I'm not sure if this is maintaining legacy apps, but it certainly scared me!
Python jobs are hardly for legacy app maintenance. More like rapid development of cutting edge stuff, prototyping, exploring, enterprise application integration... and Agile development in general. I introduced Python to my previous workplace, and after the guys there learned it, they didn't switch back (even though their chief python advocate/fascist, i.e. your truly, left:-).
Python can be used for very large problems (hundreds of modules, and much more classes), in addition to trivial scripts (0 functions). It is *fun* as hell. Python programmer is always an architect, there is very little monkey-level "grunt work", which tends to form most of your day-to-day C++/Java programming.
You really have no clue about OOP before you have tried one of the dynamic OOP languages: Python, Smalltalk, or Ruby. Smalltalk has fallen to a legacy role these days, while Ruby is much less mature and has a smaller community than Python. Additionally, Ruby is less "tasteful", in that it borrows more heavily from perl, but that is a matter of controversy;-).
Additionally, Python is an embodiment of Open Source, because the code is actually readable and concise enough to lower the barrier of reading it. In fact I have taken a look at the source code of several Open Source projects that use Python "just for kicks", while I hardly bother in case of e.g. C programs. One line of Python is equivalent of 10-20 lines of C++, so you can digest more with the typical geek attention span (i.e. borderline ADD;-).
Debian should release a stable SERVER subsystem, then build a rapidly improving desktop subsystem that remains compatible with the *stable* server subsystem. Kinda like the UnitedLinux idea, which isn't all that bad. People can tolerate when their desktop apps crash every now and again if their server side is rock solid, as we have come to expect from debian Stable. That server subsystem could also be a basis for various Debian derivatives, commercial and non-commercial.
An example release could be "Debian 4.2, based on Debian_base_3.4"
they don't have the money to carry on this long litigation.
Dunno. Boies is not paid by the hour, but by a portion of their hypothetical winnings from this case. They can stretch it forever, and the money pumped in by MSFT and (possibly) Sun isn't hurting.
If there is any justice in the world:
1) SCO will be no more after this is over
2) McBride and Sonntag will be serving jail time in a maximum security penitentiary. (ah, well, one can always dream)
Calling them morons is kinda missing the point; they value the HP and MS partnerships. Nothing else is under discussion for the current project.
But they could as well use Linux via their HP partnership - that's what we did. I only gave suggestions as to what ways of persuasion your people could use... Nobody likes to be the ignorant victim who is "being had", the way HP-UX customers are.
Home office: We are Microsoft and HP partners. We will not be using Linux.
Tell them that HP-UX is looking into an open grave, along with the PA-RISC architecture (Linux will 0wn Itanium). Tell them that migration is unevitable, and just a matter of timing. Tell them that Red Hat AS costs $1500, if they feel like paying money for support. And yes, you can buy the RHAS support from HP now, that's what my old company did.
Besides, HP is very serious about Linux - especially the people from Compaq side, which grok the "industry standard" thing. Some factions at HP will enthusiastically sell HP-UX to any moron that is buying, though. Wanna be one of those?
The point is that credit should be given where credit is due.
But why keep bringing it up everytime? It mostly serves to alienate people, and create an atmosphere of "conflict". FSF would get (even) more respect if they chose to give a little bit friendlier image.
As I said before, the two groups happen to agree with eachother on 99% of practical recommendations (so far).
Indeed, and I assume nobody has any problems with the practical recommendations of FSF. It's the ideological rantings that prove problematic to many, especially people who make decisions. I "sold" Linux to my company (replacing a proprietary Unix) by explaining how Linux is becoming the standard Unix (and thus the one with the best support and apps) and proprietary ones will be doomed to be just legacy, and showing some SPEC numbers and price comparisons. Explaining the FSF party line would have got me laughed out of the office.
The FSF keeps whining about GNU and Linux terminology is starting to get tiresome. They always bring up the topic, no matter what the context is. Additionally, they occasionally seem like they would be ready to throw linux-the-kernel to the wolves in order to boost Hurd. They just don't seem to realize that damage to linux would be a huge blow to free software overall, since it would slow the development and nullify a lot of the momentum open source OS has.
Linux is attacked because it has the most marketplace momentum. If *BSD was at the same position, it would be attacked as well. Microsoft, Sun and the likes would just love to see hobbyists dabbling with *BSD, Hurd and other environments with no significant market impact (that your boss would call "toy" operating systems). *BSD is not a realistic migration path until it has mature & solid SMP, and driver support.
