Composite your image in another buffer. Then blt from the composite buffer into the linear back buffer and when the blt is complete, swap the front and back buffers on the next vsync? I feel that many people are starting to ignore tearing just because they want the highest possible framerate and it concerns me. Tearing is an artifact that ought to be solved
I'm pretty sure that we don't have to flip pages to remove tearing. Just do the blit on vsync instead of immediatly after frame is ready and you are done. As we can only have refresh rates of 100Hz (ok, mayby you have 200Hz display - it really doesn't make difference) we have no problems to blit fast enough that blit is ready in time CRT starts drawing next frame. One can blit 320x200x8b VGA display with a Pentium 75 MHz fast enough to make it in vsync without tearing - why couldn't you make it in 1024x768 with your GeForce or better?
The problem with your suggestion (IMO) is that your two (or more) linear buffers must be in sync (you have to draw whole screen - including all the other windows in all of your linear buffers - every time something changes. It will not increase your performance!)
_________________________
My point is, that there is no solution to this problem in general for web application developers. If you're writing a traditional application where the user is running the executable locally you can always detect document changes, set a dirty flag, and check the dirty flag before exiting. There's just no way to do this in a web browser.
Perhaps there is no way to do that in Lynx but with javascript events that should be quite easy. There is OnChange event in all form input elements that can be used to set dirty flag and then you can use OnExit (or something - is there OnClose or so?) event where you prompt() user and perhaps submit() the form for him/her if wanted. Of course some people opt to not enable javascript but IMO that's not developers fault. (I haven't checked this though and it may be that javascript cannot force window to stay open - at least if it's not signed.)
OTOH, how to overcome the problem that textarea must be defined in characters instead of percentage of available space or even pixels I have no idea. I hate when I have to use textarea of 60x10 characters on a 1600x1200 display just because there could be someone with old portable with 640x480 or less that developer has choosed to support. How about if I have huge window my textarea is also huge (say 150x80 characters with font I have selected)?
_________________________
Don't have a certified driver? Can't play the game.
I think this cannot fix the problem. If cheater wants to use some special driver (s)he will use it and there will be crack to remove that driver check.
Right way could be to send driver information string to server where it could be compared to blacklist of "cheater drivers". After identifying driver as blacklisted it's up to server what to do. It could mark mark points as cheated or disconnect after random period of time with text "cheater".
This wouldn't prevent cheaters to play and therefore there would be much less effort to crack executables or so. Making cheating public would cut the edge out off it and perhaps lesser people would do that. _________________________
It all depends on graphics adapter when it's about playing movies. For example I have big problems to play more than one movie at time with my G400. Not because my processor (K6-2 500Mhz) couldn't handle it but because G400 is unable (or only damn slow) to use overlay graphics for more than one window at time.
Playing MP3 isn't that much different - for example my computer uses much more CPU power (sometimes near 10%) playing MP3 now that I have SB Live instead of SB 64 I had earlier - but it's not because SB Live been worser but because it's drivers are worser (in CPU usage). That's all in Windows.
Under Linux it takes about 0.5% to play MP3 stream with the same computer - I can easily play that many streams I want (SB Live can mix them in hardware) but there is no reason to play multiple MP3 streams (not at least for me).
If your hardware cannot provide any support for movie playback (namely scaling) or audio channels mixing you need much more CPU power no matter how nice your OS is.
And finally IMHO Windows Media Player beats QuickTime hands down any day (Where the heck is my menu bar? Why I cannot see this in fullscreen?). I just hope we could see something like DivX;-), legally. _________________________
I would suggest near diskless solution: mount all filesystems via NFS, but run programs locally. And add harddisk for swapping in every workstation - you don't need that much RAM that way.
That way you don't have to put that much traffic in network but administration is still easy. If harddrive fails you don't lose anything - workstation may crash but you can just replace HD to restrore swap area. _________________________
Ah, so you want a completely new system that forces every program to adhere to the same interface and everything will look the same, despite the users wishes.
No, what I want is new system where fonts look good (MS Windows quality is fine for me that is good kerning and anti-aliasing), clipboard is usable and applications feel and look same. Oh yes, I want OpenGL support also. Only way to make this effectively is to implement new protocol. Anything on top of X11 won't make it right.
