This won't catch on, because people generally use IM while doing something else. I can type a message and then read some email (or type a message to someone else) while waiting for a response. When the other person does respond, the window icon blinks or jumps around or whatever is usual in your chosen environment to get your attention.
These 3D environments (I've tried a few of them) generally require more attention, since firstly there are generally lots more people involved in conversation, and your "relationship" with other people in the space is a lot more hazey, based on where everyone is in the environment rather than just who they have chosen to talk to today. Also, many of them don't have an easily-usable scrollback buffer, so you can get lost in the conversation if you look away for a few minutes.
IM will always be text-based just because it's more efficient that way. Systems like Second Life and There have their place, but it's not as a replacement for IM. (Side note: There actually has IM integrated into it, and surprise, surprise... it's text-based!)
Slashcode uses Template Toolkit, which is quite a flexible (and separate) templating system written in Perl. Anyone who knows template toolkit could theoretically create a new slashdot theme and submit it. The garish section themes are created by swapping out a few of the template elements and introducing others in their place.
I don't know about how jumbled the perl and SQL are, but the template stuff's pretty divorced from the rest. See the Slashcode site, for example.
Given your web design philosophy, I must assume that you create your own LiveJournal styles, since the stock ones use some really bad HTML display hacks. LiveJournal's layouts seem to be focused on ig'nant kids rather than people who actually care about CSS and semantic markup.
XSLT is for tranforming XML data between different XML formats or sometimes from XML to non-XML formats. It doesn't have much to do with "structuring layout". (I routinely use XSLT to transform Simplified Docbook into HTML, LaTeX and XSL-FO)
XSL-FO, on the other hand, is an XML application for describing (loosely) typesetting parameters. It's actually almost parallel with CSS in purpose, but CSS is more rich in functions relating to on-screen interactive content, like support for links and behaviors. XSL-FO could be used, for example, as an internal data structure resulting from applying CSS to some XHTML, although of course in practice browsers just use their own stuff. The relationship between XSLT and XSL-FO is that originally they were one lump (called "XSL") which was used to translate XML documents into FO documents for rendering, but W3C noticed that XSLT has more uses outside of that and split it into two separate specs.
Incidentally, passivetex is an XSL-FO interpreter for TeX. If you're happy specifying typesetting parameters at the lowest level it can be quite useful, but I prefer to just go straight to LaTeX since I trust it to "do the right thing" with regard to page layout and presentation most of the time.
Now make a list of all of the sections which have color schemes you don't like and add them to your/etc/hosts file (or c:\windows\system32\drivers\etc\hosts if you're that way inclined) with the IP address of the Apache instance you just configured. It'll look something like this:
Those hostnames will now go to your Apache rather than to the real slashdot, and the virtualhost will bounce you off to plain ol' slashdot.org. If you've got several machines to do this to, you might like to consider setting up a nameserver and having it be the master for slashdot.org, but bear in mind that you'll have to change it if slashdot's IP changes. Also, it can make it a bit of a bitch to read the section front pages, but I don't generally do that anyway.
Alternatively, a much easier solution is just to configure your account to use the "Light" template rather than the standard template. If your browser's default rendering doesn't appeal to you, take a few moments to write a user stylesheet which you do like to look at and other sites will benefit from it too. It's best if you use a browser like Opera which allows you to switch between "what most browsers have as default" and "my user stylesheet" so that badly-written sites don't get tripped up by your unusual choices and cause white-on-white text.
I actually do both of these things, because even in the light template some of the beige continues to show through, for example on the comment form.
TOPMARGIN is a proprietary Internet Explorer attribute, not part of the standard. The same applies to LEFTMARGIN.
I seem to remember that HTML 3.2 was roughly based on Netscape's implementation of bastardized HTML, so I'd guess that the "correct" way to do this would be with MARGINHEIGHT and MARGINWIDTH, which IE does not support. HTML 3.2 isn't really a standard in anything but name.
Opera at one time had a feature to automatically reflow the page once it had loaded because it had some of these render-while-loading bugs. I've not used recent versions, so I don't know if it's still there...