I dunno, I used to respect FSF, but nowadays the term "Free Software" (especially when capitalized) bears mostly negative connotations. It always brings to mind RMS, whining how he hasn't got all the respect he thinks he deserves. Let's face it, Linux could do just fine without GNU, apart from gcc. Userland utilities could be nicked from the BSD's, at least. FSF should get their act together, stop begging for respect/acknowledgement and realize that Open Source is what is going to bring out the world of truly Open Systems.
Insisting that everything should be Free is not very realistic either, as we have to live in the real world where non-idealists call the shots. Linux is a great, scalable system and other OSS components provide an open infrastructure that you can rely. Let people build their business on top of that.
We must just be happy that there is the Open Source movement to make all this stuff sellable to corporations (so that you can use a decent OS at workplace also, instead of just your home).
To find these involved going to www.suse.com, clicking on "downloads". I'm sure the kids can handle this fine.
Not w/o fixed-price network connection. And I doubt they are willing to "evaluate" anything, as long as they can get the "real thing" from other distributors.
Actually, SuSE doesn't ship ISO's for free, so kids can't get legal SuSE even now (as CD images, that is). This is also the reason so many have been unable to try out SuSE (including your truly - had I been able to try it for free, I might have recommended it for my company over Red Hat. It's easier to spend company money than your own).
I think this arrangement is pretty lame. People should be able to get SuSE ISO's for home use for free, while charging for corporate desktops and SLES.
How did Microsoft's agreement to pay you for Unix rights happen? Darl: In the Microsoft case, they saw an opportunity. We originally approached them and said we're on a new licensing path; we have this intellectual property that we've started approaching vendors about. IBM is one we approached; Microsoft was another. We had about four big vendors in the last quarter that we talked with. With two of them, we signed deals. The other we're still talking with, and IBM we reached an impasse.
To me it feels like they are still talking with HP, and Sun decided to pay up to take a stab at linux (in the back, I might as well say). Or is there any other interpretation? Was anyone surprised at how quick Sun was to advertise that they are in the clear?
Also, SCO has said that Sun is the only company that is clear of all the violations. Even M$ is less clear.
I hope someone brings this up in an interview with Scott, so he can explicitly deny this if it is untrue.
Whilst I haven't done any background checking on the Initiative for Software Choice I think what they seem to be saying is that the best tool for any particular job should be picked
Now there's an idea! "Best tool for a job" is a tired cliche with not much basis in reality. "Best" is always determined by many things (knowledge, future plans, price etc.) and is wildly subjective in any case. And sometimes openness of file formats is what makes a tool "the best for the job", even if it lacked some less important features.
Thing is that you just reiterated for N:th time the fact that someone has paid for SCO and after that you started speculating it might be Sun.
This time there was evidence that might point to that direction. Read the interview I linked to. It is a known fact that SCO has one anonymous funder in this crusade, and this time one is able to infer from big-mouth-Darl's words that it may be Sun.
BTW, why are you being so aggressive denying that allegiation?
If you had *any* proof about this payer you post would have been interesting, now it is just genuine FUD and karmawhoring. But./ is famous of that.
Not everybody needs to karmawhore. I've got enough karma, so getting modded up is meaningless apart from improved visibility for my point (and +3 is just enough for that). I just hope some net journalist takes up the article I linked to and dissects it thoroughly.
So, that leaves HP (or dell) as the anonymous supporter. I find both of them less probable to support SCO. The fact remains that Microsoft and Sun stand to gain the most from this show, and it would not be unthinkable that they are the ones funding it.
And even if Sun got the IP from Novell, they might still choose to throw some money at SCO.
Again, I'm expecting that some journalist asks Sun whether they are the anonymous license buyer, and if they say no, great. If they decline to comment, well...
Care to elaborate why it is a flamebait? Because it says something negative about your employer?
One has to be pretty blind/naive not to think of Sun in this context. I would love to hear Scott@Sun explicitly deny this. It is not a nice thought (Sun backstabbing us and all...), and I would like nothing more than being proven false.
Felt like saying this close to the top of the article... From here:
How did Microsoft's agreement to pay you for Unix rights happen? Darl: In the Microsoft case, they saw an opportunity. We originally approached them and said we're on a new licensing path; we have this intellectual property that we've started approaching vendors about. IBM is one we approached; Microsoft was another. We had about four big vendors in the last quarter that we talked with. With two of them, we signed deals. The other we're still talking with, and IBM we reached an impasse.
