Has anybody looked into making a "guess what the JavaScript does" plugin? It would examine the JavaScript on a button or link and take a guess as to what the URL it is trying to open is, and open it.
Certainly I have been able to guess the URL by looking at one-line samples of Javascript. Is this possible in general? Would it be good enough to allow you to leave javascript off?
Your "IP" is not threatened. If you use GPL code in your program and then sell it without releasing the source, you are guilty of copyright violation. There is no precedence for somebody being guilty of copyright violation having to give up the rights to their own work.
What will happen is you will be required to cease distribution and pay a monetary fine, if anybody catches you and sues you. If you wish to redistribute the work you will have to remove the GPL portions and replace them with other code that you do have a license to redistribute.
It is true that many people are happy if you release your source code, and that could be used to avoid a lawsuit. But there is no requirement to do that, and it does not even get you out of being liable for the copyright violation (otherwise you could just release the source code after it is worthless due to age).
Furthermore, there is very little money-making opportunities on Linux that require you to use GPL code, in fact it is difficult to figure out something that could make money but requires GPL code, perhaps heavily modified kernels that can't be done with loadable modules?. Linux allows any program to run on it, and the vast majority of the libraries needed to make a working program are LGPL or BSD or similar (due to the fact that the library writers want their stuff used as much as possible).
Actually GIMP should take advantage of not having gone down the "use a short" root for pixels.
If there is going to be 16 bits per channel, it should use 16-bit floats (same format as used by ILM's exr image library and newer Nvidia cards). This gives immensely more resolution near black then 16 bits, and provides high dynamic range (ie colors outside the 0-1 range) which suddenly changes "color management" from a nightmare into an almost trivial operation.
Wasting time on a non-float format other than 8 bits is counter productive. Though they didn't plan it, it may be best that GIMP works the way it does.
The "transparency" you are complaining about will fix all the problems with the window updating. It is really double-buffering of the window image. Only completely finished windows will appear on the screen. So this does in fact address the problems you are having.
The other sluggishness is due to the seperate Window manager. Here the X people have got to learn. Window borders should be drawn by the application, not by a seperate process (Windows effectively does this by drawing the window borders in the kernel or in a library that you must pass through when doing the calls).
I am talking about legacy code. A huge Windows program in C++ can be converted to Linux as well, and I would bet the changes would be to the same 10-20% you mention. I fail to see how.net is going to improve this any.
Every time this is brought up people say "use #gtk" but I fail to see any difference between that and somebody saing "use gtk++" when writing a C++ program.
I can write code right now in C++ that compiles natively on both Windows and Linux. Microsoft and GNU both provide C++ compilers that match each other to an amazing degree. It is true that it is easy to write a C++ program that is Windows only (or to write a Linux-only one) as well.
The problem is I see *no* promise that C#/.net/mono is going to be better. I very, very, much doubt that a windows-only program is going to compile for Linux any easier than a Windows-only C++ program. Everytime they say "well use #gtk" I just don't see any possible advantage over C++. In C++ you can use fltk, gtk+, or Qt, or WxWindows, or lots of others. That does not help the program written using an obfuscated Microsoft-designed api (or an equally-obfuscated xlib). And I just cannot imagine how Mono will be any better, if it was there would not be any of this "use #gtk" statements.
Using a portable toolkit the GUI is differences are actually minor. I have FAR more difficulty with backslashes in pasted file names and with drive letters, with cut&paste to programs that barf if given filenames with forward slashes in them, with case-independent filenames making directory searches a nightmare and unportable, with local-time timestamps, and with CR+LF, and all the other ancient crap that still lives in Microsoft's crap operating system. Do you really thing those problems are going to go away, or is that horror just going to start invading Linux? Do you have any idea why there was such intense desire to get away from Microsoft? It has nothing to do with GUI, the real reasons are still there and just as blatent as ever.
Re:It's as if icons peaked 2-4 years ago
on
A History of Icons
·
· Score: 1
Good point. Even a man page that says "the docs are in/blah/foo/index.html" or even directed you to a public website, would be better than nothing.
By "vertical" I meant the text would look like this:
File
Edit
View
Whatever
It is quite incredible that somebody would think that I meant turning the graphic on it's side, or turning it on it's side and rotating the letters back. But I guess it proves my point that the horizontal menubar has brainwashed the masses into believing that no alternative is possible and thus they cannot even picture it!
