Re:Why do we measure things with money?
on
Star Wars Sickout
·
· Score: 1
I'm no economist, but shouldn't the market adjust to the constraints? Something should give, such as reduced pay or increased cost of the product.
I guess, though, that I'm not an economist for a reason.
Re:Why do we measure things with money?
on
Star Wars Sickout
·
· Score: 1
I've often thought that it would be better if everyone worked half the hours they currently do and thus twice as many people were in the workforce. Of course, it isn't as simple as that as you'd still get people who want to work full time, and they'd have more money and thus drive up the prices of everything until everyone has to work more hours.
It's a nice thought, though. It annoys me that most of the people in the western world spend more time working than they do anything else.
I wasn't really suggesting that you have to take Computing to get into a CS degree... I just wanted to make the distinction between ICT and Computing. When it comes to university applications, the fussiest they seem to get is that some of them require a maths A-level. You'd probably have to go to a college to do A-Level Computing, as it's a bit specialist for a school to cover.
Back when I did all this stuff we only did three A-Levels, and I did Physics, Maths and Geography. I turned out okay. Good luck. ;)
Remember that what you're doing is ICT, which was never supposed to be about computers in and of themselves. It's about the application of computers and other technology to do lots of things involving the words "streamline" and "optimise".
If you want to do programming and other specialist stuff like that, it's not ICT that you're after: it's computing. Later it specialises even more and becomes "computer science". At GCSE level no-one is anywhere near specialised enough to do a course on programming. Sure, it's frustrating to go through ICT when you've got an interest in computing as an end unto itself, but just grin and bear it and it'll all be over soon. I'm just coming to the end of a masters degree and looking back my IT GCSE (as it was called then) was just a small part of the rich set of subjects I touched on at GCSE level.
Get your GCSEs out of the way and then enrol in an A-Level Computing course. There's still some ICT in there at A-Level, but it'll give you a little to get your teeth into at least, and from there (along with a few other A-Levels, of course) you can go on to do Computer Science at a university of your choice, and that is where the fun starts.
I wonder how long it'll be before people create alternative implementations of the client portion of this. I think it'd be pretty neat, for example, to have it as a network-wide proxy on my server here so that everyone in the house can share the benefits of the local caching. We're all geeks, so we all load things like slashdot all the time. Of course, a ghetto solution for now would be to run it under Wine and arrange for Squid to forward requests to it.
It would also be nice to see other implementations of the server part. Sure, it won't be as fast as running it at Google because they've got a crawler cache and a fat pipe, but it would still be cool to have a bunch of different servers to choose from. ISPs could run such a server for their customers to speed up the DSL or Cable bit of the transfer, for example.
I do like the idea of AR. The simplest entertainment-oriented implementation that springs to mind is similar to the "laser gun" games that were popular in the 90s -- Quasar and LaserQuest were the big franchises here in the UK. Take a large area and dec it up with something resembling a sci-fi set, then send in a bunch of people wearing AR goggles (which can be made to look like head armour) and have them work together to fight computer-generated aliens in the environment. Unlike Quasar and LaserQuest, there could actually be visible projectiles and discharges, and a variety of different weapons. Depending on how fancy the tech is, there could even be scorch marks added to the walls.
Of course, there are still other things not accounted for. If the aliens shoot at you, how do you feel the impact? What happens if a player walks through an alien? It's still interesting, though, and I for one would probably enjoy the game even if it was a little unrealistic as long as it is fair and doesn't cause motion sickness.
If you have Windows configured so that all of the Explorer windows share a single process, all of the windows necessarily run as the same user as well. You just have to enable "Open folder windows as a separate process" which is in the options somewhere. (I don't have a Windows XP machine handy to find it right now.)
My Nokia 3410 is serving me well. The battery lasts a week (and I've had it for a year or so now), it has a black and white screen and it does phone-like stuff. They still seem to be on sale, too -- in the UK, at least.
As you wrote it, the combination of potion_a and potion_c would have been valid, because ! potion_a would have returned false. Of course, there's also the fun perl construct:
wrong_potion() unless potion_a && potion_b;
(If you don't like perl's postfix if shorthand, feel free to flip that back round "the right way" and add the braces.) (The word "potion" is one of those strange words that just always looks like it's spelled wrong. Maybe my brain is trying to read it as "option"!:))
IE blends PNG images with an alpha channel against their default background color. When there is no default, it uses one of the system colors that happens to be the default for GDI surfaces.