To me it feels like they are still talking with HP, and Sun decided to pay up to take a stab at linux (in the back, I might as well say). Or is there any other interpretation? Was anyone surprised at how quick Sun was to advertise that they are in the clear?
Boy, these Sun people don't seem like such friends of ours after all.
And I'm not talking about Web servers, but heavy database work, HPC etc. We are evidencing an era where proprietary Unix systems are brought down from their pedestal, and having good performance figures can't hurt.
Your mom will also like it, what with all the video&image editing and stuff.
Why is it that every time an increase in computing performance is reported, Slashdot is full of people whining why they don't need it.
Re:The people who build Linux don't want SCO's cod
on
What if SCO is Right?
·
· Score: 1
They only have something like two years to go anyway... less so now that their expenses are up (lawyers) and profits are down (nobody, especially the knowledgeable unix people, will ever again by anything from them).
You haven't "accidentally GPL'ed your software". You've only violated the license.
Ok. So instead of having to GPL the library, you get a cease-and-desist from TrollTech? Or can they outright sue you? If it is only a matter of withdrawing the app and fixing the problematics parts, good.
Why would that be? As the copyright holder, you can change the license any time you want. You can start it as GPL when it's in-house, and change the license later if you want to sell it outside of your company.
Not so with QT. Can't remeber the exact terms, but that procedure is explicitly forbidden. If you start it as a GPL project using QT, you can't just change license. You have to buy the QT license and develop a new app from scratch. I can't even begin to think about how ridiculous that must feel for the unfortunate programmer that receives the task, and to what extend copy-paste is allowed in such an endeavour.
You have a proprietary library. You develop an application, that can as well be open source, or proprietary, you don't care. So you use QT, because it appears to do the job well. At some point in time, someone thinks that a feature from the proprietary library might be handy for the app, and you link that library to the app, thinking that it doesn't matter, you just ship a closed source version. The application is shipped to the client, client sees that it should be GPL due to QT being used (note that you can't buy the QT license and make the app closed-source afterwards - QT doesn't allow that). Client requires the sources to the proprietary library and rights to distribute it under GPL.
accidently GPLize?, sounds like a lot of laziness, and a dash of incompetance
These things can happen when people are not watchful (or competent) and time is in short supply. With LGPL & GTK, this is simply not an issue at all.
GPL is greatest thing since sliced bread for layers of infrastructure that you don't have to link against (OS's, apps). Not so good for libraries.
if your employer cant make up its mind of whether to release it as a free or commercial app, BEFORE development commences, your employer has larger issues to worry about
Occasionally people will just play around with something that might become useful in a open source OR closed source app, without knowing in advance. I want to know that if I create something useful, I am free to use it in my day job and hobby alike. With QT I don't have that option.
QT is probably very good for what it is, but for this reason it can never become the de facto standard of Linux GUI development. GTK can, and quite probably will.
KDE is indeed very polished, snappy and comfortable, arguably more so than Gnome (apart from Red Hat's excellent Gnome). However, Gnome & GTK is a more future-proof *platform*, since you can develop a toy application with it, and if it is succesful, you can release it with whatever license you/your employer wants to use. With KDE & QT, your application will only be GPL, unless you cough up the money for QT license *before* you start developing your app.
For example, I develop Python applications in my current job. There are some python libraries that can't be released under GPL, by any means (the will of the company, not mine). In those cases, I just can't import those libraries when I develop a GUI application if I use PyQT. However, with PyGTK, I can release anything I want with any license I want.
So, the main point is that even if your application could be GPL, all the libs that the application would use can't necessarily be so. Of course one can use CORBA etc. the insulate the non-GPL portions, but it's a drag and I'd much rather use GTK. The code that uses GTK can be deployed everywhere without worries, with QT you have to keep vigilant that you don't accidentally GPL'ize anything.
In my view a library is not a "commodity" until its use is absolutely free of strings. That's the reason I avoid proprietary libraries, and GPL libraries. Liberate the infrastructure!
We're talking about what's on the periphery of the Linux kernel.
It would be interesting to know what parts in the periphery boost linux enough to get it from "bicycle" status to a full fledged enterprise unix. sysvinit? tuxracer? Nah, neither of them is by IBM...
Slashdot Mirror
I am not sure what you mean by Ruby being "less mature"
Less modules, less community support. I also seem to remember that they only recently got kernel threads, though I might be wrong.
I'm not sure if this is maintaining legacy apps, but it certainly scared me!
:-).
