Alternatively, wait for the clueless masses to invest in the latest waste of money and take the old box off their hands. I find most people who I help with their systems don't really know what to do with two computers, so usually one goes to another family member or, failing that, is handed to me in return for years of dutiful maintanence service.
It's often the case that people have trouble selling their old systems because they are percieved as underpowered. Furthermore, unknowledgable people often don't feel confident buying a computer outside of a computer retailer because they don't get that oh-so-valuable "advice", so you're only selling to the subset of techies who read classified ads. Most non-techies don't know how to properly list and describe their PCs on (say) eBay, either.
I don't always accept these "donations", of course. Even a geek can have too many computers.
...and this is why all of the "average users" have 25000GHz systems with amazing amounts of RAM which still manage to run deathly slow while the techies happily trundle on with their 500MHz Pentium 3 chips. Upgrading became pointless years ago.
My last "upgrade" was to get a second machine to be a server of sorts for my house, and I was given that box (a 700MHz P3) by someone who was chucking it out to get a computer with a 3.5GHz Celeron chip. My main machine has a Pentium 2 chip. They both continue to serve me fine.
Once the components start to die I will "upgrade", but it has been my observation that most computers of this era are much better manufactured than the latest things. I've seen so many hardware failures in modern hardware for stupid things like accidentally unplugging an LED from a motherboard header with the power on. On my old 486 system I used to do all sorts of stupid things like plugging in RAM with the power on (although that was actually an accident because I'd unplugged the fan and didn't know it was on) and that machine still runs just fine today. My friend's completely-overpowered machine caught fire because of a power surge, and that's only the most extreme of the failures I've seen in the last few years with this cheap-as-possible hardware.
That doesn't mean anything. As far as the HTML is concerned, you just have a bunch of meaningless containers all nested inside each other. It's only in the presence of a stylesheet that -- for visual browsers only -- the nesting is available.
This is like using DIV CLASS="heading" instead of H1 and giving that class the appearance of a heading. Sure, in a visual browser to a sighted person it might look like a heading, but it isn't any more a heading than the DIV CLASS="paragraph" that follows it.
The cleanest solution is to stop using our browsers as glorified dumb terminals and return to the model where the client is responsible for presenting the data provided by the server. Think of an IMAP client vs. Hotmail, for example. HTML-based interfaces are handy when you don't have the right client software available, but they should be the second-class option, not the only option available. ForumZilla was briefly the answer for semantic foruming, but despite some quick uptake from software such as slash and scoop it faded from view rather than the "protocol" being used for other client software. I once wrote a HTML-based wrapper around ForumZilla, which is in my opinion the correct order of things: spit out the data in a machine-readable way first, and worry about making the HTML interface afterwards using that data.
Slashdot has a "light" template which was probably originally intended for lynx and ilk. It uses headings reasonably sensibly, and I read slashdot with it using my user stylesheet.
An issue at this point, not just with slashdot but with all similar sites, is that there is no decent HTML construct for marking up threaded discussions, so you either get table/css hacks or (in the rare case that the author is a fanatic) you get nested ordered lists with the markers hidden in CSS, which just makes a mess in non-CSS browsers.
One possible approach would be to put the expiry time in the clear in the cookie but also to hash it with the password so that (assuming the hashing algorithm is good) you can only edit the cookie if you know the password. Of course, the easier option would just be to have the session expire on the server, just like every other system under the sun.
You can't use the remote IP address, though, because several ISPs (AOL, for instance) use a pool of HTTP proxy servers to handle HTTP requests for users, so each request from the same host might well go through a different proxy. The HTTP headers for proxy tracking are useless in this situation because the client can just forge them to look like the legitimate user was proxying through the attacker's host.
Be careful. If you put the server in a copyright-free country but continue to reside in a country which has copyright laws, you could find yourself sued or prosecuted. Servers don't break laws, people do.
An analogue stick's data isn't much different to a mouse's data. The mouse sends the X and Y delta since the last "frame" and the stick sends the current current X and Y position relative to the centre.
Imagine if I move my mouse at a constant rate of one unit per second in the positive X axis. This creates the same data as holding the analogue stick at one unit in the positive X axis. Sure, it's not as usable as a mouse, but if you make the UI elements large enough you can hit them okay, and hopefully you'll be mostly using application that have been designed or adapted for the joypad interface so you won't have to use it very often.
I'd be weary of leaving an X-Box running permanently. I doubt it was designed with consistant usage in mind, so you'd better make sure there's good airflow around it and hope none of the cheaply-manufactured components kick the bucket.
My housemate brought in his original IBM PC out of the garage yesterday. We were quite pleased to see that it was still completely functional apart from the second floppy drive being a little knackered at the floppy with DOS on it having deteriorated a little. Old hardware is quite resiliant compared to current hardware, where manufacturers cut as many corners as possible to shave off manufacturing costs so they can compete with each other.