My point, though, is that this can be quite annoying. Depending on the severity of the change when it's re-flowed, text and links can end up being quite a large distance away from where they appeared the first time, so just as I'm about to click in a text box it jumps up 10 pixels and I miss and hit a submit button or something. Very annoying. Instead, I tend to just make use of Opera's manual "Refresh Display" feature. Sure, it doesn't hide the problem under a hack, but at least if the page loads and is readable despite not being "quite right" I can get on and use it without the layout suddenly changing on me a few seconds later.
More likely it tags you with a unique ID and then maps that unique ID on to your searches on the server. I imagine after a while they start dropping entries out of the logs and killing entries for people they've not seen for a while to avoid storing redundant data for people who switched web browsers, got a new PC or whatever.
Some function calls expect to be called twice, or have a helper function - once to get the exact buffer length, then again (once you have created that buffer) to fill it.
Of course, even that must be used with care. In some cases this is a safe approach -- usually only when the result is only dependent on some other parameter passed to the function. However, if the function gets data from elsewhere (getting the caption of a button, for example) you have to be careful to make sure nothing else changes the data between you making the count call and making the real call.
This can be mitigated to a certain extent by having the helper function and specifying the buffer length in the real call, having the library warn you somehow if the data wasn't the length you expected.
Passing in a buffer was the norm in C, but since C++ now has (theoretically) standardized allocation and freeing of memory when using a purely C++ library it might well be safe just to return a buffer and have the caller delete it later. Of course, once you commit to doing this you commit to only using new to create objects/buffers and tie yourself to a particular compiler/runtime-library combo if you are shipping binaries, which is normally more pain than it's worth.
Jabber's model is excellent. It's very similar to the email distribution model where there is a network of servers and then each server has a bunch of clients hanging off it. Although email has a few security issues, the general model is sound: it's decentralised, and yet it still makes efficient use of the network unlike the current "peer-to-peer" apps which generate an extra network over the network usually with little regard to the underlying topology and proceed to shove redundant data everywhere.
The difference between Jabber and the email system, though, is that the email system happened during the early days of the Internet. Consequently, when it got enough momentum for commercial ISPs to appear providing access to homes it was a given that every ISP "had to" provide customers with an email account and run email servers to handle the customer's mail.
If Jabber had come along at the same time, it too would be one of those things ISPs just provide with every account. Some ISPs would even let customers have multiple Jabber IDs per account like they used to do with email addresses to differentiate themselves from the competition.
The problem is that, with a few notable exceptions, the only companies that actually get paid for Internet services are ISPs. There is no viable business model to provide free Jabber services just as there is no money in free email with POP3 and IMAP access. The availability of Jabber is limited to geeks who set up servers of varying scales and people who know said geeks. The only way a commercial Jabber service could be provided is to use a proprietary protocol for client-to-server communication and then bridge that onto the Jabber network at the server side. This is comparable to web-based email which uses proprietary HTML to encode the data, meaning you get the service on their terms and not on yours. (usually one of these terms is the presence of advertising)
On the other hand, if every viable comsumer ISP provided customers with a Jabber server and at least one Jabber ID it would have flourished. These days, though, there is no way to add new "required services" to ISPs and they are frantically trying to shed the few they have now. (Notice the decline in ISPs providing USENET servers and customer webspace)
It sucks, but there is no way at this point in the game to introduce a new distributed and open protocol. There's just no money in distributed and open protocols.
My brothers have a The Simpsons game for some console or other. It is similar to GTA3, but more linear and more wacky. It's been a while since I've seen it, and they put a lot less effort into the layout than these guys did, but it was okay.
The other thing which makes it less certain is that they are not regulating the airspace, they are instead regulating whether a specific class of device can be used. They can't stop people standing outside their land and providing wi-fi from there, but I would think that they can stop people errecting a physical transmitter on their property.
As an example of what I mean, do you think it would be legal for me to place a transmitter and mast on top of a shopping mall without asking the owners of the mall? I would certainly not dispute that they can't stop me transmitting, but I would expect that they could legally stop me adding that mast in the first place.
A wireless access point is just like a small mast. The university is prohibiting the device, not the transmissions. Of course, IANAL and I'm now just talking about how I would expect the law to be, not how it actually is. If the law doesn't agree with me here then fair enough, but I think it's unreasonable that any property owner should be legally powerless to stop an item being brought onto their property just because it happens to emit radio waves on a particular frequency.
