No, Cairo is more like a new graphics api for drawing shapes and text and lines, etc. It is similar to Quartz, or Display PostScript. Both GTK and Qt on OS/X use Quartz to draw everything, they will probably both use this api to draw anything on X once it is commonly installed. However this will not change the API to GTK or Qt any more than the OS/X port does, and they will not resemble each other much more than before.
Normal person: oh shit, that accident means the 135 is blocked and I'll have to take the detour, and I'll be late, and those guys at work will eat all the donuts before I get to the meeting.
False person: oh, but don't you feel sorry for the family of the person who was killed in that accident?
Substitute "Microsoft Word" for "DOS/Windows". DOS/Windows came for "free" with the computer for most buyers, so they did not pirate it, and Microsoft made their money. What they did do is make sure Word/Excel/etc were trivially easy to pirate, for the reasons you state. They also forced the manufacturers to not include "free word processor!" on the machines to avoid this cutting into the pirated word.
I think the paper has to come out of the machine and the voter then takes it and puts it into a ballot box, or can be returned to the disposal box (with the voter seeing it get torn up) to get the right to try to vote again.
The enclosed paper leads to the possibility that the paper seen is not the one put into the box. Granted that seems unlikely, but having the voter move the paper removes this possibility, and also eliminates the need for a complex mechanism to redirect the paper to two different bins.
That's no good, as the vote-buyer could just insist that the voter tell them the voter id number. This can be done before the list is printed. The voter may be able to pick a random number, but they are in trouble if that random number either does not appear or does not indicate a vote for the "correct" candidate.
Actually text config files are fine, for any user capable of running a text editor, provided:
1. The program works if the file is missing, or if the file contains syntax errors.
2. Comments are supported and lines can be commented out by putting a # or something at the start of the file.
3. A fully-commented example file is provided, listing every option and what the default value is and what the possible values are. Ideally the program itself can create this if asked.
4. It is easy to load a new configuration. If necessary the program should have command to kill and restart itself.
A gui is nicer but the above will get you 50% of the usablility for about 1% of the work.
Scripting languages are a pain in these files. You can get all the power of it by having a configuration that says "run this script file". Obviously the sourced script can do anything, but all the constant settings can be in a flat file that both the GUI and text editor can update.
Re:We don't need any steenkin' new paradigms...
on
GUIs Get a Makeover
·
· Score: 1
There are TWO "clipboards". One called CLIPBOARD and one called SELECTION.
SELECTION is really drag & drop, with the advantage that you can rearrange and open/close windows between when you start to drag and when you drop. To "drag" you select the text. To "drop" you click the middle mouse button at the drop location.
Unfortunatly originally there was no clipboard in X, and everybody started calling the selection "cut and paste". This led to programs that copied Mac software using the clipboard shortcuts to set or insert the contents of the selection. Eventually the clipboard was added and programs started changin, but you will occasionally run into such an old program. Not sure what ones are left, what programs have you had trouble with?
Only company Z has to get the source code. However they have to get it with the GPL license, which means they can give it to anybody else they want to.
One way around this is to sell customized services that are not much use except for company Z. I would think you can get company Z to sign a contract saying they won't enter into your customized-services business so the code is useless to them except for maintinence. Or the customized parts consist of trade secrets of company Z, and thus they won't want to re-release the code.
For commercial software the only solution is to bundle other things of value. Embedding a lot of trademarked images which have to be stripped, or including software that does not have a redistribution license, works quite well (this is how the commercial Linux distributions work).
I agree somewhat, although for me "duplex" does not mean anything, I would look for a term containing "both sides" or "2 sided" or something like that. Maybe I don't print enough.
In any case, you are right, such descriptive labels make it very hard to find something in the case where the user knows something is there and wants to locate it. Also there is the need to quickly communicate, by voice or text, the identity of a control. In these cases very short, and cryptic, but unique, terms are actually best!
I would think the solution is combination labels:
[] duplex (print on both sides of the paper)
[] ICS (allow other computers to use this one's internet connection)
[] grep (search for a piece of text in all your files)
This would allow easy identification of which item is which by the short cryptic name, and just as much help. In fact it could allow a lot more help because other items can be referred to by the short name, producing readable sentences that indicate relations between them.
That company could take a GPL program and use it and then claim "no, it's something we wrote ourselves" just as easily, or even easier.