What I mean with feel is that I can select if keyboard shortcuts use CTRL or ALT (or META etc) instead of current style of ALT/META with Netscape, CTRL with KDE apps and perhaps something else with others. I want to be able to change any menu shortcut in the same way like in GIMP today (with the possibility to rearrange menus easily).
What I mean with look is that when I change theme in one place all applications honor it immediatly. I think that mac users would tell that themeing must include option to move menus to top of screen instead of to top of window.
If look and feel is implemented with something like GTK or QT on top of X11 there will always be apps that make it wrong. If your windowing protocols forces all applications (in the same server desktop) to look the same there won't be those ugly looking menus that differ from anything else (netscape, xemacs anyone?).
Programs like XMMS can be implemented through application specific themeing or bitmaps if really needed. Sure XMMS looks cool but it doesn't make that much difference. Now if I could select all my apps looked like XMMS it would make the difference.
"Everything will look the same" doesn't (have to) mean "everything which runs on *nix will look the same". _________________________
I wish I had moderator points left. Please moderate parent up.
I would just add that don't limit displaying to pixels. OpenGL like style with viewport defination and positioning relative to that would be much better. In a few years we will have displays with 100 - 400 ppi (perhaps stereo) and even if only our icon sizes are based on pixels in our displays we are in deep shit.
of the complaints are about values that aren't encapsulated in quotes, which is completely legal
That's right. But they are really legal only if string contains only basic characters ([a-zA-Z0-9_.:-]). Notice that those characters do not contain slash used in hrefs so those errors are correct.
There is come "things" in validator like that it doesn't allow wrap="virtual" in <textarea> tag when using XHTML 1.0 Transitional. I have no idea how to make textarea working equally to that parameter with CSS (is there some other way?) and still maintaining compability with NS 4.x.
Another point worth mentioning is that even in those enclosed strings use of & is not allowed because of possible extensions in future; So you have to write <a href="./script.cgi?key1=value1&key2=value2"> instead. Yes, you have to replace & with &.
And about how great mozilla is: try to make XHTML 1.0 compatible document with a form (remember to include <!DOCTYPE... defination.) Look at those weird forms. My own cgi scripts return XHTML compatible documents but replace DOCTYPE with HTML 4.0 Transitional if client is Mozilla/5.0 to handle this bug.
W3C standards are one thing and real world browsers are another. I try to follow W3C rules but I have to bend them slightly to make pages work with current crop of browsers.
And Mozilla is poised to become the browser of choice for embedded applications. It's light, it's fast, and it's free.
I hope it too. It will be free - no browser can cost today (even Opera is trying it and to be honest their browser is actually pretty good). I'm not sure about fast part - sure it will be faster than NS4.x. One has hard times to be slower. But will it be faster than IE5.x or Opera? And finally I wouldn't call Mozilla light, at least now. It takes 20 MB of memory to just to start and memory usage like 150 MB isn't that unusual - even with pretty small documents.
Sure Mozilla is still in development but I'm afraid that it won't be even close to ready any time soon. Perhaps after a year or so.
And yes - I have tried Mozilla (milestones and nightly builds) myself both in Linux and Windows (and reported bugs) and in my experience it takes very much memory and is much slower than IE5. Bottom line: IE5.x is not good but it is still better than any of its competitors.
The good thing about Netscape 4.x is that when you do CGI programming and you succeed to make it work with NS it will work with other browsers too. _________________________
With technology like ClearType, there is no reason that a page can't be displayed at just as clearly to the human eye at 100ppi as at 200ppi
Sorry but this is certainly wrong. Say we have 10"x10" 200ppi display (2000x2000 pixels) and we have 500 lines of text on it. It has now 4 pixels for each line of text - with good antialiasing it might be barely readable. With 100 ppi you have only 2 pixels per line - there is no way text could be recognizable. Now if we could have 400 ppi display still with 10"x10" we would have 8 pixels per line which could be readable even without antialiasing (not looking great though). So display with more ppi is obviously better. Why do you think that 1200 dpi printer makes better result than 300 dpi printer?
