I have nothing against standardization process itself. But I believe to be eligible the draft to standardize should either be very good or not go through "fast track".
OK, I do have a little. There really should not be too many standards. OOXML is unnecessary.
But I am very convinced that Microsoft will not implement their software according to the standard, especially if the standard is fixed. Why should they? They can just claim to implement it "better than anybody else". And everybody else must implement 1900 as a leap year too or they are not not *Microsoft* compatible (compatibility with OOXML standard will not matter a bit).
I do believe that "the rest of the world" can keep OOXML from getting standardisation. But that is beside the point. So is NIH - you could say ODF is as much "NIH" as OOXML.
My point is: I do not believe Gnome Foundation should lose time and effort trying to implement all the "features" and bugs Microsoft will have in all the undocumented (and other) places. Especially I do not think Gnome Foundation should waste time before OOXML is standardized and widely used (if ever).
Do you honestly believe that you can faithfully implement the XML in the way Microsoft will use it? You see, they will not implement it according to the standard. They have said it themselves: "are not committed". So I think neither of us believe that Microsoft is going to change the format they use no matter how OOXML is changed before approval (even if it isn't). There will never be "OOXML" to use, just a bunch of variants (depending on Office versions).
BTW, how are you (Gnome) going to implement the bugs in Excel? Is 1900 a leap year in my calculations or not? Can I use dates before 1900? How "ceiling" function is going to behave, mathematically correctly or according to the bug in Excel?
So the most important question becomes: can I trust Gnome software to work mathematically correctly or is it going to "brownnose"?
P.S. I would understand external converter, that would help everybody (esp. developers of other spreadsheets). This one just makes their job harder.
Why should there be two standards for essentially one thing?
If more the merrier, why not 200? This would of course mean not a single one gets adopted. If less is better what does OOXML bring that is useful to the users and other developers (non Gnumeric)? So useful that developers have to implement support for both, so good a support that user's are not inconvenienced?
You are in the position to say how big task it is to support both instead of one and if it is really worthwhile effort, not me. I am just asking.
There is small but important difference. Gnome is LGPL, KDE is GPL.
This means I cannot make proprietary software to KDE (unless I pay TrollTech, but this is not expensive). It also means that if I write a free utility to KDE and if TrollTech "does SCO" - then I cannot sell proprietary spin-off of the free utility (without rewrite).
As you can see the difference is small but non-negligible.
You are wrong. There were a lot of people who did not upgrade from NT (3.5). One of the reasons was BSOD. It was prevalent before SP2 (or SP4 - hard to say).
One reason for instability was the move of the GUI into kernel.
I want my car to drive me home when I'm too drunk to drive myself! I too want your car to drive you home when you are too drunk to drive. (I never go drinking with a car anymore, it used to be annoying to fetch it the next day...).
More seriously, I am waiting for the time when cars can drive themselves. It would solve so many problems. I might be able sell my car (and join a car pool) getting perhaps $300 savings a month. Unfortunately it seems I'll be dead before that happens.
All of those programs have a natural fall-back mechanism, that is their own source. Now I am certain that either you are joking or have never ever administered a single server.
The last thing to do when shit hits the fan is to start reading source code. I'd rather go to shop and buy AIX, Windows and Mac OS combined.
No offense but there are abundant amount of F/OSS which has half assed man page, if even that. Actually every single GNU tool has half assed man page (FSF insist on the brain dead "info").
Lets take Fedora Core 6 as an example. "ps -eo comm | sort -u | wc" gives 130 processes running. How many of them you'd expect to have a manual page? How many of them really does?
Only slightly over one third has! (49 has, 81 does not).
Please do not claim *nixes are in any shape better than Windows in this sense (the need of arcane knowledge).
[...] your.net knowlege, in the future, will be just as useful as FoxPro knowlege is now. It has already once happened:.net2 is not compatible with.net1. This is not so much problem to the programmer than to the user: he must install runtime for both.
I cannot even imagine the reason why.net2 cannot emulate nor include runtime for.net1 programs, but this is immensely idiotic.
Thank god Sun has not done anything that stupid with Java.
Why cannot you do the second in GPL2? Or the third?
I am not sure if the second point is legal with GPL3. It does state that your modifications must be runnable. If it cannot do anything it is questionable if you can convince a jury that it adheres GPL3.
Besides the third point is much less a problem than with GPL2 - the DRM cannot be for the SW itself. You cannot lock out my modifications. With GPL2 you can.
There are lot of issues, many of which boils down to the problem: neither Windows (GUI) nor Windows applications know anything about virtual desktops (no API) and can do nothing to help the desktop manager for best user experience. Microsoft can change the GUI, but they cannot change all the programs.
