One other thing. Toolkits like Qt and wxWidgets are complex beasts with lots of corner cases that can screw things up. It takes a lot of testing to get it right. That's another reason I choose Qt over wxWidgets. Qt gets hammered on by the KDE project all the time. Millions of lines of code and just about every type of app you can imagine. What is wxWidgets used in? I can only think of Audacity and VLC, neither of which has a particularly good UI.
Like you, I have also worked with all three of those toolkits (Gtk+, wxWidgets, and Qt), and had quite the opposite experience. Granted I was working from C++, so maybe there are some PyQt bugs that I'm not seeing.
I agree that Qt apps don't look quite 100% native, but it's so close that I bet most users wont notice. Think about how many toolkits are in use on Windows already. Every second app rolls their own, so users are accustomed to little differences.
GTK I can't use, because it looks and behaves horribly on Windows, which is one of our primary targets, and thus is right out.
wxWidgets would have been a contender, but last time I tried it, I found a bug (in the Timer) within a few hours of learning it. Now it may have just been bad luck, but finding a bug that quickly indicates to me that actual serious use would uncover a lot more. Also I can't stand the horrible MFC style message maps for events. Maybe if they get rid of that I'll give it another look.
"Better UI design tools"? I've heard nothing but bad things about them, even from the most hardcore Qt advocates.
Personally, I like Designer. It does the job for me, and I like the way of generating code from the UI file. I always thought loading an XML file for the UI at runtime was a waste of time. In 99% of the cases, you won't be changing the UI at runtime, so why load it every time? It adds overhead both in terms of load time, and in terms of you having to have a full featured XML parser included with your program. It makes much more sense to compile the UI to code. Qt 4.2 allows you to theme your UI with stylesheets at runtime if you really want to.
I don't understand your complaint about the event system either. For custom widgets you always use the inheritance method. Whether that is to handle a click, or to handle a drag event. There is no inconsistency there. If you want to get an event from another object, you use the signal/slot method for obvious reasons. There is no overlap between the two.
What _possible_ ROI benefit is there to optimising an OS for people who won't even prepared to spend US$50ish upgrading their 5 year old PCs with a better video card an a gig of RAM ?
I can think of several: 1. Laptops. They are selling more than desktops these days, and you can't upgrade the graphics card on them (usually). Also, even modern laptops rarely have high performance video cards. 2. Power consumption. GPU processing isn't free like some people want to pretend. More processing = less time that the GPU can be in power saving mode.
Also, since you didn't read my post properly, my laptop is just over 1 year old, and yet Vista feels sluggish on it. Yes the video card isn't the best (ATI Radeon Xpress 200M, 128MB dedicated RAM) but considering it runs reasonably modern games like HL2, I don't see why it is sluggish on the effects in Vista.
I have to second this. I don't use mac's (the UI annoys me), but they really got the performance of their GUI right. At work I once had an old 500Mhz G3, (running 10.3 or something) and the UI performance was great. Resizing windows was a bit choppy, but everything else worked very smoothly. Now if I run Vista on my laptop from last year (AMD64 3200, 1GB RAM, ATI Xpress200m), it is not as smooth as that ancient iBook. I can guarantee you every component in my laptop is far far faster than anything in the iBook.
So the question remains, why is the Aero UI such a pig when it has been demonstrated by both Apple and Linux distros that the same sort of effects can be done with garbage hardware.
Well given that it has already been (partially) done, by one or two people working in their spare time, I doubt it's as difficult as you think. See here (it seems to be now defunct, but they did have a working release out at one point): http://en.wikipedia.org/wiki/Swift_(web_browser)/
Are you so blinded by partisan stupidity that you believe there aren't just as many rabid bush haters doing the exact same thing?
Wow. If people doing the same thing on the other side, there is no problem? The problem is that people are able to do this in the first place! Which party is doing it is completely immaterial. (especially to me, I don't even live in the US)
"The one area of Linux ownership and use where it becomes apparent that there's an assumption that everyone who uses Linux is an expert is hardware support. Your average user doesn't have the time, the energy or the inclination to deal with uncertainty. Also, they usually only have the one PC to play with. Hardware just has to work. There's a very good reason why Microsoft spends a lot of time on hardware compatibility - it's what people want."
