example to me. It's a proprietary platform, controlled by Apple. It's also a machine with a very low-powered CPU compared to the typical desktop system, so flash probably wouldn't be viable on it even if Apple wanted it to be.
Apple has a chance here to keep the Web becoming even more proprietary as Silverlight and Flash battle it out to lock the Web application market into one proprietary format or another.
It's not true that Flash is completely proprietary. There are multiple open-source compilers, and there's an open-source browser plugin. You do have to work hard to develop in flash using an OSS software stack, but there are people doing it. Gnash, the open-source browser plugin, has gotten to the point where it can play you-tube videos, provided you have the right hardware and sacrifice an unblemished calf. Adobe has also been slowly moving in the right direction as far as open-sourcing some of their code, and relaxing some of the more onerous licensing restrictions. A lot of the problems with making flash more open are actually problems with codecs, and that situation is also showing signs of improving, with support for less patent-encumbered codecs being added to newer versions of flash.
Originally, yes, email was text only, but like everything else it evolves as the users and developers want.
Which users? Not all users want to receive emails that are nothing but a line of text and an MS-Word attachment, or that have text in cyan letters overlaid on a background of ocean waves crashing on the beach.
HTML is pretty much the standard in most corporate mail systems, as far as I can see, and I do myself use it for basic formatting and inline images. The most common usage I see is for quoting and inline replies with multiple colours. While this is partly generated by the limitations of Outlook, I actually find it easier to read as well because each quote is identifiable by color, rather than just the level of indent. In fact there is a TBird plugin to do the same thing automatically...
No harm done, as long as you also send a plain-text version, and as long as you realize that not all users will see your html special effects.
Thanks for the correction. I checked my software, and actually the bounce message I've been sending out is, "The mail was sent only in html format. Normal practice is to send mail either in plain text format or in both plain text and html."
As various people have pointed out, this would only really work if you sent html-only email, and if the recipient was guaranteed to have client software that executed javascript or something. I use mutt, a text-only email reader, and I have my mail software set up so it bounces html-only email (that it doesn't think is spam) back to the sender with an error message explaining that html-only email violates internet standards. I've never understood why anyone sends html-only email. Seems hard to believe that there would be service providers so clueless that they'd make html-only the default, and it also seems hard to believe that people would be clueless enough to want to send html-only email, but clueful enough to switch to html-only if it wasn't the default.
I have to admit that the concept of being able to get a return receipt for email has a certain allure. Recently, for example, my boss got pissed off at me and made a big scene because he thought I hadn't notified him about something. I happened to have a copy of the email in which I notified him, and I also happened to have saved his reply to it. But what if I hadn't saved the reply, or if he hadn't replied?
A lot of people send CYA emails, e.g., "Okay, this is to confirm that you want me to put the uranium in the crisper drawer of the fridge, and that you take responsibility for the results." But the recipient can pretend he never got it.
I clicked through to your link to BitNabber. Their cheapest service is $20/mo. I don't think most people want to pay $20/mo for something that their ISP told them was going to be included in their standard service.
Cool! Thanks for the reply. I was really starting to think that mathml and svg were going to end up never being supported in any standard, cross-browser way, and it would just be a total defeat for standards. This is good news.
It works just fine and is as standards compliant as IE lets it be.
I realize that MS's poor support for the standard isn't your fault, but basically IE doesn't support xhtml in any standard way, and doesn't make it possible for mathml to be supported in a standard way by third-party developers like you.
There are guidelines for serving up MathML in a way that is compatible with both IE and FF on our website: www.dessci.com/mathplayer.
Yes, but IIRC it's a technique that violates standards, and it doesn't pass the test I described above of behaving well (possibly with degradation) in the three most important cases: a default install of IE, FF with mathml support, and IE with mathplayer. If you have a single html or xhtml file that passes that three-part test, when served with a single filename, I'd honestly be very, very happy to see it, because it would solve a lot of problems for me. (I believe there is a technique that the w3c suggests that does this, but it uses CSS tricks in an incredibly obscure, complex, and ugly way, and I think it's different from what your site suggests.)