I did that, too. For the Commodore 64's processor, in fact! It was quite fun, although the result wasn't particularly brilliant. Writing a compiler is certainly an interesting exercise.
I don't really remember my learning experience on my Commodore 64 and my Amiga because I was too young, but I remember as a teen encountering my first "IBM Compatible PC" (a Compaq 286 system). I spent the next week poking around in DOS learning to do things, with only my previous experience of AmigaDOS to guide me. dir lead to a directory of other things and eventually I encountered the fullscreen apps like MS-DOS Help, MS-DOS Editor and eventually QBasic. Finding QBasic started it all off again!
I had a similar experience the first time I used a linux box. That box is actually still sitting on top of the cupboard to the side of me as I type this, still with an old version of Red Hat Linux installed on it. It was given to me by a friend already installed. This was an old system and X wasn't installed, so it was command line all the way.
Poking around command line systems is so much more fun than poking around GUIs. A friend of mine bought an iBook and I had got bored of tinkering with OS X within a few minutes. (the command line didn't offer much since these days UNIX-like systems are old hat.) I don't really remember learning Windows, which indicates to me that it must have been a pretty boring experience.
Applications can find out that it has been pressed (although they have to go to a bit of effort to do so) but they can't prevent Windows from recieving the message. Windows then switches to a completely different desktop (or "Window Station" as it calls them) to display the options window, so there is no way that running applications can display anything onscreen until the user dismisses the Windows Security window and is switched back to the primary window station.
Of course, in order for this to be effective users must know this is the case, and in most cases they don't want to know about window stations and interrupts and things like that so it's enough simply to give a login screen which doesn't require Ctrl+Alt+Del; how many users are going to remember that they normally have to press this key combo and notice that today they didn't? Even some of those who do notice will assume that someone already pressed Ctrl+Alt+Del but walked off without clicking cancel after failing to log in.
IMG was added for Mosaic in one of the first short-sighted decisions which led to the mess that was later called HTML 3.2. When submitted for peer-review (your so-called "proper channels") people suggested that maybe it would be better to make a more general embedding element and that it should be a container so that backward-compatible content could be included inside. The ALT, element, if you think about it, is useless for this because no previous browser knew to look there! Also, it can only support plain text, so more "rich" alternatives such as a link to another document or even just a few paragraphs of text are not an option.
Mosaic developer said "tough luck, I've already written it" and released it as it was. Since Mosaic was the biggest browser at the time people quickly sucked down this new IMG element and everyone was forced to co-operate.
To me, this sounds like exactly what Microsoft is doing. IMG was a quick hack with little forward planning just to get images into web pages as quickly as possible. Mosaic and then later Netscape continued this trend by adding poorly-considered features such as FONT, APPLET, EMBED and framesets which are now, thankfully, deprecated in current W3C specifications in favor of better thought-out and more general features. Sadly, W3C's versions have buggy support in most current browsers. (They haven't replaced frames, but I suspect that's because it was a pretty bad idea in the first place, so the best thing to do is just to avoid it.)
To sum up: in GTk+, a "layout manager" is just a widget with other widgets contained in it, while in Java a layout manager is a special construct separate from widgets which is assigned to any widget which is a generic container.
In practice, I generally prefer GTk's approach, but I see merits in both approaches.
Dijkstra was almost certainly referring to "traditional" basic. I suppose you could use Microsoft's BASIC for the Commodore 64 as an example. There are a few control structures (IF and FOR) but essentially it's just high-level assembler: subroutines are call/return pairs (GOSUB/RETURN) and jump instructions (GOTO) are the primary means for creating loops.
Today's Visual Basic has seen several stages of evolution since then, and now has structured functions and subroutines, proper looping constructs and a vaugue attempt at encapsulation. Visual Basic.NET is just as structured as C# or Java.
As someone who started with BASIC and is now happy with everything from Pascal to Perl to C, I thank BASIC (CBM BASIC in particular) for showing me in the early days things which helped me understand how the underlying machine worked, which can't be said for more modern languages which abstract those details away by design. Abstractions are good if you understand what's being abstracted and why, but using an abstraction to avoid learning something is, in my opinion, a mistake, since every abstraction will leak from time to time and you won't be able to work out what happened unless you are able to reason about what's going on inside your pretty black box.