Ooooh! So that's what people want! Well that's easy then, let's just fix that issue right now. Silly linux developers, making hardware support poor on purpose because they didn't know it was important to people.
Seriously though, Windows hardware support is horrible out of the box. The only reason hardware works on it is because the hardware manufacturers spend a lot of time making drivers for it. Of course, for an end user, the reality is that Linux doesn't support their gadget and they don't care why, but blaming linux developers for that is completely backwards. Until hardware manufacturers start writing linux drivers, there will always be some hardware that doesn't work as well on Linux as Windows. No sense blaming the Linux devs for what is essentially a market share problem (chicken/egg).
By the way, try upgrading a random Win98/WinMe era laptop to W2K to see how nice the hardware support is on Windows. At least one or two devices won't work because the drivers are proprietary and were never released for W2K/XP. And that's it, no chance of ever getting it working.
Does it achieve a goal that couldn't have been achieved within the GIMP codebase with less effort?
Perhaps, perhaps not. In any case, Krita is surpassing the Gimp in some areas that people have been complaining about for years in Gimp, and nothing was done. As a developer, what would you rather do, argue on the Gimp mailing list until your face turns blue about wanting to change the interface, or just start your own project? Sometimes you have to make a clean break to get new ideas implemented.
Whether FF renders faster than IE is true or not, I have no idea, but it is a good question.
A good question perhaps, but not a good statement. The original poster was claiming that IE is faster than Firefox at rendering without any proof. Since that is not what benchmarks have shown (search on google, it's pretty much a toss up between the two), and doesn't reflect many people's experiences, it is only natural that his conclusions are being questioned. Of course, it may be that IE is faster than Firefox on his computer, but it could be due to any number of external factors (buggy extensions come to mind).
Well I checked it out, and it absolutely sucks. You're trying to get an interactive framework where every request has to do a roundtrip to the server. That's completely backwards, and it shows. Try the demo and see how laggy everything is. Sure it's nice you don't have to rely on the browser's buggy javascript implementation, but the lag associated with this approach is unacceptable. It's funny how they state that not having to rely on user's possibly slow machines is an advantage of their approach. Never mind that any reasonable machine (ie, built in the last 10 years) will be far faster than the network latency involved in doing it their way.
I don't know about Charter, but I do know that MSP (Medical Services Plan in canada) does this. For 15 minutes I tried to wrestle with the voice activated menus they have, and absolutely could not figure out how to speak to a representative. Well eventually I just got pissed and started swearing at the thing. It paused for a while, then said "Ok, a representative, one moment please"
Best. system. ever.:)
Re:Yes: I, a KDE fan, can't use KWord: no Word imp
on
KOffice 1.6 Released
·
· Score: 1
As an aside, I use AbiWord. It's free, and it has a passable MS Word import/export facility.
And out of curiosity, how many of those jobs where you created your resume in Abiword have you actually gotten?
"AbiWord can currently save in an MS Word compatible ".doc" format. This is done by saving as Rich Text Format (.rtf) but with a.doc extension. (...) There are no plans to support binary MS Word export."
Development has certainly been active lately. Cyrille Berger has a blog where he talks about some of the new features that have gone into Krita lately. http://cyrilleberger.blogspot.com/
Debian with KDE 3.5.4 on my old box, 32MB of ram used. That's almost umbelievable - I mean, that's only possible with a stripped down KDE installation, with not a single service running.
Yeah, I was amazed myself. I think it also had something to do with a recent XOrg upgrade, as before that, memory usage was around 45MB on fresh boot. The KDE itself is not overly stripped. Although this is running on my second box, which is a PII 266 w/ 196 mb of ram, so I disabled the background image, disabled all UI effects, and turned off antialiasing. Also, things like klipper (which I have no use for anyway) are disabled. Other than that, it's fairly standard KDE. Keep in mind that this is running on Debian, which by default has nothing running (I think before starting X, it uses about 10MB of RAM). A similar setup on a more desktop oriented distribution like Kubuntu will be using around 100MB, mostly due to little applets running, and more system services.
Speed. KDE (and Gnome) need better speed optimizations.