Re:It's as if icons peaked 2-4 years ago
on
A History of Icons
·
· Score: 2, Interesting
Are you kidding? There are a lot of KDE/Gnome applications where picking help gets you irrelevant or useless information. I am afraid too many of the developers think it is ok if some HTML document appears.
Most recent example is in Gnome control panel, I picked "Sessions", and I got a little window with a "Help" button. I hit that "Help" button and I got a page "Part I Setting Appearance and Personal Preference...". This document did a great job of describing *some* of the items on the Control panel, and would perhaps be acceptable for a "Help" button *ON THE CONTROL PANEL* (which unfortunatly did not have a help button...). However it did not have a chapter on the "Sessions" popup, and I wasted time seeing if it was buried in one of the other chapters.
Suggested improvement to the GUI: if no help is written, pop up a message that says "no help is available". To be really clever, have it google for answers (that would actually work pretty good for Linux, too!). Don't pop up a page of other text.
I have to say it is incredibly annoying that "man" from 1970 still works better for getting information than any of these modern systems. With "man" I can quickly and unambigously determine if the information exists (so I don't waste time looking for nonexistent stuff), and if it does it always has the information I need (ie it lists *how to run that command*).
He is talking about "toolbars" as popularized by Microsoft software, not desktop icons. That idea is certainly Microsofts, whether it is good or bad. I do believe it is *way* overused, and am rather annoyed that they still have not figured out that there is no difference between the "menubar" and the "toolbar" and have failed to make them graphically match or merge them together. To be fair, Microsoft also came up with the popup tooltip that works (Apple's earlier version was too graphically intense and I believe had some other problems that made it hard to use as a quick reference), these tooltips make the "toolbar" as usable as a text bar and the screen realestate savings may be worth it. A vertical menubar with text would be far better, but you can probably blame Apple for convincing all the morons that a menu must be horizontal.
Both platforms are quite guilty of having inscrutable *appliation* icons, which I think you were talking about. This is mostly commercial software that insists on using their company logo rather than any representation of what it does. Microsoft has apparently given up, their nice looking modern icons use "W" to indicate Word, which is not really an icon, if you think about it!
How many times do people have to come up with crazy ideas like XML and Registries or even elektra.
It's easy: make a file for each key and put the literal value (not encrypted or encoded or anything) into the file. Want the "registry key" foo/bar/baz? Look in ~/configuration/foo/bar/baz, and if it is no there look in/configuration/foo/bar/baz. Read the contents of the file. DONE!
The filesystem already supports a totally arbitrary name hierarchy, and the data is allowed to be blocks of bytes of arbitrary size. YOU DO NOT NEED ANYTHING ELSE!
Comments could be supported by renaming the file baz to baz.comment. You could also document the expected structure of each directory with a.template file in the parent directory. Many other ideas could be added. The important thing is all this structure can be IGNORED by a simple program that just wants to read configuration.
Elektra is close, but the files are not raw data.
I don't understand why this solution is not absolutely obvious to everybody. What is the problem? Why is anybody proposing anything other than this?
I write a program from scratch. I place it under the GPL and sell binaries. Can I refuse to supply source?
Sounds like this is allowed. You are allowed to do anything you want to the source, including not distributing any new copies. The fact that you "stopped distributing new copies" before even a single one was distributed seems to fit. You certainly have not violated the GPL, as it is your copyright and it is impossible to violate your own copyright, and the GPL only grants extra methods of violating a copyright.
Of course it will be quickly pointed out that you really didn't GPL your code. If you continue to say it is GPL after a reasonable demonstration that it is not, perhaps you are liable for false advertising?
Since the company is not giving those employees the right to distribute the software outside the company they do not have to give the employees GPL rights to redistribute the source code either.
Your employer's an idiot and has bought into the FUD. They should be infintely more worried about hiring somebody who worked at another company before, since that person is also "infected" and this time with secret IP of that other company. GPL code is allowed to be read and examined, this employee is far, far worse, since even looking at the code is illegal.
In fact if "infection" was a legal worry, all software companies would have to hire everybody when they are about 3 years old and train them in-house on a special company-developed curriculum, with no ability to leave the premises or contact anybody outside.
I belive the original author can only sue for a copyright violation. Normally this is for monetary damages and a ceast & desist distribution order. I very much doubt any court is going to force a company to give up their ip.
There have been lots of cases where companies have voluntarily released their code to comply with the GPL. But this has only been due to threats of litigation which was then dropped, not from a court order. Generally they have judged the cost of releasing their code (often only trivial modifications, anyway) as much less than the legal costs, potential damages, and bad public-relations. So you could say that the GPL has made some organizations release code, but it hasn't really forced them.