There's more than a standard lease between a student and a university. They also sign some kind of agreement to comply with university rules which has nothing to do with tenency laws. It is, of course, still covered by contract law.
Note that I'm not claiming that the university is okay to do this, I'm simply saying that it isn't quite as open-and-shut as it would be in the case of a simple tenency agreement. It could go either way in court, I think.
The problem here is that there's more than the lease affecting this. Students also sign a separate contract with the university agreeing to comply with university regulations. The sanctions for this aren't eviction but things like witholding degrees and whatnot. Of course, the lease probably states that if you cease to be a student you automatically give up your lease, so the university could theoretically expell students and on those grounds terminate the lease.
They could try, anyway. I'm sure if they did it the affected student take it to court and then this discussion would be moot. More likely lesser santions such as fines or grade reductions would be used instead, allowing the student to continue the lease and thus not affecting that particular contract at all.
The difference here is that they are not laying down the law. You won't go to jail for using a wireless access point in your dorm room, but the university has a lot of other sanctions it can apply to a student which are not law-related.
Students, when registering with a university, agree to a set of rules. A bunch of these rules wil detail things the university is allowed to do if you break the other rules. An example might be withholding your degree or expelling you from the college completely. Current students of the university have agreed to do what the uni says or face whatever sanctions were outlined in this university's regulations. This isn't an FCC issue because we aren't talking about the law we are talking about rules.
Just to make clear the distinction here: smashing up a lecture theater is against the law, but cheating on a test is against the rules. You can go to jail for the former, but the uni deals with the latter in accordance with their regulations.
These days I think they're getting pretty close to modelling skin, from things like how it reflects light to including the small imperfections like scars, moles and asymmetry without going overboard.
What still bothers me about computer-generated people, though, is the animation. They generally get simple body motions like walking and running right, but they haven't quite mastered realistic "body language" motions from leaning back against walls to indicate relaxation to things as apparently-simple as the face expressing fear or happiness. Computer-generated characters in situations where they are interacting socially always look awkward and overdone. They are usually fine in purely-practical situations like fight scenes and the like.
The first Animatrix short, as mentioned by another poster, demonstrates this quite well. I recommend checking it out.
The other thing that isn't quite "there" yet is the voice acting, although I think in many cases that's due to bad voice actors more than anything else.
Why does the application have to look like a hi-fi? It's thinking like that which brings us twisty knob widgets to operate with the mouse, and millions of users who try to carefully articulate the mouse in small circles to operate it, only to be frustrated and confused when the bottom part of the motion reverses what the top part did.
Metaphors can be nice, but just having buttons with the familiar play, stop and pause buttons would be enough for this kind of interface.
Your last paragraph reminds me of something one of my friends told me about a while back. In one of these new-fangled "virtual world" things -- I think it was There -- they encourage players to purchase different textures/models for their avatars, of course calling them "clothes". At the time I was told, they had recently added a special new "shoe" product which was named after a major brand of training shoes. Which one I forget.
The point is, though, that like you said players "wearing" these trainers can run slightly faster than other players. The thing that amused me most was that apparently players were buying them like crazy even though running speed has no real bearing on this particular game. If they'll do it for this, then it stands to reason that players will be even more willing to go for this in a game where running faster actually provides some kind of advantage over other players.
The other day I visited my local cinema with some friends to watch I, Robot. We went to the pub for a drink first and after queueing and my friends buying expensive food products we got in to the theatre proper about 35 minutes after the billed start time, expecting to have missed the start. We were quite shocked (and, on that particular occasion, relieved) to find that we arrived in time to see the last preview trailer as well as the "don't let mobile phones ruin your movie" and the "Love Movies? Hate Piracy!" (Aarr!) stuff.
That's 40 minutes of junk before the film starts! Normally I don't notice the length of time because I'm chatting to my friends during this time. The ticket cost five pounds (roughly 9 US Dollars) and then once the film finally started the main character kept going on about his basketball shoes, which were of a brand I've never heard of and thus promptly forgot. The blatant nature of that product-placement was actually quite amusing, which was probably what they were going for.