If you assumme the company is going to lie and get away with it, then it does not matter if it is LGPL or GPL or if it is commercial software that they are supposed to buy a license for.
Actually the machine should give the voter the ballot, cutting it off the paper roll, which the voter then place in a seperate ballot box. Technically the reason for this is to prevent the (admittedly unlikely) possible fraud of having the machine destroy the shown ballot and put another ballot (matching the fixed vote) into the collection bin. A more obvious reason is that if the ballot is wrong, the user can tear it up and present it to the collector and ask for a new one, and be quite confident that the incorrect ballot is not collected.
I remember hearing the barbeques contributing a suprising percentage of some pollutant, along with dry-cleaning establishments. However this was only in densely populated suburban areas (ie there isn't much else other than cars polluting there), and it was due to the lighter fluid, not the charcoal/gas/wood, and it was a specific pollutant such as NO or something.
I heard the terminal had arrows printed on those keys, which is why they were chosen. It might also make sense to avoid using a punctuation mark (the semicolon) as one of the arrows.
Although I kind of doubt any reciept should be allowed, a possible solution to this is that the machines refuses to make identical real and fake receipts.
In theory on a very small number of randomly-selected precients need to be hand-counted. If any descrepency is discovered then that triggers the big recount of everybody, but that should not happen unless somebody actually hacked the voting machines.
The idea is that it would be impossible to fix the election because the odds that the precients you chose to fix don't include one of these randomly-selected ones is *extremely* small (or if it is not small, then you have not fixed enough votes to make a difference in the election).
Unfortunatly, as you state, the general public is quite ignorant of probability and will probably demand a recount always, even if you can show that mathematically the odds that the electronic result is wrong are astronomically small.
Sure it's proof. The fact that you "have" the reciept, even with no identifying information, is pretty good proof you voted. Maybe you stole it from somebody else, but that seems an unlikely way to avoid voter coercion. And if it is easy to make a fake one, then the reciept is useless to prove anything, so you might as well leave it in the voting booth.
Yes, I meant that the "open" program should also accept a switch to say what program to actually use, and another switch to have it print to stdout the list of programs so that the caller can make a list of them.
It would help if these things had discoverable names, and were actually used by these systems.
I have a somewhat dated gnome, and there is no "gnome-open". And I certainly would not expect "zenity", which also does not exist on my machine.
My expected names would be "open" and "ask" and/or "message" or "dialog". And it would be really nice if some programs actually called these things so that they are debugged. For instance I don't think it should be allowed to have the file chooser do anything other than call the "open" program when you double click, so that they are forced to put the necessary functionality into "open" and debug it.
Re:But does it have a useable file-save dialogue?
on
GNOME 2.16 Released
·
· Score: 1
As another responder has mentioned, what I want is to not change the *order* of the list based on information returned by stat(). There is no problem with adding an icon or other indicator using a parallel thread, but there are problems with having the list rearrange itself after being drawn the first time.
I agree with you that Gnome is probably *thousands* of times slower than necessary and that if they solved that (which looks hopeless) it would hide the problem in most cases. But doing stat() of every file on a remote file system can still take a significant amount of time.
If putting directories first is an absolute requirement, then the best answer is to just display "reading..." or something like that, until the stat() is done and the display can be shown then. It should accept user input (ie typing) during this time, and if they go to another directory or close the file chooser, it should abort the current stat() pass and start over.
Re:But does it have a useable file-save dialogue?
on
GNOME 2.16 Released
·
· Score: 1
You both seem to have missed the fact that "is it a directory" changes the sort order. Thus determining whether a file is a directory would cause it to move in the list if that determination is done by a parallel thread in the background. Believe me, I tried this, and it is not user-friendly at all, in particular you can try to click on some name and get a different one because it rearranged at that moment. Fortunatly in my software I can ignore the whining and just sort by name and not put the directories first.
As you suggest, changing the icon using a parallel thread is easy. The file displays the icon it would have if it was not a directory, and that changes when the determination is made. People do not seem to mind this.
Tango might want to work a bit on there web site security! (it's been hacked, apparently by somebodying editing all the lead wiki pages to say "HA HA").
Slashdot Mirror
No, Cairo is more like a new graphics api for drawing shapes and text and lines, etc. It is similar to Quartz, or Display PostScript. Both GTK and Qt on OS/X use Quartz to draw everything, they will probably both use this api to draw anything on X once it is commonly installed. However this will not change the API to GTK or Qt any more than the OS/X port does, and they will not resemble each other much more than before.