;-).
;-).
Python jobs are hardly for legacy app maintenance. More like rapid development of cutting edge stuff, prototyping, exploring, enterprise application integration... and Agile development in general. I introduced Python to my previous workplace, and after the guys there learned it, they didn't switch back (even though their chief python advocate/fascist, i.e. your truly, left
Python can be used for very large problems (hundreds of modules, and much more classes), in addition to trivial scripts (0 functions). It is *fun* as hell. Python programmer is always an architect, there is very little monkey-level "grunt work", which tends to form most of your day-to-day C++/Java programming.
You really have no clue about OOP before you have tried one of the dynamic OOP languages: Python, Smalltalk, or Ruby. Smalltalk has fallen to a legacy role these days, while Ruby is much less mature and has a smaller community than Python. Additionally, Ruby is less "tasteful", in that it borrows more heavily from perl, but that is a matter of controversy
Additionally, Python is an embodiment of Open Source, because the code is actually readable and concise enough to lower the barrier of reading it. In fact I have taken a look at the source code of several Open Source projects that use Python "just for kicks", while I hardly bother in case of e.g. C programs. One line of Python is equivalent of 10-20 lines of C++, so you can digest more with the typical geek attention span (i.e. borderline ADD
Debian should release a stable SERVER subsystem, then build a rapidly improving desktop subsystem that remains compatible with the *stable* server subsystem. Kinda like the UnitedLinux idea, which isn't all that bad. People can tolerate when their desktop apps crash every now and again if their server side is rock solid, as we have come to expect from debian Stable. That server subsystem could also be a basis for various Debian derivatives, commercial and non-commercial.
An example release could be "Debian 4.2, based on Debian_base_3.4"
they don't have the money to carry on this long litigation.
Dunno. Boies is not paid by the hour, but by a portion of their hypothetical winnings from this case. They can stretch it forever, and the money pumped in by MSFT and (possibly) Sun isn't hurting.
If there is any justice in the world:
1) SCO will be no more after this is over
2) McBride and Sonntag will be serving jail time in a maximum security penitentiary. (ah, well, one can always dream)
Calling them morons is kinda missing the point; they value the HP and MS partnerships. Nothing else is under discussion for the current project.
But they could as well use Linux via their HP partnership - that's what we did. I only gave suggestions as to what ways of persuasion your people could use... Nobody likes to be the ignorant victim who is "being had", the way HP-UX customers are.
Home office: We are Microsoft and HP partners. We will not be using Linux.
Tell them that HP-UX is looking into an open grave, along with the PA-RISC architecture (Linux will 0wn Itanium). Tell them that migration is unevitable, and just a matter of timing. Tell them that Red Hat AS costs $1500, if they feel like paying money for support. And yes, you can buy the RHAS support from HP now, that's what my old company did.
Besides, HP is very serious about Linux - especially the people from Compaq side, which grok the "industry standard" thing. Some factions at HP will enthusiastically sell HP-UX to any moron that is buying, though. Wanna be one of those?
The point is that credit should be given where credit is due.
But why keep bringing it up everytime? It mostly serves to alienate people, and create an atmosphere of "conflict". FSF would get (even) more respect if they chose to give a little bit friendlier image.
As I said before, the two groups happen to agree with eachother on 99% of practical recommendations (so far).
Indeed, and I assume nobody has any problems with the practical recommendations of FSF. It's the ideological rantings that prove problematic to many, especially people who make decisions. I "sold" Linux to my company (replacing a proprietary Unix) by explaining how Linux is becoming the standard Unix (and thus the one with the best support and apps) and proprietary ones will be doomed to be just legacy, and showing some SPEC numbers and price comparisons. Explaining the FSF party line would have got me laughed out of the office.
The FSF keeps whining about GNU and Linux terminology is starting to get tiresome. They always bring up the topic, no matter what the context is. Additionally, they occasionally seem like they would be ready to throw linux-the-kernel to the wolves in order to boost Hurd. They just don't seem to realize that damage to linux would be a huge blow to free software overall, since it would slow the development and nullify a lot of the momentum open source OS has.
Linux is attacked because it has the most marketplace momentum. If *BSD was at the same position, it would be attacked as well. Microsoft, Sun and the likes would just love to see hobbyists dabbling with *BSD, Hurd and other environments with no significant market impact (that your boss would call "toy" operating systems). *BSD is not a realistic migration path until it has mature & solid SMP, and driver support.