Slashdot Mirror
This won't catch on, because people generally use IM while doing something else. I can type a message and then read some email (or type a message to someone else) while waiting for a response. When the other person does respond, the window icon blinks or jumps around or whatever is usual in your chosen environment to get your attention.
These 3D environments (I've tried a few of them) generally require more attention, since firstly there are generally lots more people involved in conversation, and your "relationship" with other people in the space is a lot more hazey, based on where everyone is in the environment rather than just who they have chosen to talk to today. Also, many of them don't have an easily-usable scrollback buffer, so you can get lost in the conversation if you look away for a few minutes.
IM will always be text-based just because it's more efficient that way. Systems like Second Life and There have their place, but it's not as a replacement for IM. (Side note: There actually has IM integrated into it, and surprise, surprise... it's text-based!)
Slashcode uses Template Toolkit, which is quite a flexible (and separate) templating system written in Perl. Anyone who knows template toolkit could theoretically create a new slashdot theme and submit it. The garish section themes are created by swapping out a few of the template elements and introducing others in their place.
I don't know about how jumbled the perl and SQL are, but the template stuff's pretty divorced from the rest. See the Slashcode site, for example.
Given your web design philosophy, I must assume that you create your own LiveJournal styles, since the stock ones use some really bad HTML display hacks. LiveJournal's layouts seem to be focused on ig'nant kids rather than people who actually care about CSS and semantic markup.
XSLT is for tranforming XML data between different XML formats or sometimes from XML to non-XML formats. It doesn't have much to do with "structuring layout". (I routinely use XSLT to transform Simplified Docbook into HTML, LaTeX and XSL-FO)
XSL-FO, on the other hand, is an XML application for describing (loosely) typesetting parameters. It's actually almost parallel with CSS in purpose, but CSS is more rich in functions relating to on-screen interactive content, like support for links and behaviors. XSL-FO could be used, for example, as an internal data structure resulting from applying CSS to some XHTML, although of course in practice browsers just use their own stuff. The relationship between XSLT and XSL-FO is that originally they were one lump (called "XSL") which was used to translate XML documents into FO documents for rendering, but W3C noticed that XSLT has more uses outside of that and split it into two separate specs.
Incidentally, passivetex is an XSL-FO interpreter for TeX. If you're happy specifying typesetting parameters at the lowest level it can be quite useful, but I prefer to just go straight to LaTeX since I trust it to "do the right thing" with regard to page layout and presentation most of the time.
I do this with a filtering proxy, but here's a version for those who don't use such a thing:
Run an Apache instance somewhere on your network. Add to it a VirtualHost with an alias configured as *.slashdot.org, as follows:
Now make a list of all of the sections which have color schemes you don't like and add them to your /etc/hosts file (or c:\windows\system32\drivers\etc\hosts if you're that way inclined) with the IP address of the Apache instance you just configured. It'll look something like this:
Those hostnames will now go to your Apache rather than to the real slashdot, and the virtualhost will bounce you off to plain ol' slashdot.org. If you've got several machines to do this to, you might like to consider setting up a nameserver and having it be the master for slashdot.org, but bear in mind that you'll have to change it if slashdot's IP changes. Also, it can make it a bit of a bitch to read the section front pages, but I don't generally do that anyway.
Alternatively, a much easier solution is just to configure your account to use the "Light" template rather than the standard template. If your browser's default rendering doesn't appeal to you, take a few moments to write a user stylesheet which you do like to look at and other sites will benefit from it too. It's best if you use a browser like Opera which allows you to switch between "what most browsers have as default" and "my user stylesheet" so that badly-written sites don't get tripped up by your unusual choices and cause white-on-white text.
I actually do both of these things, because even in the light template some of the beige continues to show through, for example on the comment form.
TOPMARGIN is a proprietary Internet Explorer attribute, not part of the standard. The same applies to LEFTMARGIN.
I seem to remember that HTML 3.2 was roughly based on Netscape's implementation of bastardized HTML, so I'd guess that the "correct" way to do this would be with MARGINHEIGHT and MARGINWIDTH, which IE does not support. HTML 3.2 isn't really a standard in anything but name.
Opera at one time had a feature to automatically reflow the page once it had loaded because it had some of these render-while-loading bugs. I've not used recent versions, so I don't know if it's still there...