While making IE properly display math takes downloading and installing a small plugin, makeing Firefox properly display math takes downloading and installing fonts.
I think you're right that at the present time it's slightly easier to get MathML working in IE with your plugin than it is to get it working in FF with open-source software. The difference is that MathML support will probably be built in to Firefox by default very soon, since the STIX fonts are due to have a final production version released in August.
The article includes a lot of confusion and/or FUD about licensing.
"There's a great fear sometimes, that if I use open source, will I lose my intellectual property?" acknowledged Novell's Levy. Other panelists Randy Hergett, director of engineering for the Open Source and Linux Organizations at HP, and Marcus Rex, CTO at the Linux Foundation, sought to assuage those fears.
"The current license for Linux requires you give back any changes you make to the open source community, but there's no way anyone can require those assurances and there's no way we'd know," Rex said.
Someone needs to sit down with some of these people and explain to them what the GPL actually says. It doesn't require software written to run on Linux to be GPL'd. Even if you had some reason why you wanted to modify the Linux kernel itself (and why the hell would a Wall Street firm want to!?), you wouldn't need to GPL your modifications unless you were turning around and selling or distributing the modified version publicly.
We seem to be getting a lot of this kind of idiocy recently. Maybe it's good news -- it might just be a sign that a lot of PHBs are getting open source on their radar for the first time. But you'd think that lawyers and journalists would at least get it straight before they published their thoughts on the web.
As far as I can tell, MS has no intention ever to support xhtml, which means it can never properly support inline mathml. Essentially a lot of people are treating xhtml as a failure, and focusing on html 5 instead. The problem with that is that it makes it impossible to support inline mathml or svg.
It's really a shame that mathml is so poorly supported in browsers. There's basically no practical, reasonable way to do write a single xhtml page that will do something reasonable in most browsers and display inline mathml. Firefox requires the file to have extension.xhtml, and standards say to serve it as application/xhtml+xml; but if you do this, a default install of IE will display a file download dialog, with a warning that "some files can harm your computer." IE wants it served as text/html, and will only display the mathml if the user has installed the MathPlayer plugin. The MathPlayer plugin also implements mathml in way that isn't standards-compliant. The bare minimum you really need is:
The majority of users, who have IE with no plugin, should see some kind of graceful degradation.
Firefox users should see the math displayed correctly.
The tiny minority of users who have MathPlayer+IE should see the math displayed correctly.
Unfortunately, you just can't accomplish this by any reasonable technique. The technique I've ended up using for the web browser version of my own physics textbooks is to use mod_rewrite to serve mathml to Firefox 3+ users, and serve a version with bitmapped renderings of the equations to everyone else. This also seems to be what DLMF has done. What a disaster.
I (intentionally) flew cross country without ID in 1987. They didn't even ask for it back then. But like you, I had assumed it was simply impossible to fly without ID in the post-9/11 era.
In fact, if the mini notebooks are sold with service, that offers the chance for more differentiation.
It sounds to me like you're completely in agreement with the claims in the article. There is differentiation in subnotebooks if they're sold with other products or services. He compares with cell phones, which have differentiation because they're sold with other products or services. Cell phone hardware is basically free, but cell phone service is expensive. He's predicting the same thing with subnotebooks.
I'd also suggest that the state variables to describe each neuron and synaptic connection would be fairly complex, so the 16,000 times bigger probably shrinks quite a bit (hint - 1,000 separate connections per neuron can't be efficiently represented in less than 1,000 bits - and if we need FP accuracy, we're talking 32Kb / neuron).
Sure, let's go with your assumption of 32 kb/neuron rather than 10 kb/neuron. That means you add 2 bits on, and now the estimate is that you need 52 bits. It doesn't affect the result in any significant way. The whole thing is just an order-of-magnitude estimate, and I think you're sort of missing the point. If you could directly simulate a human brain on a computer, it would mean immortality, the end of history, the transformation of the human race into something completely different. The fact that you can do that in a 50- or 52-bit address space means IMO that it's kind of silly even to talk about 128-bit pointers.