About physical size I would only comment that it only makes difference if you cannot select distance you look at your display from. You may use 1280x1024 head mounted display and it feels equally big (if not bigger) with 19" display on your desktop.
IMHO 10"x10"x200ppi is equal to 20"x20"x100ppi when it comes to for what you can use it. You just need to use it from different distance.
And bigger monitor with the same resolution (instead of ppi value) will be more expensive. Reason: you cannot decrease amount of light you create per area unit (because display would look darker otherwise). Because all known ways need more energy for brighter light when using same technology it will cost more to produce bigger display (for example in case of normal CRT you need to shoot much more electrons).
This isn't to say that smaller display is cheaper because there is limit when it's too hard to create smaller device. I think I would be happy with 300ppi 19" monitor. _________________________
Interesting, but how do we (the Linux community) avoid doing the same thing?
I think we are doing the opposite already. Think about X11 protocol: everybody agrees that it sucks - no alpha, anti-aliasing etc. Creating a new protocol (X12?) would be a solution but we are not doing it because it would broke our existing applications (including but not limited to gnome and kde). Other examples exist.
PS. To be exact X11 is a protocol and xlib is an API and we could change to X12 or something if we port xlib to that. Why this hasn't happened already? Is it too hard? Are we happy with current situation (I'm not!)? I don't know. I'm sure you will tell if you do. _________________________
I think the real reason that people aren't using SMS messages that much compared to talking in Finland is that one SMS (max 160 characters) costs 0.17 EUR. On the other hand it costs about the same to talk for one minute. Many people can speak worth more than 160 characters in a minute (and calls are charged on second based - if you don't spend whole minute it's cheaper).
SMS messages are still used despite the fact that they cost much more. Mayby it's because of asyncronous nature of those. Plus cases where you are sending a phone number of somebody or street address or something that usually is written down while listening.
SMS messages *could* be practically free (I would be happy with 0.01 EUR/message) but because Finns are using those no matter how expensive those are there is no reason for operators to cut down the prizes. After all we are even paying 1.3 EUR for a litre of gasoline. _________________________
the hockey puck has coloured sides and a dent in the button
Maybe you didn't notice that he especially stated he doesn't want to *look* at the mouse. I don't know about you but I don't look at my keyboard while typing - why should I look at my mouse to see where it's front is? No matter how colorful its sides were. Take a look at Logitech mouses - you never have to wonder where the mouse is pointing. Just put your hand on top of it and you immediatly know where the mouse's front is.
Nowadays I find it hard to live even without wheel in my mouse - how could I live with only one button?
By the way do you ever wonder why there is those steps in the wheel? Wheel could be very accurate tool (z-axis adjustments) if it also had optical sensors like normal mouse movement. (I know... it's because that ancient protocol only has 2 axis). _________________________
The drivers are reasonably solid and good, but running GL screensavers is asking for trouble
IMHO "solid and good" doesn't equal "disable OpenGL to make it work". One reason I use Linux instead of windows is that I can run broken program (read my own program with memory allocation error or something) without taking system down - instead I get a nice core dump for debugging if I want so. Now you are suggesting that because XFree 4.X is buggy I should limit my choice of programs to start because my userland programs can take my system down with the assistance of XFree 4.X. Uhh, no thanks.
I'm aware that XFree 3.X.X isn't that flawless either but I don't have to pray each time I use glx extension... (I have g400). _________________________
need 256MB of RAM and I could afford 196MB of real RAM or 128MB of real RAM plus the doubler
If you can afford 128MB ram and a application/solution that requires 256MB I'm pretty sure you can afford that another 128MB also. Everybody seems to agree that scsi (and raid?) are required for servers - now how can you guys buy that scsi adapter but cannot afford 512MB of memory?
Only case where this makes sense is when your hardware cannot support required amount of memory - say you have x86 platform with 2GB of memory and your mobo cannot support more. If your requirements are 2.2GB compressing would be the solution. However I would suggest better mobo in case like this.
Hardware for swapping could be much better idea. Though I cannot imagine how hardware (in MMU?) could make swapping faster if you are still using same hard drives. Perhaps specialized harddrive for swapping (smaller and faster) and own bus for swapping. But things like that costs and you could buy real memory instead.