I would require e.g. that pop-ups happen on the desktop the application is, not somewhere else. And the pop-ups should be pop-up, not pop-unders (problem in virtuawin). BTW, the problem of Firefox changing desktop is in VirtuaWin too, apparently for the reason you gave.
Then there is the problem of session management. Unfortunately for unknown reason Firefox team refuses to implement proper session management hooks, but in principle (and with other programs) when exiting a session (log-off) you will log in as the way you left (if you so prefer). No API -> it cannot work flawlessly.
BTW, I have never tried Dexpot, but VirtuaWin is good enough (for me). I couldn't live without it. Yes, it is just a shadow of Gnome/KDE/Xfce/tvwm/...
I agree completely, VirtuaWin beats the Microsoft offering in every respect.
But even VirtuaWin falls short, far far away short, of *nixes virtual desktops. The fault is definitely not in the VirtuaWin however. If the applications and windowing api are not designed with the virtual desktops in mind there is only so much you can do.
How much would you pay for the iCal (yearly)? $10?
Now assume 10 million users willing to pay same amount. You are still several orders of magnitude short of 15 billion (or milliard as it is called here... giga:-)
There is Java in every Nokia phone (except for N770/N800, but they are not phones...).
I really would like to have midpssh and E61. Unfortunately I do not have money for that E61:-( VNC would be huge addition (I do not know of any port to JavaME).
Anyway the number of Java applications limits iPhone out of my choices - unless Apple decides to include it in Europe.
P.S. MidpSSH is PITA in at least 6630 and probably in anything without keyboard and bigger display.
But will Microsoft use the situation to their advantage now? They have already bought and blackmailed for sabotaging in some countries, so they definitely have no policy against it. ISO just might be too visible and might get a huge retribution from EU or like.
Slashdot Mirror
I have nothing against standardization process itself. But I believe to be eligible the draft to standardize should either be very good or not go through "fast track".
OK, I do have a little. There really should not be too many standards. OOXML is unnecessary.
But I am very convinced that Microsoft will not implement their software according to the standard, especially if the standard is fixed. Why should they? They can just claim to implement it "better than anybody else". And everybody else must implement 1900 as a leap year too or they are not not *Microsoft* compatible (compatibility with OOXML standard will not matter a bit).
I do believe that "the rest of the world" can keep OOXML from getting standardisation. But that is beside the point. So is NIH - you could say ODF is as much "NIH" as OOXML.
My point is: I do not believe Gnome Foundation should lose time and effort trying to implement all the "features" and bugs Microsoft will have in all the undocumented (and other) places. Especially I do not think Gnome Foundation should waste time before OOXML is standardized and widely used (if ever).
Do you honestly believe that you can faithfully implement the XML in the way Microsoft will use it? You see, they will not implement it according to the standard. They have said it themselves: "are not committed". So I think neither of us believe that Microsoft is going to change the format they use no matter how OOXML is changed before approval (even if it isn't). There will never be "OOXML" to use, just a bunch of variants (depending on Office versions).
BTW, how are you (Gnome) going to implement the bugs in Excel? Is 1900 a leap year in my calculations or not? Can I use dates before 1900? How "ceiling" function is going to behave, mathematically correctly or according to the bug in Excel?
So the most important question becomes: can I trust Gnome software to work mathematically correctly or is it going to "brownnose"?
P.S. I would understand external converter, that would help everybody (esp. developers of other spreadsheets). This one just makes their job harder.
No, I'm younger (40 something). From parking to *legal* driving I think will take over 40 years.
I'd love to be proved wrong.
Why?
Why should there be two standards for essentially one thing?
If more the merrier, why not 200? This would of course mean not a single one gets adopted.
If less is better what does OOXML bring that is useful to the users and other developers (non Gnumeric)? So useful that developers have to implement support for both, so good a support that user's are not inconvenienced?
You are in the position to say how big task it is to support both instead of one and if it is really worthwhile effort, not me. I am just asking.
There is small but important difference. Gnome is LGPL, KDE is GPL.
This means I cannot make proprietary software to KDE (unless I pay TrollTech, but this is not expensive). It also means that if I write a free utility to KDE and if TrollTech "does SCO" - then I cannot sell proprietary spin-off of the free utility (without rewrite).
As you can see the difference is small but non-negligible.
You are wrong. There were a lot of people who did not upgrade from NT (3.5). One of the reasons was BSOD. It was prevalent before SP2 (or SP4 - hard to say).
One reason for instability was the move of the GUI into kernel.