For perspective, let's imagine what it would take to fill up a 256-bit address space. The number of atoms in the observable universe is estimated to be about 10^80. A 256-bit address space would have 10^77 addresses. In other words, if you wanted to manufacture 1000 computers, each of which had enough memory to exhaust a 256-bit address space, you would need to use up all the matter in the observable universe, assuming you could manufacture one bit of memory out of one hydrogen atom. The point here is that if the nth generation of computer chips uses pointers with 8x2^n bits (where n=1 for 16-bit machines in 1980, n=2 for 32-bit machines today, etc.), then the size of the address space varies like O(2^(2^n)), which just gets big ridiculously fast.
IIRC 1.5 bits per year address space bloat is from Hennessy and Patterson. [...] At this point we have 30 unused bits of address space, assuming current apps need 32GB tops. That gives 64 bit x64 another 20 years lifetime!
Empirically, it hasn't been growing at anywhere near that rate. Ca. 1980 my TRS-80 had a 16-bit address space, and had enough memory to exhaust all of the addresses. Today, I'm using computers that have 1 Gb of memory, which is 30 bits worth of address space. That's less than 0.5 bits per year.
Also, in order to keep the actual used address space growing at a constant number of bits per year, Moore's law would have to continue indefinitely. But most experts are saying it will probably stop in 10 to 30 years. If we keep growing at 0.5 bits per year, starting now at 30 bits, and stop growing at the Moore's law rate in 2038, then we'll only be using 45 bits worth of actual address space.
It's hard to grok how big a 64-bit address space would really be. As a reality check, let's say that I want to own every movie that's ever been listed on IMDB, and store every single one of those in my computer's RAM simultaneously. If each one takes as much storage as a 5 Gb DVD, and IMDB has 400,000 movies listed, then that's a total of 2x10^15 bytes, which is 50 bits. That's 16,000 times smaller than a 64-bit address space.
As another example, the human brain has about 10^11 neurons. Each of those may be connected to 10^4 other neurons, so the total number of connections is about 10^15. That suggests that the total amount of RAM needed for direct, brute-force modeling of a human brain (assuming we knew enough to program such a model, which we don't, and had parallel processors that could run such a simulation, which we don't) might be about 10^15 bytes, which is a 50-bit address space. A 64-bit address space is 16,000 times bigger than that.
I think we're likely to see flying cars, Turing-level AI, and vacations on the moon before we need 128-bit pointers.
Furthermore, on the philosophical side, I should be allowed to take the many powerful GPL command line utilities out there, write a user-friendly GUI for it, and sell it. I am adding incredible value for the end user. If someone wants the command-line utility, they are still free to download it. If someone wants a GUI that's easy to use, they can pay me for my time to develop the GUI.
The GPL doesn't prohibit you from doing this. You can write your GUI, license it under the GPL, and sell it. The GPL doesn't prohibit selling software. For instance, Red Hat sells GPL'd software.
Oh, wait, did you mean that you wanted your GUI to be closed source? Well, actually,
I don't think the GPL prohibits you from doing that either. The license refers to "modified versions" of a work "based on" that work. I believe that this is interpreted to apply to linking to libraries, but not to executing programs via a shell. It's certainly not interpreted to apply to calling the linux kernel's interfaces, for example -- if it did, then it would be illegal to run closed-source code on a linux box.
All right, so did you actually mean that you wanted to take someone else's GPL'd code, link your code to it, and make your code closed source? Well in that case you're still off base. It's not the GPL that's preventing you from doing that, it's copyright law. The original author of the code didn't choose to give you permission to copy their code under the conditions you have in mind. If you're upset about that, you could contact the author and see if he's interested in being paid to offer his code to you under some other terms.
In fact, you can use GPL'd software without even agreeing to the license.
What you said here was also correct:
If you want to download, install and run GPL software, you need permission from the copyright owner, just as you need that permission for non-GPL software.
The GPL distinguishes three different levels of things you can do with the software.
"Executing it on a computer or modifying a private copy" doesn't require you to agree to the GPL at all. "Propagating" it requires you to agree to the GPL unless you think it falls under one of the exceptions to copyright law. "Conveying" it requires you to agree to the GPL, and imposes the highest requirements on you.