I still use a machine with a 350MHz Pentium II chip for almost everything I do, and I do a much wider variety of things than your average user would do.
Unfortunately, I don't have a word processor installed so I can't do what I wanted to do and load an "average user"'s set of software and see how fast it runs, but suffice it to say that currently I have a bunch of text editors, my web browser, a VoIP client, an IM client, my email client, an IRC client, a whole tonne of terminals running application-wise, and in the background Apache and mySQL running, occasionally serving requests from other users of my network.
The thing all of these pieces of software have in common is that they are interactive: they don't do anything unless the user is actively using them. The ones I'm not currently looking at are using a minimal amount of CPU perhaps processing the occasional packet, or whatever. The foreground application might occasionally have quite an intensive task to perform, but it's usually over within 30 seconds tops, and the scheduler ensures that the other apps get a chance to run anyway.
It's applications like games, video playback and so on that beat the CPU constantly that become a problem. Having said that, I regularly play back video files from over the network fullscreen in mplayer and with the use of some cache and the hardframedrop option there's no discernable degredation apart from the occasional sound stutter or decompression artifact where a key frame gets skipped. Realistically no action game since Quake III Arena would run on this machine, which is its only real downfall. I don't generally play computer action games, though, so it works well enough for me.
The hard part, of course, is finding a 400MHz CPU to buy new. Second hand could work (and that's how I got my CPU three years ago), but new you'd probably be looking at a lot more than something a thousand MHz faster just because there are no economies of scale attached to such a CPU anymore. If people became interested in a low-cost PC to the point where there was a demand for such CPUs, I'd assume that today with the lessons learned from the faster CPUs companies could make a much leaner, meaner "slow" CPU that runs a lot cooler and with much less power consumption than the Pentium 2 family did. The CPUs on the fanless EPIA Mini-ITX motherboards are a good example of this, but you can't buy those separately of course.
Imagine that I am watching television. (It doesn't happen often, but this is hypothetical.) You and a friend sit down next to me and have a conversation which distracts me from the television. Clearly, your conversation is going to bother me. Who is in the wrong here?
I've seen several compilers using object-oriented design. The most famous example is the Mono C# compiler. Most of them aren't entirely object-oriented, but there's not really any reason why you can't just make a small chunk of your program procedural if that design makes more sense.
Having said all that, I'm thinking of object-oriented in the more "static" sense that people are used to from Java, C++ and C#. Writing a compiler in a prototype-based OO language like JavaScript or Lua would perhaps be more "interesting"... there are some things which would get easier, and others which would get harder. That's true when you switch between procedural and OO as well, though.
I'm not disagreeing with you that OO isn't the answer to everything, though. Clearly there are problems where the "object" paradigm isn't the best approach.
In my humble opinion Photoshop has a terrible user interface. I used it for a short time on a friend's computer because it was all that was available and the interface was very non-intuitive and often required me to dig around menus (which were very oddly set-out -- presumably because they were too lazy to do a real Windows version) to find options with really bizarre names.
In fact, I'd say this is true of most Adobe software. Illustrator and InDesign, which I also dealt with for a short time, are similarly confusing. Perhaps a little confusion comes with having so many different features, but I think that partly the reason why people are attached to Photoshop is because, despite its odd UI, they've already learnt it. It would be nicer to try to create a better interface than Adobe's floating palettes and wacky menus.
(I always found that bits of my document got lost under the floating palettes. Why are these things considered to be so great? I guess it's just because of the MDI interface in Windows... it's probably better on a Mac.)
I'm currently running an automated information line for a small non-profit in the UK using VoIP. The only cost is running the server, which hasn't made a major impact on outgoings since the server was already there. In the UK, we have the special 0845 and 0870 dialling codes which are local and national rate respectively, regardless of where the caller is located.
The benefit of these numbers is that when they are used a slice of the call charge goes to the callee, meaning that there are companies willing to provide free services related to these numbers. One such service is for mapping an 0870 number to a SIP endpoint, although frustratingly I can't remember their name nor find them again right now. Asterisk is listening for incoming SIP connections from the incoming PSTN provider and then running an AGI script to deal with the incoming calls.
One nice thing about this is that this organisation has many PSTN lines and the non-profit has none, (well, except the one they make phone calls on!) so many calls can be handled simultaneously with the available PSTN lines shared between several SIP endpoints. The peak number of concurrent users has been five, though, so this is not a major thing I'm talking about. If you're going to be dealing with hundreds of concurrent calls this solution is probably not for you.
There are two obvious attacks on this system. The first is automatable, and just involves searching for common words like "the", "and" and "but" and scraping the results to work out which pages you're getting. Keep on at it until you have every page (or at least every page you are interested in) and you can rebuild an entire chapter or even an entire book locally. Since the highlighting is predictable it would be trivial to have a mapping from the highlight colors to the non-highlight colors.