My point, though, is that this can be quite annoying. Depending on the severity of the change when it's re-flowed, text and links can end up being quite a large distance away from where they appeared the first time, so just as I'm about to click in a text box it jumps up 10 pixels and I miss and hit a submit button or something. Very annoying. Instead, I tend to just make use of Opera's manual "Refresh Display" feature. Sure, it doesn't hide the problem under a hack, but at least if the page loads and is readable despite not being "quite right" I can get on and use it without the layout suddenly changing on me a few seconds later.
More likely it tags you with a unique ID and then maps that unique ID on to your searches on the server. I imagine after a while they start dropping entries out of the logs and killing entries for people they've not seen for a while to avoid storing redundant data for people who switched web browsers, got a new PC or whatever.
Of course, even that must be used with care. In some cases this is a safe approach -- usually only when the result is only dependent on some other parameter passed to the function. However, if the function gets data from elsewhere (getting the caption of a button, for example) you have to be careful to make sure nothing else changes the data between you making the count call and making the real call.
This can be mitigated to a certain extent by having the helper function and specifying the buffer length in the real call, having the library warn you somehow if the data wasn't the length you expected.
Passing in a buffer was the norm in C, but since C++ now has (theoretically) standardized allocation and freeing of memory when using a purely C++ library it might well be safe just to return a buffer and have the caller delete it later. Of course, once you commit to doing this you commit to only using new to create objects/buffers and tie yourself to a particular compiler/runtime-library combo if you are shipping binaries, which is normally more pain than it's worth.
Jabber's model is excellent. It's very similar to the email distribution model where there is a network of servers and then each server has a bunch of clients hanging off it. Although email has a few security issues, the general model is sound: it's decentralised, and yet it still makes efficient use of the network unlike the current "peer-to-peer" apps which generate an extra network over the network usually with little regard to the underlying topology and proceed to shove redundant data everywhere.
The difference between Jabber and the email system, though, is that the email system happened during the early days of the Internet. Consequently, when it got enough momentum for commercial ISPs to appear providing access to homes it was a given that every ISP "had to" provide customers with an email account and run email servers to handle the customer's mail.
If Jabber had come along at the same time, it too would be one of those things ISPs just provide with every account. Some ISPs would even let customers have multiple Jabber IDs per account like they used to do with email addresses to differentiate themselves from the competition.
The problem is that, with a few notable exceptions, the only companies that actually get paid for Internet services are ISPs. There is no viable business model to provide free Jabber services just as there is no money in free email with POP3 and IMAP access. The availability of Jabber is limited to geeks who set up servers of varying scales and people who know said geeks. The only way a commercial Jabber service could be provided is to use a proprietary protocol for client-to-server communication and then bridge that onto the Jabber network at the server side. This is comparable to web-based email which uses proprietary HTML to encode the data, meaning you get the service on their terms and not on yours. (usually one of these terms is the presence of advertising)
On the other hand, if every viable comsumer ISP provided customers with a Jabber server and at least one Jabber ID it would have flourished. These days, though, there is no way to add new "required services" to ISPs and they are frantically trying to shed the few they have now. (Notice the decline in ISPs providing USENET servers and customer webspace)
It sucks, but there is no way at this point in the game to introduce a new distributed and open protocol. There's just no money in distributed and open protocols.
My brothers have a The Simpsons game for some console or other. It is similar to GTA3, but more linear and more wacky. It's been a while since I've seen it, and they put a lot less effort into the layout than these guys did, but it was okay.
The other thing which makes it less certain is that they are not regulating the airspace, they are instead regulating whether a specific class of device can be used. They can't stop people standing outside their land and providing wi-fi from there, but I would think that they can stop people errecting a physical transmitter on their property.
As an example of what I mean, do you think it would be legal for me to place a transmitter and mast on top of a shopping mall without asking the owners of the mall? I would certainly not dispute that they can't stop me transmitting, but I would expect that they could legally stop me adding that mast in the first place.
A wireless access point is just like a small mast. The university is prohibiting the device, not the transmissions. Of course, IANAL and I'm now just talking about how I would expect the law to be, not how it actually is. If the law doesn't agree with me here then fair enough, but I think it's unreasonable that any property owner should be legally powerless to stop an item being brought onto their property just because it happens to emit radio waves on a particular frequency.