Bottom line: having enough real memory is the only way to go. _________________________
Encode the first song "Shaky Ground" at 128kbps/44000Hz.
One thing that should be remembered is that MP3 is a lossy method and therefore it does matter which program/algorithm you use to generate your mp3 stream. Sure if you use some poor encoder your results are similarly poor. I am pretty pleased with lame quality and you should try that also before bashing mp3. IMHO if you think music encoded with lame sounds terrible you are unfortunate enough to be born with golden ears.
You should try a few decoders also but it shouldn't make that much difference. _________________________
Just one question. How can Microsoft hurt your typical graphics card manufacturer ( say Matrox or 3dfx)?
Take for example DX8. There are many new features close to nVidia's hardware (because nVidia currently has the best hardware features and nVidia is in close co-operation with MS). Now if MS doesn't like ATI and opts to not include API for vertex blending (supported by ATI hardware) then game developers won't use vertex blending and there is no advantage for ATI cards to include that feature. No matter how much faster/better looking games could be if it was supported by DX. _________________________
In theory, an output from one vendor's fully-compliant HTML/CSS user agent should be identical to another vendor's.
This works also in practise. But we still need to see first fully-compliant HTML/CSS user agent. Mozilla could be it after 8000 bug fixes. _________________________
What? BR-tag marks forced line break. AFAIK spec says that repeated BR-tags should be replaced with only one. You cannot design HTML pages - and you aren't supposed to - with absolute values in everything. Viewer possibly doesn't have even font you would want to use or cannot use color you defined. Possibly he/she cannot even see the page - I hope your ALT-tags are correct.
As stated before HTML (or at least should) describes only document structure (no matter there is formatting tags/attributes in spec) and CSS should be used for ALL formatting. Note that those tables are not for formatting. Though I use those for formatting too. But I do it only to support those brain dead browsers. Any Netscape 5- users out there?
About point 2) I think that vector graphics should be distributed in EPS (Encapsulated PostScript) format. It doesn't support animation (at least that I know) but I yet have to see page that has any extra information in vector graphics animation. (No, if your menus roll, blink and zoom when you come visible it is NOT extra information I want.) There could be better format but it needs to be open.
Item 3) in your list is true but I don't think good design has anything to do with absolute positioning and stuff like that. We should NOT mess up with GUI and UI either because as I see it HTML is not limited to 2D displays with paper like formatting (think about blind people for example). After all it's only headers and text paragraphs. _________________________
That's because there aren't any "additions." Word 2000 is 100% backward-compatible with the Word 97 format.
AFAIK word2k supports tables inside table cells and word97 doesn't - how that can be 100% compatible? There is some other (small?) changes also. I for one wouldn't edit word2k document with word97 and still hope to not broke "formatting". I admit that you can see information inside word2k file with word97 and that's much better than with word95/word97. However, opening.doc file format would be The Right Thing. _________________________
IMHO MPEG4 with the same bit rate as MPEG2 does look a lot better. The problem isn't the image quality but processing requirements. I'm currently not aware of system that could decode 9Mbps MPEG4 stream (real time) - if you have one I would want to know what it includes. According to what I have seen I would say MPEG4 has MPEG2 quality with half (or less) the bit rate.
Of course because MPEG4 is able to decode stream with decreased image quality with less processing power you may get lower quality output if don't have enough processing power. And because there are many ways to generate MPEG4 just like there are many ways to generate MP3, bad encoding phase will ruin your image quality. _________________________
Everybody knows this allready but I wanted to remind: Alpha may have faster fp ops than x86 but it shows in prize tag. There is no better speed/$$$ ratio than that of x86.
Scientific applications are often highly paraller and should be using clusters anyways. _________________________
Re:x86 is popular to hate, but not that bad really
on
Is The x86 Obsolete?
·
· Score: 1
It seems to me it would be insane to abandon backward compatibility for the sake of increased efficiency. Not only would it be insane, but to give up on hundreds of millions lines of code just to have your system run slightly faster would be a commercial suicide.
Correction: give up on hundreds of millions lines of code just to have your system run slightly slower would be commercial suicide.
Reason: do you really think that creating an OS with a good compiler for a new processor is that easy? Think about gcc - it compiles pretty nice code for x86 but performace is poor for example for MIPS or so. IA-64 will have new core but I think it won't be fast with first generation compilers.