It seems to me that his distributing the code GPL before being hired is irrelevant. *that* version remains GPL, but it has absolutly no bearing on what he can do with his own ip. He can modify the code and release the next version GPL, or public domain, or as closed source, or not release it at all.
Now it seems he has signed some legal agreement assigning all his rights to the code to the company. Therefore the company owns it. They can modify the code and release the next version GPL, or public domain, or as closed source, or not release it at all. He is now stuck, he cannot release it GPL, as it is not his.
*His* code is owned by Daimaou, as he signed the rights away to it. It does not matter that he may have released it GPL. Lots of FUD-masters want to push the fantasy that releasing something GPL somehow is giving up your rights to the code. It doesn't, you can change your mind and release all future versions under a different license or not at all. So the company can do anything they want with *his* code.
However the IBM (and probably other) GPL code is *not* his and thus could not be signed over to the company. Therefore for using that part of the code they are in violation of the GPL. They will have to cleanly seperate the two parts, if possible, or maybe go to IBM and find the original copyright holders and get their permission (perhaps with payment) to use the code in anything other than a GPL product.
The story author is totally screwed, and should have gotten something better than a verbal agreement. The company is also stuck unless they are willing to obey the GPL.
You can't write such a license, as that would be a contract, because it specifies restrictions that are not required by copyright law.
However you can write a GPL-like license that says "you must chew green gum while *redistributing* this software". Normally redistributing the code would violate copyright laws. However you have now stated that they may violate the copyright, *if* they chew green gum. But you have not said anything about redistributing it without chewing green gum, thus that is illegal because it is still a copyright violation.
However if they just write software using your code they are not violating copyright law. Unless you get them to sign a contract saying they will chew the green gum (in which case they are violating contract law), you have no power over them.
No it is not a contract. The GPL is a giving up by the copyright holder of *some* rights to the work. Public domain is a giving up of *all* rights to the work. For that reason, any argument that the GPL is somehow illegal would mean that public domain is also illegal.
Sorry, that is bogus. By your argument, public-domain work is illegal.
The GPL says "you can violate the copyright if you do X". It is not a restriction at all, because if the reciever does not do "X" then they are in the exact same situation as though there was no GPL on the code: they are bound by copyright. So they are not "forced" to do "X".
Claiming the author of a work cannot put such a "you can violate the copyright if you do this" rule on their work is nonsense. If it was illegal (like you claim) it would be impossible for any creation to be put in the public domain by the author. It would also mean the BSD license would be illegal. It would mean "copying for educational use only" is illegal. It would mean "you can copy this if you pay me a million dollars" is illegal. It just doesn't work that way, except in bizarro land.
If you assumme some device that stores one bit per atom on the surface of a crystal of, say, silicon, exactly what storage density are we talking about?
Assumme only a 2D array as I really suspect getting at the internal atoms of a cube will never happen. (Though it is likely these devices can be stacked so eventually there may be engineering done to make them as thin as possible...)
My guess is that this device is still many orders of magnitude away, but I really don't know.
Sorry I didn't mean to sound hostile. It is just from a Linux point of view, I am still very confused as to what this offers. It must offer something, but too much of the description is how to work around bad ideas in Windows. I would like to know how it really is superior to hav
In Unix it is assummed that a program is quite able to look in arbitrary places for it's configuration. The user has to set an environment variable that says "look here", but *that* can be set by another script. Even totally broken programs are fooled into running by symbolic links or by making a script to run it do the necessary rsync. In all cases you can make running a program lead to something the administrator can change on the central machine, so most of "administration" is "edit this file and save it". Now I'm the first to admit this leads to a spagetti mess of shell scripts, and cleaning it up would certainly be nice.
But it just seems the whole design of this remote administration is backwards from a Unix point of view. Even you admit that a big part of the design is because "many of the services are not available on Windows". On Unix you only "go to the other machine" to fix things when they are *broken*, meaning you are experimenting and thus a remote shell allowing arbitrary commands is the ideal interface. All automatic things are done by doing something *local* to the administration machine.
Some of the documentation I read described as "pull" verses "push". I guess then classic Unix administration is *all* "pull". The idea that anything needs "push" is foreign, and thus leading to the many questions people with Unix experience have about these systems.
Actually the only thing I really understand is the re-imaging code, but that is a "pull" system as well, in that it checks on reboot, and just-in-time, in that it gets the new image just before it uses it.