Slashdot Mirror
I don't really get this... it's only "suggesting" what you're probably trying to type. If you prefer to just type out the whole url yourself, it's not going to stop you or overwrite it with their suggestion. They are just there for convenience -- if the algorithm is good enough, it will actually save you typing at least some of the time, and will never require you to type more than if it weren't there.
The problem is that it already had this feature, implemented in a way that I, and many other people, liked. Now they've reimplemented it in a way that we hate. For instance, I used to be in the habit of typing "en", and having it complete that to "en.wikipedia.org", so I could go to wikipedia with just a few keystrokes. Now that no longer works, because it pops up a dozen possibilities, such as clusty.com because its title is "clusty, the clustering search engine."
I'd just like to point out that it adaptively learns how to sort the results, so you shouldn't discard it on first use. Give it some time to come up with the most relevant URLs (for you) on top.
I've been using it for months now, and it hasn't adapted to what I want. Anyway, I don't want to have to train software by painstaking repetition, and with uncertain prospects of success, like it's a puppy that needs to be housebroken. I want to be able to configure it to operate the way I want. Software that tries is guess what you want is a bad idea. Users don't want software that behaves unpredicatably, they want software that behaves predictably.
All it does is change the graphics back to the old style. It doesn't change the new matching algorithm. There's a whole thread about this above.
I hate the awesome bar too.
In the version of FF I'm running (3.0), I don't have a boolean browser.urlbar.richResults, I have a browser.urlbar.maxRichResults, which is an integer.
Here's a summary of what I've been able to figure out about how to get rid of the awesome bar:
What I really want is a way to make it search only on urls, not titles. When I type a url in the url bar, I have a url in mind. I don't want it to match titles.
I'm with you all the way as far as preferring standards over proprietary stuff. However, the iPhone seems like a bad example to me. It's a proprietary platform, controlled by Apple. It's also a machine with a very low-powered CPU compared to the typical desktop system, so flash probably wouldn't be viable on it even if Apple wanted it to be.
It's not true that Flash is completely proprietary. There are multiple open-source compilers, and there's an open-source browser plugin. You do have to work hard to develop in flash using an OSS software stack, but there are people doing it. Gnash, the open-source browser plugin, has gotten to the point where it can play you-tube videos, provided you have the right hardware and sacrifice an unblemished calf. Adobe has also been slowly moving in the right direction as far as open-sourcing some of their code, and relaxing some of the more onerous licensing restrictions. A lot of the problems with making flash more open are actually problems with codecs, and that situation is also showing signs of improving, with support for less patent-encumbered codecs being added to newer versions of flash.
Like I said in the comment you quoted, I have the software set up so it only bounces html-only mail if it thinks it's not spam.
Which users? Not all users want to receive emails that are nothing but a line of text and an MS-Word attachment, or that have text in cyan letters overlaid on a background of ocean waves crashing on the beach.
No harm done, as long as you also send a plain-text version, and as long as you realize that not all users will see your html special effects.
Thanks for the correction. I checked my software, and actually the bounce message I've been sending out is, "The mail was sent only in html format. Normal practice is to send mail either in plain text format or in both plain text and html."
As various people have pointed out, this would only really work if you sent html-only email, and if the recipient was guaranteed to have client software that executed javascript or something. I use mutt, a text-only email reader, and I have my mail software set up so it bounces html-only email (that it doesn't think is spam) back to the sender with an error message explaining that html-only email violates internet standards. I've never understood why anyone sends html-only email. Seems hard to believe that there would be service providers so clueless that they'd make html-only the default, and it also seems hard to believe that people would be clueless enough to want to send html-only email, but clueful enough to switch to html-only if it wasn't the default.
I have to admit that the concept of being able to get a return receipt for email has a certain allure. Recently, for example, my boss got pissed off at me and made a big scene because he thought I hadn't notified him about something. I happened to have a copy of the email in which I notified him, and I also happened to have saved his reply to it. But what if I hadn't saved the reply, or if he hadn't replied?