This increased efficiency is pretty much theoretical. It does matter for power usage but for real world computing having good compiler is what makes difference. IMHO Intel still sells their processors despite of AMD because compilers cannot generate good code for Athlon - and of course legacy programs will be a problem even when we have better compiler.
I do, however, agree that we could throw away all those older opcodes, including all those 16 bit instructions. I wish we had better way to boot our systems nowadays (that's where we need those 16 bit instructions anyway). On the other hand, I would add instructions to make operations from memory address to memory address because L1 cache is equally fast to registers - using memory addresses instead of registers you would have virtually unlimited pool of registers. _________________________
Integrating static widget set in server is Very Bad Idea(tm). Creating loadable widget set with java (processor independent, this is supposed to be remote protocol) server programs to render homegrown widgets in server end is much better idea. (I'm not counting things like rectangle, bezier line, polygon or OpenGL area as widgets.)
The problem is who/what should be allowed to upload widgets into server? Application? GUI Library? Window Manager? IMHO right way would be window manager because that's the only way to make sure all apps have the same look. Legacy programs would be pain in ass though. _________________________
Just wanted to add that it's true that this works (kind of) with analog LCD display also. The problem, however, is that with analog signal you cannot sync RGB values perfectly into RGB subpixels cells - this is why we are going towards digital connections.
Analog LCD display is in between CRT and digital LCD in image quality. Part of image quality is because of subpixel addressing and another part of quality is because of the same antialiasing as with CRT.
Because this technology does not (fully) work with colored background the only real solution is to increase real resolution. IMHO this is like compressing your harddrive instead of getting bigger one. _________________________
Slashdot Mirror
I'm pretty sure that we don't have to flip pages to remove tearing. Just do the blit on vsync instead of immediatly after frame is ready and you are done. As we can only have refresh rates of 100Hz (ok, mayby you have 200Hz display - it really doesn't make difference) we have no problems to blit fast enough that blit is ready in time CRT starts drawing next frame. One can blit 320x200x8b VGA display with a Pentium 75 MHz fast enough to make it in vsync without tearing - why couldn't you make it in 1024x768 with your GeForce or better?
The problem with your suggestion (IMO) is that your two (or more) linear buffers must be in sync (you have to draw whole screen - including all the other windows in all of your linear buffers - every time something changes. It will not increase your performance!)
_________________________
Perhaps there is no way to do that in Lynx but with javascript events that should be quite easy. There is OnChange event in all form input elements that can be used to set dirty flag and then you can use OnExit (or something - is there OnClose or so?) event where you prompt() user and perhaps submit() the form for him/her if wanted. Of course some people opt to not enable javascript but IMO that's not developers fault. (I haven't checked this though and it may be that javascript cannot force window to stay open - at least if it's not signed.)
OTOH, how to overcome the problem that textarea must be defined in characters instead of percentage of available space or even pixels I have no idea. I hate when I have to use textarea of 60x10 characters on a 1600x1200 display just because there could be someone with old portable with 640x480 or less that developer has choosed to support. How about if I have huge window my textarea is also huge (say 150x80 characters with font I have selected)?
_________________________
I think this cannot fix the problem. If cheater wants to use some special driver (s)he will use it and there will be crack to remove that driver check.
Right way could be to send driver information string to server where it could be compared to blacklist of "cheater drivers". After identifying driver as blacklisted it's up to server what to do. It could mark mark points as cheated or disconnect after random period of time with text "cheater".
This wouldn't prevent cheaters to play and therefore there would be much less effort to crack executables or so. Making cheating public would cut the edge out off it and perhaps lesser people would do that.
_________________________
Playing MP3 isn't that much different - for example my computer uses much more CPU power (sometimes near 10%) playing MP3 now that I have SB Live instead of SB 64 I had earlier - but it's not because SB Live been worser but because it's drivers are worser (in CPU usage). That's all in Windows.
Under Linux it takes about 0.5% to play MP3 stream with the same computer - I can easily play that many streams I want (SB Live can mix them in hardware) but there is no reason to play multiple MP3 streams (not at least for me).