It is very questionable to consider "continued functioning" if software cannot do anything (useful).
IMNHSO it does not "run as it would otherwise" if the HW does not respond "as it would otherwise".
I'd love to see you trying to convince your analogy to a jury, it would be hilarious. No offense.
More seriously, I am waiting for the time when cars can drive themselves. It would solve so many problems. I might be able sell my car (and join a car pool) getting perhaps $300 savings a month. Unfortunately it seems I'll be dead before that happens.
The last thing to do when shit hits the fan is to start reading source code. I'd rather go to shop and buy AIX, Windows and Mac OS combined.
No offense but there are abundant amount of F/OSS which has half assed man page, if even that. Actually every single GNU tool has half assed man page (FSF insist on the brain dead "info").
Lets take Fedora Core 6 as an example. "ps -eo comm | sort -u | wc" gives 130 processes running. How many of them you'd expect to have a manual page? How many of them really does?
Only slightly over one third has! (49 has, 81 does not).
Please do not claim *nixes are in any shape better than Windows in this sense (the need of arcane knowledge).
I cannot even imagine the reason why
Thank god Sun has not done anything that stupid with Java.
Why cannot you do the second in GPL2? Or the third?
I am not sure if the second point is legal with GPL3. It does state that your modifications must be runnable. If it cannot do anything it is questionable if you can convince a jury that it adheres GPL3.
Besides the third point is much less a problem than with GPL2 - the DRM cannot be for the SW itself. You cannot lock out my modifications. With GPL2 you can.
Strange, for me VS (2003) does not crash. Sometimes (once a week) it confuses the VirtuaWin, but neither crashes. Restarting VS solves this.
I have no clue why mine works and yours don't, sorry.
There are lot of issues, many of which boils down to the problem: neither Windows (GUI) nor Windows applications know anything about virtual desktops (no API) and can do nothing to help the desktop manager for best user experience. Microsoft can change the GUI, but they cannot change all the programs.
I would require e.g. that pop-ups happen on the desktop the application is, not somewhere else. And the pop-ups should be pop-up, not pop-unders (problem in virtuawin). BTW, the problem of Firefox changing desktop is in VirtuaWin too, apparently for the reason you gave.
Then there is the problem of session management. Unfortunately for unknown reason Firefox team refuses to implement proper session management hooks, but in principle (and with other programs) when exiting a session (log-off) you will log in as the way you left (if you so prefer). No API -> it cannot work flawlessly.
BTW, I have never tried Dexpot, but VirtuaWin is good enough (for me). I couldn't live without it. Yes, it is just a shadow of Gnome/KDE/Xfce/tvwm/...
Well, Mr. Lyons' errors were pointed in detail several times - he refused to believe. Until now.
Besides "everybody else did too" is really, REALLY not a good excuse.
I agree completely, VirtuaWin beats the Microsoft offering in every respect.
But even VirtuaWin falls short, far far away short, of *nixes virtual desktops. The fault is definitely not in the VirtuaWin however. If the applications and windowing api are not designed with the virtual desktops in mind there is only so much you can do.
Exxon-Mobile? Is that some kind of combustion engine powered mobile phone? (sorry, i've got fewer ...)
See http://slashdot.org/comments.pl?sid=316537&cid=20847323
How much would you pay for the iCal (yearly)? $10?
... giga :-)
Now assume 10 million users willing to pay same amount. You are still several orders of magnitude short of 15 billion (or milliard as it is called here
There is Java in every Nokia phone (except for N770/N800, but they are not phones ...).
:-( VNC would be huge addition (I do not know of any port to JavaME).
I really would like to have midpssh and E61. Unfortunately I do not have money for that E61
Anyway the number of Java applications limits iPhone out of my choices - unless Apple decides to include it in Europe.
P.S. MidpSSH is PITA in at least 6630 and probably in anything without keyboard and bigger display.
Neither I do believe this was planned outcome.
But will Microsoft use the situation to their advantage now? They have already bought and blackmailed for sabotaging in some countries, so they definitely have no policy against it. ISO just might be too visible and might get a huge retribution from EU or like.
With the recent increase of drunk drivers and speeding I am waiting for this to happen.
Could you please tell me how to "make" new (0.9.X) Evince on FC 6?
As it happens this is practically impossible to do (it is hugely easier to migrate to F7 or whatever).
And there goes the "Linux" ...
After what you explained we'd have two, different and incompatible Linuxes, RH and Ubuntu.
Unlike for you, the software I use does not run in both RH-3 and (not so old) Ubuntu, the libraries are too different and so are the kernels.
Linux really needs to get "binary compatibility", and so does Linux distros (LSB among others).