You can give IE a better (but still solid) color to blend against by having pngcrush write in a bKGD chunk with the desired color. This doesn't help you if you are trying to blend against a texture, but it's handy if, as in your case, the image is blended against a solid background color anyway.
In most cases, playing around with the permissions on the install directory or installing it somewhere that the user has write access allows it to work. If it needs to write to the Windows directory to run, though, it's time to get a better application.
Also, have all of the inexperienced users (and ideally all of the experienced users too) set as limited user accounts. Even if someone needs to be able to install something, it's easy enough to run an installer as another user account.
Now that I've left home my family's computer doesn't get the regular maintenence attention that it once did, so I've set everyone as a limited user account and created an "Admin" account which is used for software installation and other such tasks. I popped home over Easter and took the opportunity to check up and make sure nothing had slipped through, and sure enough everything was fine. What I did find, though, was lots of malware and worm installers in my brother's Internet Explorer cache, which would have been installed and running if he had access to write to the Windows directory.
One of these days worm and spyware authors might wise up and install their software in the user's own file area and run from the current user startup list, but at least then a user on a shared machine is only screwing himself. If it comes to it, you can just back up the documents and nuke the user account.
It's been a while since I used Word, but I'm sure there was an option to disable the squiggly lines "as you type". You can still perform a spelling and grammar check when it is convenient to you by pressing F7 and/or Shift+F7, or by choosing the appropriate options from the menus. It'll then hop through the document flagging each possible error, giving you a list of suggestions and the opportunity to edit in a much more efficient interface than that used for as-you-type corrections.
Back in the days when Microsoft's best video player was the ActiveMovie player (hah!), OEMs bundled alternative players. My family's PC came with some weird-looking application which probably isn't around anymore (I forget the name) and I knew friends who got PCs bundled with Winamp and QuickTime Player.
Microsoft could easily have released Media Player as a separate application, and assuming it was good enough to compete with the other players around at the time OEMs would have bundled it in preference. See, OEMs know that user's want a system that "does stuff" out of the box, which is why they bundle all that value-add stuff like an office application, voicemodem software and in the past sound and video playback software. Microsoft didn't have to bundle it themselves; it could have competed fairly in the marketplace for a position in the default install of Windows, just like Microsoft Office did.
You're getting Media Player and DirectShow confused. DirectShow is the standard Windows API for video playback. Windows Media Player is an application which provides a UI for video playback based on the DirectShow API. The codecs live inside DirectShow. The Media Player ActiveX control confuses things a little, but is really a way to embed the Media Player user interface into your application or document, not just the bare video playback functionality.
If MS removes Media Player, they break the ActiveX control, which breaks wep pages and documents which embed that ActiveX control. However, if those documents instead just referenced the video data and left it up to the viewer application to determine an appropriate player, there would be no problem as users could install another player and have it take over the duties that Media Player would have performed by default. Embedding Media Player directly is generally a bad idea anyway, since clearly that will only work on Windows systems. (and not even all Windows systems, now.)
In addition, applications which depend on DirectShow for video playback should go on working. This includes Winamp, mplayer (in part), DirectX Player as well as a load of games and other applications that aren't directly video-related but still playback video for some reason. If Microsoft has removed DirectShow, then it is really trying to be facetious, as clearly when the courts ruled that Media Player be unbundled they were referring to the Media Player application and not the API on which it depends. By that argument, they should have removed the Win32 API as well.
One way to embed video clips into Office documents is to embed the Media Player ActiveX control. Of course, without Media Player there is no Media Player ActiveX control and so documents using this technique won't load correctly without Media Player.
I've not used Word in years, but I'm going to assume that there's also a second way which involves embedding the video just as video data, without any particular container. Now, I'd expect those to play back through DirectShow (the API Media Player uses to play video) not Media Player itself, and so it should go on working just like any game which uses DirectShow for movie playback should go on working, and Winamp (which plays back most filetypes through DirectShow) should go on working.
Therefore there should be no dependence on any particular player frontend, but you'll still only be able to play movies which have a registered DirectShow codec. Since DirectShow is the standard Windows API for video playback, this is sensible. That a bunch of video format owners (Real, Quicktime) don't distribute DirectShow codecs is their fault, not Microsoft's. Of course, if Office applications really don't have a way to embed video directly without using a specific player ActiveX control then I would describe the developers as incompetant rather than claim sabotage.