This should help the other window managers considerably. This shared API can also be used by them.
Don't try to be so holier-than-though!
The reality is:
Normal person: oh shit, that accident means the 135 is blocked and I'll have to take the detour, and I'll be late, and those guys at work will eat all the donuts before I get to the meeting.
False person: oh, but don't you feel sorry for the family of the person who was killed in that accident?
Normal person: um, no.
That does not necessarily mean "secret from Hans", or secret from his lawyer. Perhaps it was in both parties interest not to release the information.
Substitute "Microsoft Word" for "DOS/Windows". DOS/Windows came for "free" with the computer for most buyers, so they did not pirate it, and Microsoft made their money. What they did do is make sure Word/Excel/etc were trivially easy to pirate, for the reasons you state. They also forced the manufacturers to not include "free word processor!" on the machines to avoid this cutting into the pirated word.
I think the paper has to come out of the machine and the voter then takes it and puts it into a ballot box, or can be returned to the disposal box (with the voter seeing it get torn up) to get the right to try to vote again.
The enclosed paper leads to the possibility that the paper seen is not the one put into the box. Granted that seems unlikely, but having the voter move the paper removes this possibility, and also eliminates the need for a complex mechanism to redirect the paper to two different bins.
That's no good, as the vote-buyer could just insist that the voter tell them the voter id number. This can be done before the list is printed. The voter may be able to pick a random number, but they are in trouble if that random number either does not appear or does not indicate a vote for the "correct" candidate.
Actually text config files are fine, for any user capable of running a text editor, provided:
1. The program works if the file is missing, or if the file contains syntax errors.
2. Comments are supported and lines can be commented out by putting a # or something at the start of the file.
3. A fully-commented example file is provided, listing every option and what the default value is and what the possible values are. Ideally the program itself can create this if asked.
4. It is easy to load a new configuration. If necessary the program should have command to kill and restart itself.
A gui is nicer but the above will get you 50% of the usablility for about 1% of the work.
Scripting languages are a pain in these files. You can get all the power of it by having a configuration that says "run this script file". Obviously the sourced script can do anything, but all the constant settings can be in a flat file that both the GUI and text editor can update.
There are TWO "clipboards". One called CLIPBOARD and one called SELECTION.
SELECTION is really drag & drop, with the advantage that you can rearrange and open/close windows between when you start to drag and when you drop. To "drag" you select the text. To "drop" you click the middle mouse button at the drop location.
Unfortunatly originally there was no clipboard in X, and everybody started calling the selection "cut and paste". This led to programs that copied Mac software using the clipboard shortcuts to set or insert the contents of the selection. Eventually the clipboard was added and programs started changin, but you will occasionally run into such an old program. Not sure what ones are left, what programs have you had trouble with?
Congratulations, you have just reinvented the X multi-screen support from 1985. You cannot move a window from one screen to another.
Despite the fact that it sucked in 1985, apparently because Microsoft allows it to happen in 2006 it is a great thing, huh.
Also I am a little confused by the earlier poster. I though you *could* replicate the taskbar on each screen. It works in W2k, did they remove that?
Only company Z has to get the source code. However they have to get it with the GPL license, which means they can give it to anybody else they want to.
One way around this is to sell customized services that are not much use except for company Z. I would think you can get company Z to sign a contract saying they won't enter into your customized-services business so the code is useless to them except for maintinence. Or the customized parts consist of trade secrets of company Z, and thus they won't want to re-release the code.
For commercial software the only solution is to bundle other things of value. Embedding a lot of trademarked images which have to be stripped, or including software that does not have a redistribution license, works quite well (this is how the commercial Linux distributions work).
You missed these:
4) Replace the GPL portions with code you wrote or bought
5) Replace the NDA ones with code you wrote or bought
I agree somewhat, although for me "duplex" does not mean anything, I would look for a term containing "both sides" or "2 sided" or something like that. Maybe I don't print enough.
In any case, you are right, such descriptive labels make it very hard to find something in the case where the user knows something is there and wants to locate it. Also there is the need to quickly communicate, by voice or text, the identity of a control. In these cases very short, and cryptic, but unique, terms are actually best!