This is most likely your video card drivers. KDE is plenty fast, but if you dont have acceleration working in X, then everything will seem sluggish. My card is poorly supported (ATI Xpress 200m) and it makes everything seem slow.
Memory usage. The memory requirements of KDE and and Gnome are ridiculous.
Yes, if you mean ridiculously low. Fresh boot, Debian with KDE 3.5.4 on my old box, 32MB of ram used. Start up konversation (irc client) and it's about 45MB. Every subsequent application uses less extra ram, because the libraries are already loaded. Fresh boot on windows xp is at least 100MB, on my laptop more like 150. Most likely you have no idea on how to measure memory usage on linux. Have a look here: http://ktown.kde.org/~seli/memory/
Clutter.
You've got a point there, it's getting better with every release though. And no, sacrificing features for simplicity like Gnome did is not a good strategy.
Consistency.
That's one of the strengths of KDE actually. Everything works the same across the KDE apps. Keyboard shortcuts, look, general menu structure, colours, style, etc etc. And then we get into the even more important consistency, which is functional consistency. Just about every app that needs a text editor uses the same one, so they all behave the same. The same spell checking engine is used almost everywhere, and the password manager saves passwords for every application that has a need to store them. No other operating system is anywhere close to that consistent. Not OS X, not Windows, nothing.
An alternative to this is to just keep all your documents on a flash drive. Small enough to carry around, relatively cheap, and you can even carry your programs around with you. Eventually you will have to start paying a monthly fee for those online apps you use, and then it will really add up.
As a recent engineering grad, I know this is all too commonplace. I did it a lot, my friends did it a lot, almost everyone did it a lot. Sometimes the work load was just too much, and copying an assignment was the only way it could be completed without overworking yourself. So the alternative was to not finish some assignments and get poor marks, or copy them, and then study like mad for the final.
I feel that the important thing is to learn the material. You can't cheat on a final exam, and if you know it there, then you've obviously learned the material, copying or not. Assignments are just devices to make sure that the student keeps up with the course material. If students copy them verbatim and don't understand, then that will be their problem in the final.
Of course, none of us was dumb enough to copy an assignment line for line. There were always changes made, and if a question seemed wrong and the original author couldn't explain themselves, then it was not copied. Someone was always copying some assignent or other. I always liked people copying my work, because in doing so they would very often point out mistakes I had made in the process.
Correct. However, I would rather have the ability to do something, even if it takes a bit more figuring out, rather than not being able to do it at all. Thus, while KDE could be more intuitive, it is usable for me, while Gnome is not at all. Concentrating on usability before having basic features is a mistake.
I believe as Gnome adds the features that everyone has taken for granted in other desktops, they will have trouble retaining their simple, efficient UI.
Of course they welcome this. Just look at the results page for the Coverty scans and see how many defects have been fixed in major open source projects. http://scan.coverity.com/
What happens when the bus stops on a railway track and gets broadsided?
No really, this happened to my mom when she was younger. She just broke her leg, but lots of other people on that bus died. If you're that worried about safety, take the train!:)
The side effect of feeding the hungry is that it effectively destroys their entire local food production business. The farmers who previously supported themselves selling food can't compete with free and are suddenly themselves dependant on handouts to survive.
Do some reading on how the flood of donated clothes from the western world destroyed the textile industry in many areas of Africa. Handouts are a terrible long term solution.
Actually they probably are sensitive enough to pick up the muscle signals going to your eyes. Muscle signals are much easier to detect than brain waves. We have a similar device at work (headband with a couple electrodes) that will do the same thing if you set the filters to the correct frequency for the eye movement signals.
Next up was a device that measured my perscription. I had to stare at a little picture while it zoomed in and out of focus. Apparently this determines my exact perscription, none of that "Is this better, or that" lens swapping. I wonder why eye doctors don't use this all the time.
My optometrist has one and uses it but says it is not nearly as accurate as a manual exam with the lenses. It's just there to give him a rough estimate as a starting point, but it tends to overprescribe.