I suppose a final possibility is that Microsoft heard "Remove Media Player!" and went and stripped out DirectShow. That would be malicious in my mind since DirectShow is the standard API for video playback in Windows and so its removal would break loads of applications. It was the Media Player application that was to be removed, not the APIs it uses. By that logic, the Win32 API should be removed as well as it's clearly part of Media Player!
What I had in mind was something like the browser indicating it has support for XUL in the Accept header, or perhaps just some way to deliver both in the same response and have Mozilla ignore the HTML and others ignore the XUL. One more possibility, which does have the overhead of an additional request, is to include a LINK element in the HTML document that indicates that there is an alternative XUL version of the page. Mozilla would then spot this and use the XUL document in preference.
All of these require browser changes, but since Mozilla is currently the only lot with XUL support they get to set some standards in this respect. I think it'd be worth it since it would encourage the uptake of XUL and thus attract more users. Hopefully in the long term we'll see other browsers implementing XUL support and we can finally be rid of the ugly hacks people do to get an HTML page to act like an application.
XUL could almost certainly do what you're describing. However, no-one will actually bother to do it firstly because no-one wants to use a word-processor embedded in a web browser and secondly Mozilla's marketshare isn't big enough for anyone to consider such a product.
A slightly more realistic goal would be to get websites implementing a superior XUL interface in addition to the HTML one. Photo gallery sites could implement a proper GUI for organising arranging pictures, for example.
This would be made much easier if there was a way to transparently provide Mozilla users with a XUL interface while everyone else gets a HTML interface. Until that can be done without any input from the user (who won't necessarily even understand what XUL is) it won't get done. If someone can find a way to do that, I'll certainly push for getting it added to software and sites that I have some influence over.
Implementing a new template doesn't require touching any Perl code. The submitter simply needs to provide a new set of Template Toolkit templates.
The problem stems from the fact that slashdot's pages are generated from tons of little bits of template, and so taking someone's hack of a static front page doesn't really get you very far to having a working implementation.
INTEGER
A 16-bit signed integer variable.
LONG
A 32-bit signed integer variable.
SINGLE
A single-precision 32-bit floating-point variable.
DOUBLE
A double-precision 64-bit floating-point variable.
STRING * n%
A fixed-length string variable n% bytes long.
STRING
A variable-length string variable.
Sounds to me like you were using a DOUBLE where you should have been using a LONG. There's no such thing as a "double-precision integer", because integers don't have precision, they just have a maximum and minimum value.
(Those descriptions came straight out of QBasic's online help, which for some reason I still have lying around.)
Getting their CSS support up to par isn't an easy job. Lots of it is broken because it is just mapping CSS properties onto parameters for old rendering code, which is why there are lots of little quirks in the box model support. Lots of it is unsupported and would require massive changes to the rendering engine to implement.
I'd much rather that MS save CSS fixes for a time when they are able to do it properly and completely, otherwise what we'll end up with is another version of IE that's differently broken to the previous ones, so the IE-specific hacks that people end up using will be further complicated by differentiating between the two versions.
Of course, I've been telling people for years that using IE6-specific hacks is just committing yourself to lots more work at the next release. If you're sure you'll be around to do it then, go ahead... but I think it's far better for most people to just compromise their design so that it works in IE without the hacks. You still get the freedom to re-evaluate your position later, but you're not forced to.
Opera had an MDI interface which, with windows maximised, acts almost exactly like tabbed browsing. It has the advantage that you can display your "tabs" side-by-side if you want to, which isn't an option with Mozilla's tabs.
However, more recent versions of Opera allow the user to choose between full MDI, Mozilla-style tabs or one page per top-level window. I still use MDI, personally.
Where can I find an old-fashioned, heavy, clacky keyboard with a USB interface? This isn't a troll, it's a serious question. One of these days my last one of these things is going to break and I'm going to need to replace it, and right now I'm having a hard time finding anything but flimsy, cheap keyboards with no give on the keys and with the Ins/Del/Home/End/PgUp/PgDn keys in the wrong places.
I'd be cool if they would make it default to using 'handheld' media stylesheets rather than 'screen' as in the desktop version. That way sites can provide a simpler version of their design for small-screened devices. Of course, it'd be even better if someone would implement something like Opera's "Small Screen Rendering", which actually transforms the HTML a little to make sites which do not provide a special handheld stylesheet more usable.
As to the speed thing, the PocketPC systems I've seen have had reasonable processors, but doing intensive things like playing back highly-compressed video obviously uses up more battery than browsing your calendar.
Wouldn't having one access point feed off another be more like having an ethernet hub uplinking to another hub? My AP at least doesn't interfere with the network layer: the wireless hosts are on the same broadcast network as the wired ones.