Slashdot Mirror
Has anybody looked into making a "guess what the JavaScript does" plugin? It would examine the JavaScript on a button or link and take a guess as to what the URL it is trying to open is, and open it.
Certainly I have been able to guess the URL by looking at one-line samples of Javascript. Is this possible in general? Would it be good enough to allow you to leave javascript off?
Your "IP" is not threatened. If you use GPL code in your program and then sell it without releasing the source, you are guilty of copyright violation. There is no precedence for somebody being guilty of copyright violation having to give up the rights to their own work.
What will happen is you will be required to cease distribution and pay a monetary fine, if anybody catches you and sues you. If you wish to redistribute the work you will have to remove the GPL portions and replace them with other code that you do have a license to redistribute.
It is true that many people are happy if you release your source code, and that could be used to avoid a lawsuit. But there is no requirement to do that, and it does not even get you out of being liable for the copyright violation (otherwise you could just release the source code after it is worthless due to age).
Furthermore, there is very little money-making opportunities on Linux that require you to use GPL code, in fact it is difficult to figure out something that could make money but requires GPL code, perhaps heavily modified kernels that can't be done with loadable modules?. Linux allows any program to run on it, and the vast majority of the libraries needed to make a working program are LGPL or BSD or similar (due to the fact that the library writers want their stuff used as much as possible).
Actually GIMP should take advantage of not having gone down the "use a short" root for pixels.
If there is going to be 16 bits per channel, it should use 16-bit floats (same format as used by ILM's exr image library and newer Nvidia cards). This gives immensely more resolution near black then 16 bits, and provides high dynamic range (ie colors outside the 0-1 range) which suddenly changes "color management" from a nightmare into an almost trivial operation.
Wasting time on a non-float format other than 8 bits is counter productive. Though they didn't plan it, it may be best that GIMP works the way it does.
How about a "package" where you double-click and it does the compile for you?
The "transparency" you are complaining about will fix all the problems with the window updating. It is really double-buffering of the window image. Only completely finished windows will appear on the screen. So this does in fact address the problems you are having.
The other sluggishness is due to the seperate Window manager. Here the X people have got to learn. Window borders should be drawn by the application, not by a seperate process (Windows effectively does this by drawing the window borders in the kernel or in a library that you must pass through when doing the calls).
I am talking about legacy code. A huge Windows program in C++ can be converted to Linux as well, and I would bet the changes would be to the same 10-20% you mention. I fail to see how .net is going to improve this any.
Every time this is brought up people say "use #gtk" but I fail to see any difference between that and somebody saing "use gtk++" when writing a C++ program.
"outlawing guns is no different, in principle, to laws saying murder should be illegal"
No matter where your opinion on gun control is, I hope you can see how nonsensical this statement is.
I can write code right now in C++ that compiles natively on both Windows and Linux. Microsoft and GNU both provide C++ compilers that match each other to an amazing degree. It is true that it is easy to write a C++ program that is Windows only (or to write a Linux-only one) as well.
The problem is I see *no* promise that C#/.net/mono is going to be better. I very, very, much doubt that a windows-only program is going to compile for Linux any easier than a Windows-only C++ program. Everytime they say "well use #gtk" I just don't see any possible advantage over C++. In C++ you can use fltk, gtk+, or Qt, or WxWindows, or lots of others. That does not help the program written using an obfuscated Microsoft-designed api (or an equally-obfuscated xlib). And I just cannot imagine how Mono will be any better, if it was there would not be any of this "use #gtk" statements.
Using a portable toolkit the GUI is differences are actually minor. I have FAR more difficulty with backslashes in pasted file names and with drive letters, with cut&paste to programs that barf if given filenames with forward slashes in them, with case-independent filenames making directory searches a nightmare and unportable, with local-time timestamps, and with CR+LF, and all the other ancient crap that still lives in Microsoft's crap operating system. Do you really thing those problems are going to go away, or is that horror just going to start invading Linux? Do you have any idea why there was such intense desire to get away from Microsoft? It has nothing to do with GUI, the real reasons are still there and just as blatent as ever.
Good point. Even a man page that says "the docs are in /blah/foo/index.html" or even directed you to a public website, would be better than nothing.
By "vertical" I meant the text would look like this:
File
Edit
View
Whatever
It is quite incredible that somebody would think that I meant turning the graphic on it's side, or turning it on it's side and rotating the letters back. But I guess it proves my point that the horizontal menubar has brainwashed the masses into believing that no alternative is possible and thus they cannot even picture it!