Slashdot Mirror
One other thing. Toolkits like Qt and wxWidgets are complex beasts with lots of corner cases that can screw things up. It takes a lot of testing to get it right. That's another reason I choose Qt over wxWidgets. Qt gets hammered on by the KDE project all the time. Millions of lines of code and just about every type of app you can imagine. What is wxWidgets used in? I can only think of Audacity and VLC, neither of which has a particularly good UI.
Like you, I have also worked with all three of those toolkits (Gtk+, wxWidgets, and Qt), and had quite the opposite experience. Granted I was working from C++, so maybe there are some PyQt bugs that I'm not seeing.
I agree that Qt apps don't look quite 100% native, but it's so close that I bet most users wont notice. Think about how many toolkits are in use on Windows already. Every second app rolls their own, so users are accustomed to little differences.
GTK I can't use, because it looks and behaves horribly on Windows, which is one of our primary targets, and thus is right out.
wxWidgets would have been a contender, but last time I tried it, I found a bug (in the Timer) within a few hours of learning it. Now it may have just been bad luck, but finding a bug that quickly indicates to me that actual serious use would uncover a lot more. Also I can't stand the horrible MFC style message maps for events. Maybe if they get rid of that I'll give it another look.
"Better UI design tools"? I've heard nothing but bad things about them, even from the most hardcore Qt advocates.
Personally, I like Designer. It does the job for me, and I like the way of generating code from the UI file. I always thought loading an XML file for the UI at runtime was a waste of time. In 99% of the cases, you won't be changing the UI at runtime, so why load it every time? It adds overhead both in terms of load time, and in terms of you having to have a full featured XML parser included with your program. It makes much more sense to compile the UI to code. Qt 4.2 allows you to theme your UI with stylesheets at runtime if you really want to.
I don't understand your complaint about the event system either. For custom widgets you always use the inheritance method. Whether that is to handle a click, or to handle a drag event. There is no inconsistency there. If you want to get an event from another object, you use the signal/slot method for obvious reasons. There is no overlap between the two.
What _possible_ ROI benefit is there to optimising an OS for people who won't even prepared to spend US$50ish upgrading their 5 year old PCs with a better video card an a gig of RAM ?
I can think of several:
1. Laptops. They are selling more than desktops these days, and you can't upgrade the graphics card on them (usually). Also, even modern laptops rarely have high performance video cards.
2. Power consumption. GPU processing isn't free like some people want to pretend. More processing = less time that the GPU can be in power saving mode.
Also, since you didn't read my post properly, my laptop is just over 1 year old, and yet Vista feels sluggish on it. Yes the video card isn't the best (ATI Radeon Xpress 200M, 128MB dedicated RAM) but considering it runs reasonably modern games like HL2, I don't see why it is sluggish on the effects in Vista.
I have to second this. I don't use mac's (the UI annoys me), but they really got the performance of their GUI right. At work I once had an old 500Mhz G3, (running 10.3 or something) and the UI performance was great. Resizing windows was a bit choppy, but everything else worked very smoothly. Now if I run Vista on my laptop from last year (AMD64 3200, 1GB RAM, ATI Xpress200m), it is not as smooth as that ancient iBook. I can guarantee you every component in my laptop is far far faster than anything in the iBook.
So the question remains, why is the Aero UI such a pig when it has been demonstrated by both Apple and Linux distros that the same sort of effects can be done with garbage hardware.
Well given that it has already been (partially) done, by one or two people working in their spare time, I doubt it's as difficult as you think.
See here (it seems to be now defunct, but they did have a working release out at one point):
http://en.wikipedia.org/wiki/Swift_(web_browser)/
Are you so blinded by partisan stupidity that you believe there aren't just as many rabid bush haters doing the exact same thing?
Wow. If people doing the same thing on the other side, there is no problem? The problem is that people are able to do this in the first place! Which party is doing it is completely immaterial. (especially to me, I don't even live in the US)
"The one area of Linux ownership and use where it becomes apparent that there's an assumption that everyone who uses Linux is an expert is hardware support. Your average user doesn't have the time, the energy or the inclination to deal with uncertainty. Also, they usually only have the one PC to play with. Hardware just has to work. There's a very good reason why Microsoft spends a lot of time on hardware compatibility - it's what people want."