Slashdot Mirror
I'm no economist, but shouldn't the market adjust to the constraints? Something should give, such as reduced pay or increased cost of the product.
I guess, though, that I'm not an economist for a reason.
I've often thought that it would be better if everyone worked half the hours they currently do and thus twice as many people were in the workforce. Of course, it isn't as simple as that as you'd still get people who want to work full time, and they'd have more money and thus drive up the prices of everything until everyone has to work more hours.
It's a nice thought, though. It annoys me that most of the people in the western world spend more time working than they do anything else.
I wasn't really suggesting that you have to take Computing to get into a CS degree... I just wanted to make the distinction between ICT and Computing. When it comes to university applications, the fussiest they seem to get is that some of them require a maths A-level. You'd probably have to go to a college to do A-Level Computing, as it's a bit specialist for a school to cover.
Back when I did all this stuff we only did three A-Levels, and I did Physics, Maths and Geography. I turned out okay. Good luck. ;)
Remember that what you're doing is ICT, which was never supposed to be about computers in and of themselves. It's about the application of computers and other technology to do lots of things involving the words "streamline" and "optimise".
If you want to do programming and other specialist stuff like that, it's not ICT that you're after: it's computing. Later it specialises even more and becomes "computer science". At GCSE level no-one is anywhere near specialised enough to do a course on programming. Sure, it's frustrating to go through ICT when you've got an interest in computing as an end unto itself, but just grin and bear it and it'll all be over soon. I'm just coming to the end of a masters degree and looking back my IT GCSE (as it was called then) was just a small part of the rich set of subjects I touched on at GCSE level.
Get your GCSEs out of the way and then enrol in an A-Level Computing course. There's still some ICT in there at A-Level, but it'll give you a little to get your teeth into at least, and from there (along with a few other A-Levels, of course) you can go on to do Computer Science at a university of your choice, and that is where the fun starts.
I wonder how long it'll be before people create alternative implementations of the client portion of this. I think it'd be pretty neat, for example, to have it as a network-wide proxy on my server here so that everyone in the house can share the benefits of the local caching. We're all geeks, so we all load things like slashdot all the time. Of course, a ghetto solution for now would be to run it under Wine and arrange for Squid to forward requests to it.
It would also be nice to see other implementations of the server part. Sure, it won't be as fast as running it at Google because they've got a crawler cache and a fat pipe, but it would still be cool to have a bunch of different servers to choose from. ISPs could run such a server for their customers to speed up the DSL or Cable bit of the transfer, for example.
I do like the idea of AR. The simplest entertainment-oriented implementation that springs to mind is similar to the "laser gun" games that were popular in the 90s -- Quasar and LaserQuest were the big franchises here in the UK. Take a large area and dec it up with something resembling a sci-fi set, then send in a bunch of people wearing AR goggles (which can be made to look like head armour) and have them work together to fight computer-generated aliens in the environment. Unlike Quasar and LaserQuest, there could actually be visible projectiles and discharges, and a variety of different weapons. Depending on how fancy the tech is, there could even be scorch marks added to the walls.
Of course, there are still other things not accounted for. If the aliens shoot at you, how do you feel the impact? What happens if a player walks through an alien? It's still interesting, though, and I for one would probably enjoy the game even if it was a little unrealistic as long as it is fair and doesn't cause motion sickness.
If you have Windows configured so that all of the Explorer windows share a single process, all of the windows necessarily run as the same user as well. You just have to enable "Open folder windows as a separate process" which is in the options somewhere. (I don't have a Windows XP machine handy to find it right now.)
My Nokia 3410 is serving me well. The battery lasts a week (and I've had it for a year or so now), it has a black and white screen and it does phone-like stuff. They still seem to be on sale, too -- in the UK, at least.
I think you really mean:
...or...
As you wrote it, the combination of potion_a and potion_c would have been valid, because ! potion_a would have returned false. Of course, there's also the fun perl construct:
(If you don't like perl's postfix if shorthand, feel free to flip that back round "the right way" and add the braces.) (The word "potion" is one of those strange words that just always looks like it's spelled wrong. Maybe my brain is trying to read it as "option"! :))
IE blends PNG images with an alpha channel against their default background color. When there is no default, it uses one of the system colors that happens to be the default for GDI surfaces.
You can give IE a better (but still solid) color to blend against by having pngcrush write in a bKGD chunk with the desired color. This doesn't help you if you are trying to blend against a texture, but it's handy if, as in your case, the image is blended against a solid background color anyway.