Are you kidding? There are a lot of KDE/Gnome applications where picking help gets you irrelevant or useless information. I am afraid too many of the developers think it is ok if some HTML document appears.
Most recent example is in Gnome control panel, I picked "Sessions", and I got a little window with a "Help" button. I hit that "Help" button and I got a page "Part I Setting Appearance and Personal Preference...". This document did a great job of describing *some* of the items on the Control panel, and would perhaps be acceptable for a "Help" button *ON THE CONTROL PANEL* (which unfortunatly did not have a help button...). However it did not have a chapter on the "Sessions" popup, and I wasted time seeing if it was buried in one of the other chapters.
Suggested improvement to the GUI: if no help is written, pop up a message that says "no help is available". To be really clever, have it google for answers (that would actually work pretty good for Linux, too!). Don't pop up a page of other text.
I have to say it is incredibly annoying that "man" from 1970 still works better for getting information than any of these modern systems. With "man" I can quickly and unambigously determine if the information exists (so I don't waste time looking for nonexistent stuff), and if it does it always has the information I need (ie it lists *how to run that command*).
He is talking about "toolbars" as popularized by Microsoft software, not desktop icons. That idea is certainly Microsofts, whether it is good or bad. I do believe it is *way* overused, and am rather annoyed that they still have not figured out that there is no difference between the "menubar" and the "toolbar" and have failed to make them graphically match or merge them together. To be fair, Microsoft also came up with the popup tooltip that works (Apple's earlier version was too graphically intense and I believe had some other problems that made it hard to use as a quick reference), these tooltips make the "toolbar" as usable as a text bar and the screen realestate savings may be worth it. A vertical menubar with text would be far better, but you can probably blame Apple for convincing all the morons that a menu must be horizontal.
Both platforms are quite guilty of having inscrutable *appliation* icons, which I think you were talking about. This is mostly commercial software that insists on using their company logo rather than any representation of what it does. Microsoft has apparently given up, their nice looking modern icons use "W" to indicate Word, which is not really an icon, if you think about it!
Hurry up before they patent the business model!
How many times do people have to come up with crazy ideas like XML and Registries or even elektra.
/configuration/foo/bar/baz. Read the contents of the file. DONE!
.template file in the parent directory. Many other ideas could be added. The important thing is all this structure can be IGNORED by a simple program that just wants to read configuration.
It's easy: make a file for each key and put the literal value (not encrypted or encoded or anything) into the file. Want the "registry key" foo/bar/baz? Look in ~/configuration/foo/bar/baz, and if it is no there look in
The filesystem already supports a totally arbitrary name hierarchy, and the data is allowed to be blocks of bytes of arbitrary size. YOU DO NOT NEED ANYTHING ELSE!
Comments could be supported by renaming the file baz to baz.comment. You could also document the expected structure of each directory with a
Elektra is close, but the files are not raw data.
I don't understand why this solution is not absolutely obvious to everybody. What is the problem? Why is anybody proposing anything other than this?
I write a program from scratch. I place it under the GPL and sell binaries. Can I refuse to supply source?
Sounds like this is allowed. You are allowed to do anything you want to the source, including not distributing any new copies. The fact that you "stopped distributing new copies" before even a single one was distributed seems to fit. You certainly have not violated the GPL, as it is your copyright and it is impossible to violate your own copyright, and the GPL only grants extra methods of violating a copyright.
Of course it will be quickly pointed out that you really didn't GPL your code. If you continue to say it is GPL after a reasonable demonstration that it is not, perhaps you are liable for false advertising?
Since the company is not giving those employees the right to distribute the software outside the company they do not have to give the employees GPL rights to redistribute the source code either.
Your employer's an idiot and has bought into the FUD. They should be infintely more worried about hiring somebody who worked at another company before, since that person is also "infected" and this time with secret IP of that other company. GPL code is allowed to be read and examined, this employee is far, far worse, since even looking at the code is illegal.
In fact if "infection" was a legal worry, all software companies would have to hire everybody when they are about 3 years old and train them in-house on a special company-developed curriculum, with no ability to leave the premises or contact anybody outside.
I belive the original author can only sue for a copyright violation. Normally this is for monetary damages and a ceast & desist distribution order. I very much doubt any court is going to force a company to give up their ip.
There have been lots of cases where companies have voluntarily released their code to comply with the GPL. But this has only been due to threats of litigation which was then dropped, not from a court order. Generally they have judged the cost of releasing their code (often only trivial modifications, anyway) as much less than the legal costs, potential damages, and bad public-relations. So you could say that the GPL has made some organizations release code, but it hasn't really forced them.