Ooooh! So that's what people want! Well that's easy then, let's just fix that issue right now. Silly linux developers, making hardware support poor on purpose because they didn't know it was important to people.
Seriously though, Windows hardware support is horrible out of the box. The only reason hardware works on it is because the hardware manufacturers spend a lot of time making drivers for it. Of course, for an end user, the reality is that Linux doesn't support their gadget and they don't care why, but blaming linux developers for that is completely backwards. Until hardware manufacturers start writing linux drivers, there will always be some hardware that doesn't work as well on Linux as Windows. No sense blaming the Linux devs for what is essentially a market share problem (chicken/egg).
By the way, try upgrading a random Win98/WinMe era laptop to W2K to see how nice the hardware support is on Windows. At least one or two devices won't work because the drivers are proprietary and were never released for W2K/XP. And that's it, no chance of ever getting it working.
Does it achieve a goal that couldn't have been achieved within the GIMP codebase with less effort?
Perhaps, perhaps not. In any case, Krita is surpassing the Gimp in some areas that people have been complaining about for years in Gimp, and nothing was done. As a developer, what would you rather do, argue on the Gimp mailing list until your face turns blue about wanting to change the interface, or just start your own project? Sometimes you have to make a clean break to get new ideas implemented.
Whether FF renders faster than IE is true or not, I have no idea, but it is a good question.
A good question perhaps, but not a good statement. The original poster was claiming that IE is faster than Firefox at rendering without any proof. Since that is not what benchmarks have shown (search on google, it's pretty much a toss up between the two), and doesn't reflect many people's experiences, it is only natural that his conclusions are being questioned. Of course, it may be that IE is faster than Firefox on his computer, but it could be due to any number of external factors (buggy extensions come to mind).
Well I checked it out, and it absolutely sucks. You're trying to get an interactive framework where every request has to do a roundtrip to the server. That's completely backwards, and it shows. Try the demo and see how laggy everything is. Sure it's nice you don't have to rely on the browser's buggy javascript implementation, but the lag associated with this approach is unacceptable. It's funny how they state that not having to rely on user's possibly slow machines is an advantage of their approach. Never mind that any reasonable machine (ie, built in the last 10 years) will be far faster than the network latency involved in doing it their way.
I don't know about Charter, but I do know that MSP (Medical Services Plan in canada) does this. For 15 minutes I tried to wrestle with the voice activated menus they have, and absolutely could not figure out how to speak to a representative. Well eventually I just got pissed and started swearing at the thing. It paused for a while, then said "Ok, a representative, one moment please"
:)
Best. system. ever.
As an aside, I use AbiWord. It's free, and it has a passable MS Word import/export facility.
a qMicrosoftWordDocuments
.doc extension. (...) There are no plans to support binary MS Word export."
And out of curiosity, how many of those jobs where you created your resume in Abiword have you actually gotten?
I think you should read this (as already posted by someone else): http://www.abisource.com/twiki/bin/view/Abiword/F
"AbiWord can currently save in an MS Word compatible ".doc" format. This is done by saving as Rich Text Format (.rtf) but with a
Coincidentally, version 1.6 of KOffice was released today, with a lot of new major features for Krita.
Check it out here: http://koffice.org/announcements/announce-1.6.php
Development has certainly been active lately. Cyrille Berger has a blog where he talks about some of the new features that have gone into Krita lately. http://cyrilleberger.blogspot.com/
Debian with KDE 3.5.4 on my old box, 32MB of ram used. That's almost umbelievable - I mean, that's only possible with a stripped down KDE installation, with not a single service running.
Yeah, I was amazed myself. I think it also had something to do with a recent XOrg upgrade, as before that, memory usage was around 45MB on fresh boot. The KDE itself is not overly stripped. Although this is running on my second box, which is a PII 266 w/ 196 mb of ram, so I disabled the background image, disabled all UI effects, and turned off antialiasing. Also, things like klipper (which I have no use for anyway) are disabled. Other than that, it's fairly standard KDE.
Keep in mind that this is running on Debian, which by default has nothing running (I think before starting X, it uses about 10MB of RAM). A similar setup on a more desktop oriented distribution like Kubuntu will be using around 100MB, mostly due to little applets running, and more system services.