A lot of people send CYA emails, e.g., "Okay, this is to confirm that you want me to put the uranium in the crisper drawer of the fridge, and that you take responsibility for the results." But the recipient can pretend he never got it.
Here's a link with a higher signal to noise ratio.
I clicked through to your link to BitNabber. Their cheapest service is $20/mo. I don't think most people want to pay $20/mo for something that their ISP told them was going to be included in their standard service.
Cool! Thanks for the reply. I was really starting to think that mathml and svg were going to end up never being supported in any standard, cross-browser way, and it would just be a total defeat for standards. This is good news.
I realize that MS's poor support for the standard isn't your fault, but basically IE doesn't support xhtml in any standard way, and doesn't make it possible for mathml to be supported in a standard way by third-party developers like you.
Yes, but IIRC it's a technique that violates standards, and it doesn't pass the test I described above of behaving well (possibly with degradation) in the three most important cases: a default install of IE, FF with mathml support, and IE with mathplayer. If you have a single html or xhtml file that passes that three-part test, when served with a single filename, I'd honestly be very, very happy to see it, because it would solve a lot of problems for me. (I believe there is a technique that the w3c suggests that does this, but it uses CSS tricks in an incredibly obscure, complex, and ugly way, and I think it's different from what your site suggests.)
I think you're right that at the present time it's slightly easier to get MathML working in IE with your plugin than it is to get it working in FF with open-source software. The difference is that MathML support will probably be built in to Firefox by default very soon, since the STIX fonts are due to have a final production version released in August.
The article includes a lot of confusion and/or FUD about licensing.
Someone needs to sit down with some of these people and explain to them what the GPL actually says. It doesn't require software written to run on Linux to be GPL'd. Even if you had some reason why you wanted to modify the Linux kernel itself (and why the hell would a Wall Street firm want to!?), you wouldn't need to GPL your modifications unless you were turning around and selling or distributing the modified version publicly.
We seem to be getting a lot of this kind of idiocy recently. Maybe it's good news -- it might just be a sign that a lot of PHBs are getting open source on their radar for the first time. But you'd think that lawyers and journalists would at least get it straight before they published their thoughts on the web.
As far as I can tell, MS has no intention ever to support xhtml, which means it can never properly support inline mathml. Essentially a lot of people are treating xhtml as a failure, and focusing on html 5 instead. The problem with that is that it makes it impossible to support inline mathml or svg.
- The majority of users, who have IE with no plugin, should see some kind of graceful degradation.
- Firefox users should see the math displayed correctly.
- The tiny minority of users who have MathPlayer+IE should see the math displayed correctly.
Unfortunately, you just can't accomplish this by any reasonable technique. The technique I've ended up using for the web browser version of my own physics textbooks is to use mod_rewrite to serve mathml to Firefox 3+ users, and serve a version with bitmapped renderings of the equations to everyone else. This also seems to be what DLMF has done. What a disaster.The article the slashdot summary links to is basically a drastically shortened version of this recent article in Scientific American, plus a nutshell presentation of this paper.
I (intentionally) flew cross country without ID in 1987. They didn't even ask for it back then. But like you, I had assumed it was simply impossible to fly without ID in the post-9/11 era.
I mean, we flip-flop on issues like anti-matter every few years for physics.
Uh, no, we don't. Yes, I am a physicist.
It sounds like you don't understand the correspondence principle.
In fact, if the mini notebooks are sold with service, that offers the chance for more differentiation.
It sounds to me like you're completely in agreement with the claims in the article. There is differentiation in subnotebooks if they're sold with other products or services. He compares with cell phones, which have differentiation because they're sold with other products or services. Cell phone hardware is basically free, but cell phone service is expensive. He's predicting the same thing with subnotebooks.
I'd also suggest that the state variables to describe each neuron and synaptic connection would be fairly complex, so the 16,000 times bigger probably shrinks quite a bit (hint - 1,000 separate connections per neuron can't be efficiently represented in less than 1,000 bits - and if we need FP accuracy, we're talking 32Kb / neuron).