It seems to me that his distributing the code GPL before being hired is irrelevant. *that* version remains GPL, but it has absolutly no bearing on what he can do with his own ip. He can modify the code and release the next version GPL, or public domain, or as closed source, or not release it at all.
Now it seems he has signed some legal agreement assigning all his rights to the code to the company. Therefore the company owns it. They can modify the code and release the next version GPL, or public domain, or as closed source, or not release it at all. He is now stuck, he cannot release it GPL, as it is not his.
*His* code is owned by Daimaou, as he signed the rights away to it. It does not matter that he may have released it GPL. Lots of FUD-masters want to push the fantasy that releasing something GPL somehow is giving up your rights to the code. It doesn't, you can change your mind and release all future versions under a different license or not at all. So the company can do anything they want with *his* code.
However the IBM (and probably other) GPL code is *not* his and thus could not be signed over to the company. Therefore for using that part of the code they are in violation of the GPL. They will have to cleanly seperate the two parts, if possible, or maybe go to IBM and find the original copyright holders and get their permission (perhaps with payment) to use the code in anything other than a GPL product.
The story author is totally screwed, and should have gotten something better than a verbal agreement. The company is also stuck unless they are willing to obey the GPL.
You can't write such a license, as that would be a contract, because it specifies restrictions that are not required by copyright law.
However you can write a GPL-like license that says "you must chew green gum while *redistributing* this software". Normally redistributing the code would violate copyright laws. However you have now stated that they may violate the copyright, *if* they chew green gum. But you have not said anything about redistributing it without chewing green gum, thus that is illegal because it is still a copyright violation.
However if they just write software using your code they are not violating copyright law. Unless you get them to sign a contract saying they will chew the green gum (in which case they are violating contract law), you have no power over them.
No it is not a contract. The GPL is a giving up by the copyright holder of *some* rights to the work. Public domain is a giving up of *all* rights to the work. For that reason, any argument that the GPL is somehow illegal would mean that public domain is also illegal.
Sorry, that is bogus. By your argument, public-domain work is illegal.
The GPL says "you can violate the copyright if you do X". It is not a restriction at all, because if the reciever does not do "X" then they are in the exact same situation as though there was no GPL on the code: they are bound by copyright. So they are not "forced" to do "X".
Claiming the author of a work cannot put such a "you can violate the copyright if you do this" rule on their work is nonsense. If it was illegal (like you claim) it would be impossible for any creation to be put in the public domain by the author. It would also mean the BSD license would be illegal. It would mean "copying for educational use only" is illegal. It would mean "you can copy this if you pay me a million dollars" is illegal. It just doesn't work that way, except in bizarro land.
If you assumme some device that stores one bit per atom on the surface of a crystal of, say, silicon, exactly what storage density are we talking about?
Assumme only a 2D array as I really suspect getting at the internal atoms of a cube will never happen. (Though it is likely these devices can be stacked so eventually there may be engineering done to make them as thin as possible...)
My guess is that this device is still many orders of magnitude away, but I really don't know.
Sorry I didn't mean to sound hostile. It is just from a Linux point of view, I am still very confused as to what this offers. It must offer something, but too much of the description is how to work around bad ideas in Windows. I would like to know how it really is superior to hav
In Unix it is assummed that a program is quite able to look in arbitrary places for it's configuration. The user has to set an environment variable that says "look here", but *that* can be set by another script. Even totally broken programs are fooled into running by symbolic links or by making a script to run it do the necessary rsync. In all cases you can make running a program lead to something the administrator can change on the central machine, so most of "administration" is "edit this file and save it". Now I'm the first to admit this leads to a spagetti mess of shell scripts, and cleaning it up would certainly be nice.
But it just seems the whole design of this remote administration is backwards from a Unix point of view. Even you admit that a big part of the design is because "many of the services are not available on Windows". On Unix you only "go to the other machine" to fix things when they are *broken*, meaning you are experimenting and thus a remote shell allowing arbitrary commands is the ideal interface. All automatic things are done by doing something *local* to the administration machine.
Some of the documentation I read described as "pull" verses "push". I guess then classic Unix administration is *all* "pull". The idea that anything needs "push" is foreign, and thus leading to the many questions people with Unix experience have about these systems.
Actually the only thing I really understand is the re-imaging code, but that is a "pull" system as well, in that it checks on reboot, and just-in-time, in that it gets the new image just before it uses it.