I dunno, I used to respect FSF, but nowadays the term "Free Software" (especially when capitalized) bears mostly negative connotations. It always brings to mind RMS, whining how he hasn't got all the respect he thinks he deserves. Let's face it, Linux could do just fine without GNU, apart from gcc. Userland utilities could be nicked from the BSD's, at least. FSF should get their act together, stop begging for respect/acknowledgement and realize that Open Source is what is going to bring out the world of truly Open Systems.
Insisting that everything should be Free is not very realistic either, as we have to live in the real world where non-idealists call the shots. Linux is a great, scalable system and other OSS components provide an open infrastructure that you can rely. Let people build their business on top of that.
We must just be happy that there is the Open Source movement to make all this stuff sellable to corporations (so that you can use a decent OS at workplace also, instead of just your home).
To find these involved going to www.suse.com, clicking on "downloads". I'm sure the kids can handle this fine.
Not w/o fixed-price network connection. And I doubt they are willing to "evaluate" anything, as long as they can get the "real thing" from other distributors.
pirate copies of the latest SuSE
Actually, SuSE doesn't ship ISO's for free, so kids can't get legal SuSE even now (as CD images, that is). This is also the reason so many have been unable to try out SuSE (including your truly - had I been able to try it for free, I might have recommended it for my company over Red Hat. It's easier to spend company money than your own).
I think this arrangement is pretty lame. People should be able to get SuSE ISO's for home use for free, while charging for corporate desktops and SLES.
I'm reposting this, since the previous posting was to an article that was pretty dead already .
From McBride interview at ZDNet:
How did Microsoft's agreement to pay you for Unix rights happen?
Darl: In the Microsoft case, they saw an opportunity. We originally approached them and said we're on a new licensing path; we have this intellectual property that we've started approaching vendors about. IBM is one we approached; Microsoft was another. We had about four big vendors in the last quarter that we talked with. With two of them, we signed deals. The other we're still talking with, and IBM we reached an impasse.
To me it feels like they are still talking with HP, and Sun decided to pay up to take a stab at linux (in the back, I might as well say). Or is there any other interpretation? Was anyone surprised at how quick Sun was to advertise that they are in the clear?
Also, SCO has said that Sun is the only company that is clear of all the violations. Even M$ is less clear.
I hope someone brings this up in an interview with Scott, so he can explicitly deny this if it is untrue.
Whilst I haven't done any background checking on the Initiative for Software Choice I think what they seem to be saying is that the best tool for any particular job should be picked
Now there's an idea! "Best tool for a job" is a tired cliche with not much basis in reality. "Best" is always determined by many things (knowledge, future plans, price etc.) and is wildly subjective in any case. And sometimes openness of file formats is what makes a tool "the best for the job", even if it lacked some less important features.
Thing is that you just reiterated for N:th time the fact that someone has paid for SCO and after that you started speculating it might be Sun.
./ is famous of that.
This time there was evidence that might point to that direction. Read the interview I linked to. It is a known fact that SCO has one anonymous funder in this crusade, and this time one is able to infer from big-mouth-Darl's words that it may be Sun.
BTW, why are you being so aggressive denying that allegiation?
If you had *any* proof about this payer you post would have been interesting, now it is just genuine FUD and karmawhoring. But
Not everybody needs to karmawhore. I've got enough karma, so getting modded up is meaningless apart from improved visibility for my point (and +3 is just enough for that). I just hope some net journalist takes up the article I linked to and dissects it thoroughly.
So, that leaves HP (or dell) as the anonymous supporter. I find both of them less probable to support SCO. The fact remains that Microsoft and Sun stand to gain the most from this show, and it would not be unthinkable that they are the ones funding it.
And even if Sun got the IP from Novell, they might still choose to throw some money at SCO.
Again, I'm expecting that some journalist asks Sun whether they are the anonymous license buyer, and if they say no, great. If they decline to comment, well...
Care to elaborate why it is a flamebait? Because it says something negative about your employer?
One has to be pretty blind/naive not to think of Sun in this context. I would love to hear Scott@Sun explicitly deny this. It is not a nice thought (Sun backstabbing us and all...), and I would like nothing more than being proven false.
Felt like saying this close to the top of the article... From here:
How did Microsoft's agreement to pay you for Unix rights happen?
Darl: In the Microsoft case, they saw an opportunity. We originally approached them and said we're on a new licensing path; we have this intellectual property that we've started approaching vendors about. IBM is one we approached; Microsoft was another. We had about four big vendors in the last quarter that we talked with. With two of them, we signed deals. The other we're still talking with, and IBM we reached an impasse.