In most cases, playing around with the permissions on the install directory or installing it somewhere that the user has write access allows it to work. If it needs to write to the Windows directory to run, though, it's time to get a better application.
Also, have all of the inexperienced users (and ideally all of the experienced users too) set as limited user accounts. Even if someone needs to be able to install something, it's easy enough to run an installer as another user account.
Now that I've left home my family's computer doesn't get the regular maintenence attention that it once did, so I've set everyone as a limited user account and created an "Admin" account which is used for software installation and other such tasks. I popped home over Easter and took the opportunity to check up and make sure nothing had slipped through, and sure enough everything was fine. What I did find, though, was lots of malware and worm installers in my brother's Internet Explorer cache, which would have been installed and running if he had access to write to the Windows directory.
One of these days worm and spyware authors might wise up and install their software in the user's own file area and run from the current user startup list, but at least then a user on a shared machine is only screwing himself. If it comes to it, you can just back up the documents and nuke the user account.
It's been a while since I used Word, but I'm sure there was an option to disable the squiggly lines "as you type". You can still perform a spelling and grammar check when it is convenient to you by pressing F7 and/or Shift+F7, or by choosing the appropriate options from the menus. It'll then hop through the document flagging each possible error, giving you a list of suggestions and the opportunity to edit in a much more efficient interface than that used for as-you-type corrections.
Back in the days when Microsoft's best video player was the ActiveMovie player (hah!), OEMs bundled alternative players. My family's PC came with some weird-looking application which probably isn't around anymore (I forget the name) and I knew friends who got PCs bundled with Winamp and QuickTime Player.
Microsoft could easily have released Media Player as a separate application, and assuming it was good enough to compete with the other players around at the time OEMs would have bundled it in preference. See, OEMs know that user's want a system that "does stuff" out of the box, which is why they bundle all that value-add stuff like an office application, voicemodem software and in the past sound and video playback software. Microsoft didn't have to bundle it themselves; it could have competed fairly in the marketplace for a position in the default install of Windows, just like Microsoft Office did.
You're getting Media Player and DirectShow confused. DirectShow is the standard Windows API for video playback. Windows Media Player is an application which provides a UI for video playback based on the DirectShow API. The codecs live inside DirectShow. The Media Player ActiveX control confuses things a little, but is really a way to embed the Media Player user interface into your application or document, not just the bare video playback functionality.
If MS removes Media Player, they break the ActiveX control, which breaks wep pages and documents which embed that ActiveX control. However, if those documents instead just referenced the video data and left it up to the viewer application to determine an appropriate player, there would be no problem as users could install another player and have it take over the duties that Media Player would have performed by default. Embedding Media Player directly is generally a bad idea anyway, since clearly that will only work on Windows systems. (and not even all Windows systems, now.)
In addition, applications which depend on DirectShow for video playback should go on working. This includes Winamp, mplayer (in part), DirectX Player as well as a load of games and other applications that aren't directly video-related but still playback video for some reason. If Microsoft has removed DirectShow, then it is really trying to be facetious, as clearly when the courts ruled that Media Player be unbundled they were referring to the Media Player application and not the API on which it depends. By that argument, they should have removed the Win32 API as well.
One way to embed video clips into Office documents is to embed the Media Player ActiveX control. Of course, without Media Player there is no Media Player ActiveX control and so documents using this technique won't load correctly without Media Player.
I've not used Word in years, but I'm going to assume that there's also a second way which involves embedding the video just as video data, without any particular container. Now, I'd expect those to play back through DirectShow (the API Media Player uses to play video) not Media Player itself, and so it should go on working just like any game which uses DirectShow for movie playback should go on working, and Winamp (which plays back most filetypes through DirectShow) should go on working.
Therefore there should be no dependence on any particular player frontend, but you'll still only be able to play movies which have a registered DirectShow codec. Since DirectShow is the standard Windows API for video playback, this is sensible. That a bunch of video format owners (Real, Quicktime) don't distribute DirectShow codecs is their fault, not Microsoft's. Of course, if Office applications really don't have a way to embed video directly without using a specific player ActiveX control then I would describe the developers as incompetant rather than claim sabotage.
I suppose a final possibility is that Microsoft heard "Remove Media Player!" and went and stripped out DirectShow. That would be malicious in my mind since DirectShow is the standard API for video playback in Windows and so its removal would break loads of applications. It was the Media Player application that was to be removed, not the APIs it uses. By that logic, the Win32 API should be removed as well as it's clearly part of Media Player!