I would think the solution is combination labels:
[] duplex (print on both sides of the paper)
[] ICS (allow other computers to use this one's internet connection)
[] grep (search for a piece of text in all your files)
This would allow easy identification of which item is which by the short cryptic name, and just as much help. In fact it could allow a lot more help because other items can be referred to by the short name, producing readable sentences that indicate relations between them.
That company could take a GPL program and use it and then claim "no, it's something we wrote ourselves" just as easily, or even easier.
If you assumme the company is going to lie and get away with it, then it does not matter if it is LGPL or GPL or if it is commercial software that they are supposed to buy a license for.
Actually the machine should give the voter the ballot, cutting it off the paper roll, which the voter then place in a seperate ballot box. Technically the reason for this is to prevent the (admittedly unlikely) possible fraud of having the machine destroy the shown ballot and put another ballot (matching the fixed vote) into the collection bin. A more obvious reason is that if the ballot is wrong, the user can tear it up and present it to the collector and ask for a new one, and be quite confident that the incorrect ballot is not collected.
I remember hearing the barbeques contributing a suprising percentage of some pollutant, along with dry-cleaning establishments. However this was only in densely populated suburban areas (ie there isn't much else other than cars polluting there), and it was due to the lighter fluid, not the charcoal/gas/wood, and it was a specific pollutant such as NO or something.
I heard the terminal had arrows printed on those keys, which is why they were chosen. It might also make sense to avoid using a punctuation mark (the semicolon) as one of the arrows.
Although I kind of doubt any reciept should be allowed, a possible solution to this is that the machines refuses to make identical real and fake receipts.
In theory on a very small number of randomly-selected precients need to be hand-counted. If any descrepency is discovered then that triggers the big recount of everybody, but that should not happen unless somebody actually hacked the voting machines.
The idea is that it would be impossible to fix the election because the odds that the precients you chose to fix don't include one of these randomly-selected ones is *extremely* small (or if it is not small, then you have not fixed enough votes to make a difference in the election).
Unfortunatly, as you state, the general public is quite ignorant of probability and will probably demand a recount always, even if you can show that mathematically the odds that the electronic result is wrong are astronomically small.
Sure it's proof. The fact that you "have" the reciept, even with no identifying information, is pretty good proof you voted. Maybe you stole it from somebody else, but that seems an unlikely way to avoid voter coercion. And if it is easy to make a fake one, then the reciept is useless to prove anything, so you might as well leave it in the voting booth.
Yes, I meant that the "open" program should also accept a switch to say what program to actually use, and another switch to have it print to stdout the list of programs so that the caller can make a list of them.
It would help if these things had discoverable names, and were actually used by these systems.
I have a somewhat dated gnome, and there is no "gnome-open". And I certainly would not expect "zenity", which also does not exist on my machine.
My expected names would be "open" and "ask" and/or "message" or "dialog". And it would be really nice if some programs actually called these things so that they are debugged. For instance I don't think it should be allowed to have the file chooser do anything other than call the "open" program when you double click, so that they are forced to put the necessary functionality into "open" and debug it.
As another responder has mentioned, what I want is to not change the *order* of the list based on information returned by stat(). There is no problem with adding an icon or other indicator using a parallel thread, but there are problems with having the list rearrange itself after being drawn the first time.
I agree with you that Gnome is probably *thousands* of times slower than necessary and that if they solved that (which looks hopeless) it would hide the problem in most cases. But doing stat() of every file on a remote file system can still take a significant amount of time.
If putting directories first is an absolute requirement, then the best answer is to just display "reading..." or something like that, until the stat() is done and the display can be shown then. It should accept user input (ie typing) during this time, and if they go to another directory or close the file chooser, it should abort the current stat() pass and start over.
You both seem to have missed the fact that "is it a directory" changes the sort order. Thus determining whether a file is a directory would cause it to move in the list if that determination is done by a parallel thread in the background. Believe me, I tried this, and it is not user-friendly at all, in particular you can try to click on some name and get a different one because it rearranged at that moment. Fortunatly in my software I can ignore the whining and just sort by name and not put the directories first.
As you suggest, changing the icon using a parallel thread is easy. The file displays the icon it would have if it was not a directory, and that changes when the determination is made. People do not seem to mind this.
Tango might want to work a bit on there web site security! (it's been hacked, apparently by somebodying editing all the lead wiki pages to say "HA HA").