To me it feels like they are still talking with HP, and Sun decided to pay up to take a stab at linux (in the back, I might as well say). Or is there any other interpretation? Was anyone surprised at how quick Sun was to advertise that they are in the clear?
Boy, these Sun people don't seem like such friends of ours after all.
Faster CPU's = More free time for 'Working' Nerds?
I dunno, having a long-running compile is a great time to refresh slashdot.
People running serious server-side stuff can.
And I'm not talking about Web servers, but heavy database work, HPC etc. We are evidencing an era where proprietary Unix systems are brought down from their pedestal, and having good performance figures can't hurt.
Your mom will also like it, what with all the video&image editing and stuff.
Why is it that every time an increase in computing performance is reported, Slashdot is full of people whining why they don't need it.
They only have something like two years to go anyway... less so now that their expenses are up (lawyers) and profits are down (nobody, especially the knowledgeable unix people, will ever again by anything from them).
You haven't "accidentally GPL'ed your software". You've only violated the license.
:-).
Ok. So instead of having to GPL the library, you get a cease-and-desist from TrollTech? Or can they outright sue you? If it is only a matter of withdrawing the app and fixing the problematics parts, good.
Still, I prefer the lawyer-free going of LGPL
Why would that be? As the copyright holder, you can change the license any time you want. You can start it as GPL when it's in-house, and change the license later if you want to sell it outside of your company.
Not so with QT. Can't remeber the exact terms, but that procedure is explicitly forbidden. If you start it as a GPL project using QT, you can't just change license. You have to buy the QT license and develop a new app from scratch. I can't even begin to think about how ridiculous that must feel for the unfortunate programmer that receives the task, and to what extend copy-paste is allowed in such an endeavour.
how do you accidently GPL anything?
You have a proprietary library. You develop an application, that can as well be open source, or proprietary, you don't care. So you use QT, because it appears to do the job well. At some point in time, someone thinks that a feature from the proprietary library might be handy for the app, and you link that library to the app, thinking that it doesn't matter, you just ship a closed source version. The application is shipped to the client, client sees that it should be GPL due to QT being used (note that you can't buy the QT license and make the app closed-source afterwards - QT doesn't allow that). Client requires the sources to the proprietary library and rights to distribute it under GPL.
accidently GPLize?, sounds like a lot of laziness, and a dash of incompetance
These things can happen when people are not watchful (or competent) and time is in short supply. With LGPL & GTK, this is simply not an issue at all.
GPL is greatest thing since sliced bread for layers of infrastructure that you don't have to link against (OS's, apps). Not so good for libraries.
if your employer cant make up its mind of whether to release it as a free or commercial app, BEFORE development commences, your employer has larger issues to worry about
Occasionally people will just play around with something that might become useful in a open source OR closed source app, without knowing in advance. I want to know that if I create something useful, I am free to use it in my day job and hobby alike. With QT I don't have that option.
QT is probably very good for what it is, but for this reason it can never become the de facto standard of Linux GUI development. GTK can, and quite probably will.
KDE is indeed very polished, snappy and comfortable, arguably more so than Gnome (apart from Red Hat's excellent Gnome). However, Gnome & GTK is a more future-proof *platform*, since you can develop a toy application with it, and if it is succesful, you can release it with whatever license you/your employer wants to use. With KDE & QT, your application will only be GPL, unless you cough up the money for QT license *before* you start developing your app.
For example, I develop Python applications in my current job. There are some python libraries that can't be released under GPL, by any means (the will of the company, not mine). In those cases, I just can't import those libraries when I develop a GUI application if I use PyQT. However, with PyGTK, I can release anything I want with any license I want.
So, the main point is that even if your application could be GPL, all the libs that the application would use can't necessarily be so. Of course one can use CORBA etc. the insulate the non-GPL portions, but it's a drag and I'd much rather use GTK. The code that uses GTK can be deployed everywhere without worries, with QT you have to keep vigilant that you don't accidentally GPL'ize anything.
In my view a library is not a "commodity" until its use is absolutely free of strings. That's the reason I avoid proprietary libraries, and GPL libraries. Liberate the infrastructure!
This should be told also to the jury in the SCO trial.
We're talking about what's on the periphery of the Linux kernel.
It would be interesting to know what parts in the periphery boost linux enough to get it from "bicycle" status to a full fledged enterprise unix. sysvinit? tuxracer? Nah, neither of them is by IBM...