What I had in mind was something like the browser indicating it has support for XUL in the Accept header, or perhaps just some way to deliver both in the same response and have Mozilla ignore the HTML and others ignore the XUL. One more possibility, which does have the overhead of an additional request, is to include a LINK element in the HTML document that indicates that there is an alternative XUL version of the page. Mozilla would then spot this and use the XUL document in preference.
All of these require browser changes, but since Mozilla is currently the only lot with XUL support they get to set some standards in this respect. I think it'd be worth it since it would encourage the uptake of XUL and thus attract more users. Hopefully in the long term we'll see other browsers implementing XUL support and we can finally be rid of the ugly hacks people do to get an HTML page to act like an application.
XUL could almost certainly do what you're describing. However, no-one will actually bother to do it firstly because no-one wants to use a word-processor embedded in a web browser and secondly Mozilla's marketshare isn't big enough for anyone to consider such a product.
A slightly more realistic goal would be to get websites implementing a superior XUL interface in addition to the HTML one. Photo gallery sites could implement a proper GUI for organising arranging pictures, for example.
This would be made much easier if there was a way to transparently provide Mozilla users with a XUL interface while everyone else gets a HTML interface. Until that can be done without any input from the user (who won't necessarily even understand what XUL is) it won't get done. If someone can find a way to do that, I'll certainly push for getting it added to software and sites that I have some influence over.
Implementing a new template doesn't require touching any Perl code. The submitter simply needs to provide a new set of Template Toolkit templates.
The problem stems from the fact that slashdot's pages are generated from tons of little bits of template, and so taking someone's hack of a static front page doesn't really get you very far to having a working implementation.
Intrinsic types in QBasic:
INTEGER A 16-bit signed integer variable. LONG A 32-bit signed integer variable. SINGLE A single-precision 32-bit floating-point variable. DOUBLE A double-precision 64-bit floating-point variable. STRING * n% A fixed-length string variable n% bytes long. STRING A variable-length string variable.Sounds to me like you were using a DOUBLE where you should have been using a LONG. There's no such thing as a "double-precision integer", because integers don't have precision, they just have a maximum and minimum value.
(Those descriptions came straight out of QBasic's online help, which for some reason I still have lying around.)
Getting their CSS support up to par isn't an easy job. Lots of it is broken because it is just mapping CSS properties onto parameters for old rendering code, which is why there are lots of little quirks in the box model support. Lots of it is unsupported and would require massive changes to the rendering engine to implement.
I'd much rather that MS save CSS fixes for a time when they are able to do it properly and completely, otherwise what we'll end up with is another version of IE that's differently broken to the previous ones, so the IE-specific hacks that people end up using will be further complicated by differentiating between the two versions.
Of course, I've been telling people for years that using IE6-specific hacks is just committing yourself to lots more work at the next release. If you're sure you'll be around to do it then, go ahead... but I think it's far better for most people to just compromise their design so that it works in IE without the hacks. You still get the freedom to re-evaluate your position later, but you're not forced to.
Opera had an MDI interface which, with windows maximised, acts almost exactly like tabbed browsing. It has the advantage that you can display your "tabs" side-by-side if you want to, which isn't an option with Mozilla's tabs.
However, more recent versions of Opera allow the user to choose between full MDI, Mozilla-style tabs or one page per top-level window. I still use MDI, personally.
Where can I find an old-fashioned, heavy, clacky keyboard with a USB interface? This isn't a troll, it's a serious question. One of these days my last one of these things is going to break and I'm going to need to replace it, and right now I'm having a hard time finding anything but flimsy, cheap keyboards with no give on the keys and with the Ins/Del/Home/End/PgUp/PgDn keys in the wrong places.
I'd be cool if they would make it default to using 'handheld' media stylesheets rather than 'screen' as in the desktop version. That way sites can provide a simpler version of their design for small-screened devices. Of course, it'd be even better if someone would implement something like Opera's "Small Screen Rendering", which actually transforms the HTML a little to make sites which do not provide a special handheld stylesheet more usable.
As to the speed thing, the PocketPC systems I've seen have had reasonable processors, but doing intensive things like playing back highly-compressed video obviously uses up more battery than browsing your calendar.
Wouldn't having one access point feed off another be more like having an ethernet hub uplinking to another hub? My AP at least doesn't interfere with the network layer: the wireless hosts are on the same broadcast network as the wired ones.