Sure, let's go with your assumption of 32 kb/neuron rather than 10 kb/neuron. That means you add 2 bits on, and now the estimate is that you need 52 bits. It doesn't affect the result in any significant way. The whole thing is just an order-of-magnitude estimate, and I think you're sort of missing the point. If you could directly simulate a human brain on a computer, it would mean immortality, the end of history, the transformation of the human race into something completely different. The fact that you can do that in a 50- or 52-bit address space means IMO that it's kind of silly even to talk about 128-bit pointers.
For perspective, let's imagine what it would take to fill up a 256-bit address space. The number of atoms in the observable universe is estimated to be about 10^80. A 256-bit address space would have 10^77 addresses. In other words, if you wanted to manufacture 1000 computers, each of which had enough memory to exhaust a 256-bit address space, you would need to use up all the matter in the observable universe, assuming you could manufacture one bit of memory out of one hydrogen atom. The point here is that if the nth generation of computer chips uses pointers with 8x2^n bits (where n=1 for 16-bit machines in 1980, n=2 for 32-bit machines today, etc.), then the size of the address space varies like O(2^(2^n)), which just gets big ridiculously fast.
IIRC 1.5 bits per year address space bloat is from Hennessy and Patterson. [...] At this point we have 30 unused bits of address space, assuming current apps need 32GB tops. That gives 64 bit x64 another 20 years lifetime!
Empirically, it hasn't been growing at anywhere near that rate. Ca. 1980 my TRS-80 had a 16-bit address space, and had enough memory to exhaust all of the addresses. Today, I'm using computers that have 1 Gb of memory, which is 30 bits worth of address space. That's less than 0.5 bits per year.
Also, in order to keep the actual used address space growing at a constant number of bits per year, Moore's law would have to continue indefinitely. But most experts are saying it will probably stop in 10 to 30 years. If we keep growing at 0.5 bits per year, starting now at 30 bits, and stop growing at the Moore's law rate in 2038, then we'll only be using 45 bits worth of actual address space.
It's hard to grok how big a 64-bit address space would really be. As a reality check, let's say that I want to own every movie that's ever been listed on IMDB, and store every single one of those in my computer's RAM simultaneously. If each one takes as much storage as a 5 Gb DVD, and IMDB has 400,000 movies listed, then that's a total of 2x10^15 bytes, which is 50 bits. That's 16,000 times smaller than a 64-bit address space.
As another example, the human brain has about 10^11 neurons. Each of those may be connected to 10^4 other neurons, so the total number of connections is about 10^15. That suggests that the total amount of RAM needed for direct, brute-force modeling of a human brain (assuming we knew enough to program such a model, which we don't, and had parallel processors that could run such a simulation, which we don't) might be about 10^15 bytes, which is a 50-bit address space. A 64-bit address space is 16,000 times bigger than that.
I think we're likely to see flying cars, Turing-level AI, and vacations on the moon before we need 128-bit pointers.
The GPL doesn't prohibit you from doing this. You can write your GUI, license it under the GPL, and sell it. The GPL doesn't prohibit selling software. For instance, Red Hat sells GPL'd software.
Oh, wait, did you mean that you wanted your GUI to be closed source? Well, actually, I don't think the GPL prohibits you from doing that either. The license refers to "modified versions" of a work "based on" that work. I believe that this is interpreted to apply to linking to libraries, but not to executing programs via a shell. It's certainly not interpreted to apply to calling the linux kernel's interfaces, for example -- if it did, then it would be illegal to run closed-source code on a linux box.
All right, so did you actually mean that you wanted to take someone else's GPL'd code, link your code to it, and make your code closed source? Well in that case you're still off base. It's not the GPL that's preventing you from doing that, it's copyright law. The original author of the code didn't choose to give you permission to copy their code under the conditions you have in mind. If you're upset about that, you could contact the author and see if he's interested in being paid to offer his code to you under some other terms.
What I said here was correct:
What you said here was also correct:
The GPL distinguishes three different levels of things you can do with the software. "Executing it on a computer or modifying a private copy" doesn't require you to agree to the GPL at all. "Propagating" it requires you to agree to the GPL unless you think it falls under one of the exceptions to copyright law. "Conveying" it requires you to agree to the GPL, and imposes the highest requirements on you.