If your hardware cannot provide any support for movie playback (namely scaling) or audio channels mixing you need much more CPU power no matter how nice your OS is.
And finally IMHO Windows Media Player beats QuickTime hands down any day (Where the heck is my menu bar? Why I cannot see this in fullscreen?). I just hope we could see something like DivX;-), legally.
_________________________
That way you don't have to put that much traffic in network but administration is still easy. If harddrive fails you don't lose anything - workstation may crash but you can just replace HD to restrore swap area.
_________________________
No, what I want is new system where fonts look good (MS Windows quality is fine for me that is good kerning and anti-aliasing), clipboard is usable and applications feel and look same. Oh yes, I want OpenGL support also. Only way to make this effectively is to implement new protocol. Anything on top of X11 won't make it right.
What I mean with feel is that I can select if keyboard shortcuts use CTRL or ALT (or META etc) instead of current style of ALT/META with Netscape, CTRL with KDE apps and perhaps something else with others. I want to be able to change any menu shortcut in the same way like in GIMP today (with the possibility to rearrange menus easily).
What I mean with look is that when I change theme in one place all applications honor it immediatly. I think that mac users would tell that themeing must include option to move menus to top of screen instead of to top of window.
If look and feel is implemented with something like GTK or QT on top of X11 there will always be apps that make it wrong. If your windowing protocols forces all applications (in the same server desktop) to look the same there won't be those ugly looking menus that differ from anything else (netscape, xemacs anyone?).
Programs like XMMS can be implemented through application specific themeing or bitmaps if really needed. Sure XMMS looks cool but it doesn't make that much difference. Now if I could select all my apps looked like XMMS it would make the difference.
"Everything will look the same" doesn't (have to) mean "everything which runs on *nix will look the same".
_________________________
I wish I had moderator points left. Please moderate parent up.
I would just add that don't limit displaying to pixels. OpenGL like style with viewport defination and positioning relative to that would be much better. In a few years we will have displays with 100 - 400 ppi (perhaps stereo) and even if only our icon sizes are based on pixels in our displays we are in deep shit.
_________________________
That's right. But they are really legal only if string contains only basic characters ([a-zA-Z0-9_.:-]). Notice that those characters do not contain slash used in hrefs so those errors are correct.
There is come "things" in validator like that it doesn't allow wrap="virtual" in <textarea> tag when using XHTML 1.0 Transitional. I have no idea how to make textarea working equally to that parameter with CSS (is there some other way?) and still maintaining compability with NS 4.x.
Another point worth mentioning is that even in those enclosed strings use of & is not allowed because of possible extensions in future; So you have to write <a href="./script.cgi?key1=value1&key2=value2"> instead. Yes, you have to replace & with &.
And about how great mozilla is: try to make XHTML 1.0 compatible document with a form (remember to include <!DOCTYPE... defination.) Look at those weird forms. My own cgi scripts return XHTML compatible documents but replace DOCTYPE with HTML 4.0 Transitional if client is Mozilla/5.0 to handle this bug.
W3C standards are one thing and real world browsers are another. I try to follow W3C rules but I have to bend them slightly to make pages work with current crop of browsers.
This comment is XHTML 1.0 compatible.
_________________________
I hope it too. It will be free - no browser can cost today (even Opera is trying it and to be honest their browser is actually pretty good). I'm not sure about fast part - sure it will be faster than NS4.x. One has hard times to be slower. But will it be faster than IE5.x or Opera? And finally I wouldn't call Mozilla light, at least now. It takes 20 MB of memory to just to start and memory usage like 150 MB isn't that unusual - even with pretty small documents.
Sure Mozilla is still in development but I'm afraid that it won't be even close to ready any time soon. Perhaps after a year or so.
And yes - I have tried Mozilla (milestones and nightly builds) myself both in Linux and Windows (and reported bugs) and in my experience it takes very much memory and is much slower than IE5. Bottom line: IE5.x is not good but it is still better than any of its competitors.
The good thing about Netscape 4.x is that when you do CGI programming and you succeed to make it work with NS it will work with other browsers too.