I did that, too. For the Commodore 64's processor, in fact! It was quite fun, although the result wasn't particularly brilliant. Writing a compiler is certainly an interesting exercise.
I don't really remember my learning experience on my Commodore 64 and my Amiga because I was too young, but I remember as a teen encountering my first "IBM Compatible PC" (a Compaq 286 system). I spent the next week poking around in DOS learning to do things, with only my previous experience of AmigaDOS to guide me. dir lead to a directory of other things and eventually I encountered the fullscreen apps like MS-DOS Help, MS-DOS Editor and eventually QBasic. Finding QBasic started it all off again!
I had a similar experience the first time I used a linux box. That box is actually still sitting on top of the cupboard to the side of me as I type this, still with an old version of Red Hat Linux installed on it. It was given to me by a friend already installed. This was an old system and X wasn't installed, so it was command line all the way.
Poking around command line systems is so much more fun than poking around GUIs. A friend of mine bought an iBook and I had got bored of tinkering with OS X within a few minutes. (the command line didn't offer much since these days UNIX-like systems are old hat.) I don't really remember learning Windows, which indicates to me that it must have been a pretty boring experience.
Applications can find out that it has been pressed (although they have to go to a bit of effort to do so) but they can't prevent Windows from recieving the message. Windows then switches to a completely different desktop (or "Window Station" as it calls them) to display the options window, so there is no way that running applications can display anything onscreen until the user dismisses the Windows Security window and is switched back to the primary window station.
Of course, in order for this to be effective users must know this is the case, and in most cases they don't want to know about window stations and interrupts and things like that so it's enough simply to give a login screen which doesn't require Ctrl+Alt+Del; how many users are going to remember that they normally have to press this key combo and notice that today they didn't? Even some of those who do notice will assume that someone already pressed Ctrl+Alt+Del but walked off without clicking cancel after failing to log in.
IMG was added for Mosaic in one of the first short-sighted decisions which led to the mess that was later called HTML 3.2. When submitted for peer-review (your so-called "proper channels") people suggested that maybe it would be better to make a more general embedding element and that it should be a container so that backward-compatible content could be included inside. The ALT, element, if you think about it, is useless for this because no previous browser knew to look there! Also, it can only support plain text, so more "rich" alternatives such as a link to another document or even just a few paragraphs of text are not an option.
Mosaic developer said "tough luck, I've already written it" and released it as it was. Since Mosaic was the biggest browser at the time people quickly sucked down this new IMG element and everyone was forced to co-operate.
To me, this sounds like exactly what Microsoft is doing. IMG was a quick hack with little forward planning just to get images into web pages as quickly as possible. Mosaic and then later Netscape continued this trend by adding poorly-considered features such as FONT, APPLET, EMBED and framesets which are now, thankfully, deprecated in current W3C specifications in favor of better thought-out and more general features. Sadly, W3C's versions have buggy support in most current browsers. (They haven't replaced frames, but I suspect that's because it was a pretty bad idea in the first place, so the best thing to do is just to avoid it.)
To sum up: in GTk+, a "layout manager" is just a widget with other widgets contained in it, while in Java a layout manager is a special construct separate from widgets which is assigned to any widget which is a generic container.
In practice, I generally prefer GTk's approach, but I see merits in both approaches.
Dijkstra was almost certainly referring to "traditional" basic. I suppose you could use Microsoft's BASIC for the Commodore 64 as an example. There are a few control structures (IF and FOR) but essentially it's just high-level assembler: subroutines are call/return pairs (GOSUB/RETURN) and jump instructions (GOTO) are the primary means for creating loops.
Today's Visual Basic has seen several stages of evolution since then, and now has structured functions and subroutines, proper looping constructs and a vaugue attempt at encapsulation. Visual Basic .NET is just as structured as C# or Java.
As someone who started with BASIC and is now happy with everything from Pascal to Perl to C, I thank BASIC (CBM BASIC in particular) for showing me in the early days things which helped me understand how the underlying machine worked, which can't be said for more modern languages which abstract those details away by design. Abstractions are good if you understand what's being abstracted and why, but using an abstraction to avoid learning something is, in my opinion, a mistake, since every abstraction will leak from time to time and you won't be able to work out what happened unless you are able to reason about what's going on inside your pretty black box.