Speed. KDE (and Gnome) need better speed optimizations.
This is most likely your video card drivers. KDE is plenty fast, but if you dont have acceleration working in X, then everything will seem sluggish. My card is poorly supported (ATI Xpress 200m) and it makes everything seem slow.
Memory usage. The memory requirements of KDE and and Gnome are ridiculous.
Yes, if you mean ridiculously low. Fresh boot, Debian with KDE 3.5.4 on my old box, 32MB of ram used. Start up konversation (irc client) and it's about 45MB. Every subsequent application uses less extra ram, because the libraries are already loaded. Fresh boot on windows xp is at least 100MB, on my laptop more like 150. Most likely you have no idea on how to measure memory usage on linux. Have a look here: http://ktown.kde.org/~seli/memory/
Clutter.
You've got a point there, it's getting better with every release though. And no, sacrificing features for simplicity like Gnome did is not a good strategy.
Consistency.
That's one of the strengths of KDE actually. Everything works the same across the KDE apps. Keyboard shortcuts, look, general menu structure, colours, style, etc etc. And then we get into the even more important consistency, which is functional consistency. Just about every app that needs a text editor uses the same one, so they all behave the same. The same spell checking engine is used almost everywhere, and the password manager saves passwords for every application that has a need to store them. No other operating system is anywhere close to that consistent. Not OS X, not Windows, nothing.
An alternative to this is to just keep all your documents on a flash drive. Small enough to carry around, relatively cheap, and you can even carry your programs around with you. Eventually you will have to start paying a monthly fee for those online apps you use, and then it will really add up.
As a recent engineering grad, I know this is all too commonplace. I did it a lot, my friends did it a lot, almost everyone did it a lot. Sometimes the work load was just too much, and copying an assignment was the only way it could be completed without overworking yourself.
So the alternative was to not finish some assignments and get poor marks, or copy them, and then study like mad for the final.
I feel that the important thing is to learn the material. You can't cheat on a final exam, and if you know it there, then you've obviously learned the material, copying or not. Assignments are just devices to make sure that the student keeps up with the course material. If students copy them verbatim and don't understand, then that will be their problem in the final.
Of course, none of us was dumb enough to copy an assignment line for line. There were always changes made, and if a question seemed wrong and the original author couldn't explain themselves, then it was not copied. Someone was always copying some assignent or other. I always liked people copying my work, because in doing so they would very often point out mistakes I had made in the process.
Correct. However, I would rather have the ability to do something, even if it takes a bit more figuring out, rather than not being able to do it at all. Thus, while KDE could be more intuitive, it is usable for me, while Gnome is not at all. Concentrating on usability before having basic features is a mistake.
I believe as Gnome adds the features that everyone has taken for granted in other desktops, they will have trouble retaining their simple, efficient UI.
Of course they welcome this. Just look at the results page for the Coverty scans and see how many defects have been fixed in major open source projects.
http://scan.coverity.com/
Hehe, dribble charge.
The term you are looking for is Trickle Charge
What happens when the bus stops on a railway track and gets broadsided?
:)
No really, this happened to my mom when she was younger. She just broke her leg, but lots of other people on that bus died.
If you're that worried about safety, take the train!
The side effect of feeding the hungry is that it effectively destroys their entire local food production business. The farmers who previously supported themselves selling food can't compete with free and are suddenly themselves dependant on handouts to survive.
Do some reading on how the flood of donated clothes from the western world destroyed the textile industry in many areas of Africa. Handouts are a terrible long term solution.
Coke. Outmynose.
Thanks for that.
Actually they probably are sensitive enough to pick up the muscle signals going to your eyes. Muscle signals are much easier to detect than brain waves.
We have a similar device at work (headband with a couple electrodes) that will do the same thing if you set the filters to the correct frequency for the eye movement signals.
Next up was a device that measured my perscription. I had to stare at
a little picture while it zoomed in and out of focus. Apparently this
determines my exact perscription, none of that "Is this better, or that"
lens swapping. I wonder why eye doctors don't use this all the time.
My optometrist has one and uses it but says it is not nearly as accurate as a manual exam with the lenses. It's just there to give him a rough estimate as a starting point, but it tends to overprescribe.