_________________________
Sorry but this is certainly wrong. Say we have 10"x10" 200ppi display (2000x2000 pixels) and we have 500 lines of text on it. It has now 4 pixels for each line of text - with good antialiasing it might be barely readable. With 100 ppi you have only 2 pixels per line - there is no way text could be recognizable. Now if we could have 400 ppi display still with 10"x10" we would have 8 pixels per line which could be readable even without antialiasing (not looking great though). So display with more ppi is obviously better. Why do you think that 1200 dpi printer makes better result than 300 dpi printer?
About physical size I would only comment that it only makes difference if you cannot select distance you look at your display from. You may use 1280x1024 head mounted display and it feels equally big (if not bigger) with 19" display on your desktop.
IMHO 10"x10"x200ppi is equal to 20"x20"x100ppi when it comes to for what you can use it. You just need to use it from different distance.
And bigger monitor with the same resolution (instead of ppi value) will be more expensive. Reason: you cannot decrease amount of light you create per area unit (because display would look darker otherwise). Because all known ways need more energy for brighter light when using same technology it will cost more to produce bigger display (for example in case of normal CRT you need to shoot much more electrons).
This isn't to say that smaller display is cheaper because there is limit when it's too hard to create smaller device. I think I would be happy with 300ppi 19" monitor.
_________________________
I think we are doing the opposite already. Think about X11 protocol: everybody agrees that it sucks - no alpha, anti-aliasing etc. Creating a new protocol (X12?) would be a solution but we are not doing it because it would broke our existing applications (including but not limited to gnome and kde). Other examples exist.
PS. To be exact X11 is a protocol and xlib is an API and we could change to X12 or something if we port xlib to that. Why this hasn't happened already? Is it too hard? Are we happy with current situation (I'm not!)? I don't know. I'm sure you will tell if you do.
_________________________
SMS messages are still used despite the fact that they cost much more. Mayby it's because of asyncronous nature of those. Plus cases where you are sending a phone number of somebody or street address or something that usually is written down while listening.
SMS messages *could* be practically free (I would be happy with 0.01 EUR/message) but because Finns are using those no matter how expensive those are there is no reason for operators to cut down the prizes. After all we are even paying 1.3 EUR for a litre of gasoline.
_________________________
Maybe you didn't notice that he especially stated he doesn't want to *look* at the mouse. I don't know about you but I don't look at my keyboard while typing - why should I look at my mouse to see where it's front is? No matter how colorful its sides were. Take a look at Logitech mouses - you never have to wonder where the mouse is pointing. Just put your hand on top of it and you immediatly know where the mouse's front is.
Nowadays I find it hard to live even without wheel in my mouse - how could I live with only one button?
By the way do you ever wonder why there is those steps in the wheel? Wheel could be very accurate tool (z-axis adjustments) if it also had optical sensors like normal mouse movement. (I know... it's because that ancient protocol only has 2 axis).
_________________________
IMHO "solid and good" doesn't equal "disable OpenGL to make it work". One reason I use Linux instead of windows is that I can run broken program (read my own program with memory allocation error or something) without taking system down - instead I get a nice core dump for debugging if I want so. Now you are suggesting that because XFree 4.X is buggy I should limit my choice of programs to start because my userland programs can take my system down with the assistance of XFree 4.X. Uhh, no thanks.
I'm aware that XFree 3.X.X isn't that flawless either but I don't have to pray each time I use glx extension... (I have g400).
_________________________
If you can afford 128MB ram and a application/solution that requires 256MB I'm pretty sure you can afford that another 128MB also. Everybody seems to agree that scsi (and raid?) are required for servers - now how can you guys buy that scsi adapter but cannot afford 512MB of memory?
Only case where this makes sense is when your hardware cannot support required amount of memory - say you have x86 platform with 2GB of memory and your mobo cannot support more. If your requirements are 2.2GB compressing would be the solution. However I would suggest better mobo in case like this.
Hardware for swapping could be much better idea. Though I cannot imagine how hardware (in MMU?) could make swapping faster if you are still using same hard drives. Perhaps specialized harddrive for swapping (smaller and faster) and own bus for swapping. But things like that costs and you could buy real memory instead.
Bottom line: having enough real memory is the only way to go.