There's more than a standard lease between a student and a university. They also sign some kind of agreement to comply with university rules which has nothing to do with tenency laws. It is, of course, still covered by contract law.
Note that I'm not claiming that the university is okay to do this, I'm simply saying that it isn't quite as open-and-shut as it would be in the case of a simple tenency agreement. It could go either way in court, I think.
The problem here is that there's more than the lease affecting this. Students also sign a separate contract with the university agreeing to comply with university regulations. The sanctions for this aren't eviction but things like witholding degrees and whatnot. Of course, the lease probably states that if you cease to be a student you automatically give up your lease, so the university could theoretically expell students and on those grounds terminate the lease.
They could try, anyway. I'm sure if they did it the affected student take it to court and then this discussion would be moot. More likely lesser santions such as fines or grade reductions would be used instead, allowing the student to continue the lease and thus not affecting that particular contract at all.
The difference here is that they are not laying down the law. You won't go to jail for using a wireless access point in your dorm room, but the university has a lot of other sanctions it can apply to a student which are not law-related.
Students, when registering with a university, agree to a set of rules. A bunch of these rules wil detail things the university is allowed to do if you break the other rules. An example might be withholding your degree or expelling you from the college completely. Current students of the university have agreed to do what the uni says or face whatever sanctions were outlined in this university's regulations. This isn't an FCC issue because we aren't talking about the law we are talking about rules.
Just to make clear the distinction here: smashing up a lecture theater is against the law, but cheating on a test is against the rules. You can go to jail for the former, but the uni deals with the latter in accordance with their regulations.
These days I think they're getting pretty close to modelling skin, from things like how it reflects light to including the small imperfections like scars, moles and asymmetry without going overboard.
What still bothers me about computer-generated people, though, is the animation. They generally get simple body motions like walking and running right, but they haven't quite mastered realistic "body language" motions from leaning back against walls to indicate relaxation to things as apparently-simple as the face expressing fear or happiness. Computer-generated characters in situations where they are interacting socially always look awkward and overdone. They are usually fine in purely-practical situations like fight scenes and the like.
The first Animatrix short, as mentioned by another poster, demonstrates this quite well. I recommend checking it out.
The other thing that isn't quite "there" yet is the voice acting, although I think in many cases that's due to bad voice actors more than anything else.
Why does the application have to look like a hi-fi? It's thinking like that which brings us twisty knob widgets to operate with the mouse, and millions of users who try to carefully articulate the mouse in small circles to operate it, only to be frustrated and confused when the bottom part of the motion reverses what the top part did.
Metaphors can be nice, but just having buttons with the familiar play, stop and pause buttons would be enough for this kind of interface.
Your last paragraph reminds me of something one of my friends told me about a while back. In one of these new-fangled "virtual world" things -- I think it was There -- they encourage players to purchase different textures/models for their avatars, of course calling them "clothes". At the time I was told, they had recently added a special new "shoe" product which was named after a major brand of training shoes. Which one I forget.
The point is, though, that like you said players "wearing" these trainers can run slightly faster than other players. The thing that amused me most was that apparently players were buying them like crazy even though running speed has no real bearing on this particular game. If they'll do it for this, then it stands to reason that players will be even more willing to go for this in a game where running faster actually provides some kind of advantage over other players.
The other day I visited my local cinema with some friends to watch I, Robot. We went to the pub for a drink first and after queueing and my friends buying expensive food products we got in to the theatre proper about 35 minutes after the billed start time, expecting to have missed the start. We were quite shocked (and, on that particular occasion, relieved) to find that we arrived in time to see the last preview trailer as well as the "don't let mobile phones ruin your movie" and the "Love Movies? Hate Piracy!" (Aarr!) stuff.
That's 40 minutes of junk before the film starts! Normally I don't notice the length of time because I'm chatting to my friends during this time. The ticket cost five pounds (roughly 9 US Dollars) and then once the film finally started the main character kept going on about his basketball shoes, which were of a brand I've never heard of and thus promptly forgot. The blatant nature of that product-placement was actually quite amusing, which was probably what they were going for.