_________________________
One thing that should be remembered is that MP3 is a lossy method and therefore it does matter which program/algorithm you use to generate your mp3 stream. Sure if you use some poor encoder your results are similarly poor. I am pretty pleased with lame quality and you should try that also before bashing mp3. IMHO if you think music encoded with lame sounds terrible you are unfortunate enough to be born with golden ears.
You should try a few decoders also but it shouldn't make that much difference.
_________________________
Take for example DX8. There are many new features close to nVidia's hardware (because nVidia currently has the best hardware features and nVidia is in close co-operation with MS). Now if MS doesn't like ATI and opts to not include API for vertex blending (supported by ATI hardware) then game developers won't use vertex blending and there is no advantage for ATI cards to include that feature. No matter how much faster/better looking games could be if it was supported by DX.
_________________________
This works also in practise. But we still need to see first fully-compliant HTML/CSS user agent. Mozilla could be it after 8000 bug fixes.
_________________________
What? BR-tag marks forced line break. AFAIK spec says that repeated BR-tags should be replaced with only one. You cannot design HTML pages - and you aren't supposed to - with absolute values in everything. Viewer possibly doesn't have even font you would want to use or cannot use color you defined. Possibly he/she cannot even see the page - I hope your ALT-tags are correct.
As stated before HTML (or at least should) describes only document structure (no matter there is formatting tags/attributes in spec) and CSS should be used for ALL formatting. Note that those tables are not for formatting. Though I use those for formatting too. But I do it only to support those brain dead browsers. Any Netscape 5- users out there?
About point 2) I think that vector graphics should be distributed in EPS (Encapsulated PostScript) format. It doesn't support animation (at least that I know) but I yet have to see page that has any extra information in vector graphics animation. (No, if your menus roll, blink and zoom when you come visible it is NOT extra information I want.) There could be better format but it needs to be open.
Item 3) in your list is true but I don't think good design has anything to do with absolute positioning and stuff like that. We should NOT mess up with GUI and UI either because as I see it HTML is not limited to 2D displays with paper like formatting (think about blind people for example). After all it's only headers and text paragraphs.
_________________________
AFAIK word2k supports tables inside table cells and word97 doesn't - how that can be 100% compatible? There is some other (small?) changes also. I for one wouldn't edit word2k document with word97 and still hope to not broke "formatting". I admit that you can see information inside word2k file with word97 and that's much better than with word95/word97. However, opening .doc file format would be The Right Thing.
_________________________
Of course because MPEG4 is able to decode stream with decreased image quality with less processing power you may get lower quality output if don't have enough processing power. And because there are many ways to generate MPEG4 just like there are many ways to generate MP3, bad encoding phase will ruin your image quality.
_________________________
Scientific applications are often highly paraller and should be using clusters anyways.
_________________________
Correction: give up on hundreds of millions lines of code just to have your system run slightly slower would be commercial suicide.
Reason: do you really think that creating an OS with a good compiler for a new processor is that easy? Think about gcc - it compiles pretty nice code for x86 but performace is poor for example for MIPS or so. IA-64 will have new core but I think it won't be fast with first generation compilers.
This increased efficiency is pretty much theoretical. It does matter for power usage but for real world computing having good compiler is what makes difference. IMHO Intel still sells their processors despite of AMD because compilers cannot generate good code for Athlon - and of course legacy programs will be a problem even when we have better compiler.
I do, however, agree that we could throw away all those older opcodes, including all those 16 bit instructions. I wish we had better way to boot our systems nowadays (that's where we need those 16 bit instructions anyway). On the other hand, I would add instructions to make operations from memory address to memory address because L1 cache is equally fast to registers - using memory addresses instead of registers you would have virtually unlimited pool of registers.
_________________________
The problem is who/what should be allowed to upload widgets into server? Application? GUI Library? Window Manager? IMHO right way would be window manager because that's the only way to make sure all apps have the same look. Legacy programs would be pain in ass though.
_________________________
Analog LCD display is in between CRT and digital LCD in image quality. Part of image quality is because of subpixel addressing and another part of quality is because of the same antialiasing as with CRT.
Because this technology does not (fully) work with colored background the only real solution is to increase real resolution. IMHO this is like compressing your harddrive instead of getting bigger one.
_________________________