The second attack is an interactive one, useful if you just want to read a chapter onscreen. What you do is find some way of finding the first, second or third page of the chapter (an exercise for the reader) and then use excerpts from page+2 to shift your "viewing margin" along each time you hit the limit.
The Perl bit is already written. You just need to write a set of Template Toolkit templates. I seem to remember from looking before that the way they are used in Slash is pretty obvious once you find the template files in the source distribution.
Write a little script to just feed junk data directly into one of their SMTP servers. Upload a significant amount (say, 250MB) and then just use the service to forward the message to yourself over and over. The local-to-local delivery should hopefully be relatively fast compared to your upload speed, so you'll be able to make progress quickly.
The articles on A List Apart create a static HTML mockup of the front page. However, there's more to Slashdot than the front page, and it's not just a matter of copying that mockup onto the site and having done with it. The Template Toolkit templates have to be rewritten to use the new code, and similar new markup and CSS must be written for things like comments, the comment form, the nutty little boxes and so on.
CowboyNeal has said repeatedly that if someone was to submit a complete, working template he'd consider making use of it. Also, more recently it was claimed that someone was working on one. The software that powers Slashdot is an open source project, and Template Toolkit is not specific to Slashdot and pretty well documented. If it really bothers you, scratch your own itch and submit a patch.
If a majority of files on Freenet were being illegally distributed, it could be argued that users of Freenet know that by using Freenet they are knowingly assisting with copyright infringment by providing part of the work at hand as there is a high probability that part of the data you are hosting is unlicenced. Fortunately, Freenet isn't widely used for copyright infringement right now, so no-one's really paying much attention to it.
I'd much rather that UK employers employ me once I'm done with my masters degree than try to encourage people from abroad to take all of the jobs. I was actually offered a job in the US before I started my masters degree but I was denied a VISA because the US (quite rightly, I'd say) gives preferential treatment to US citizens. Until there is a plethora of available jobs, I'd obviously prefer that they employ me rather than US citizens jumping ship to get away from stifling laws.
Microsoft will offer both versions, and OEMs will probably just take the version with media player attached, since otherwise they'd have to research other players to bundle instead and that's probably a lot more effort than they want to put in.
In this day and age, a PC vendor can hardly sell a PC without the ability to play music and video. That is can be added later is irrelevant to the masses who are used to just double-clicking the icon for some track in explorer and having it start playing in a program that doesn't even appear to have a name if you're running it in Skin Mode. It's just "the music thingy".
What made you think it was a good idea to put an apostrophe in "Gates", by the way?
Slashdot Mirror
Alternatively, wait for the clueless masses to invest in the latest waste of money and take the old box off their hands. I find most people who I help with their systems don't really know what to do with two computers, so usually one goes to another family member or, failing that, is handed to me in return for years of dutiful maintanence service.
It's often the case that people have trouble selling their old systems because they are percieved as underpowered. Furthermore, unknowledgable people often don't feel confident buying a computer outside of a computer retailer because they don't get that oh-so-valuable "advice", so you're only selling to the subset of techies who read classified ads. Most non-techies don't know how to properly list and describe their PCs on (say) eBay, either.
I don't always accept these "donations", of course. Even a geek can have too many computers.
...and this is why all of the "average users" have 25000GHz systems with amazing amounts of RAM which still manage to run deathly slow while the techies happily trundle on with their 500MHz Pentium 3 chips. Upgrading became pointless years ago.
My last "upgrade" was to get a second machine to be a server of sorts for my house, and I was given that box (a 700MHz P3) by someone who was chucking it out to get a computer with a 3.5GHz Celeron chip. My main machine has a Pentium 2 chip. They both continue to serve me fine.
Once the components start to die I will "upgrade", but it has been my observation that most computers of this era are much better manufactured than the latest things. I've seen so many hardware failures in modern hardware for stupid things like accidentally unplugging an LED from a motherboard header with the power on. On my old 486 system I used to do all sorts of stupid things like plugging in RAM with the power on (although that was actually an accident because I'd unplugged the fan and didn't know it was on) and that machine still runs just fine today. My friend's completely-overpowered machine caught fire because of a power surge, and that's only the most extreme of the failures I've seen in the last few years with this cheap-as-possible hardware.
Indeed. After spending a little while starting up (30-40s) both Firefox and Thunderbird run just fine on my P2-350.
That doesn't mean anything. As far as the HTML is concerned, you just have a bunch of meaningless containers all nested inside each other. It's only in the presence of a stylesheet that -- for visual browsers only -- the nesting is available.
This is like using DIV CLASS="heading" instead of H1 and giving that class the appearance of a heading. Sure, in a visual browser to a sighted person it might look like a heading, but it isn't any more a heading than the DIV CLASS="paragraph" that follows it.
The cleanest solution is to stop using our browsers as glorified dumb terminals and return to the model where the client is responsible for presenting the data provided by the server. Think of an IMAP client vs. Hotmail, for example. HTML-based interfaces are handy when you don't have the right client software available, but they should be the second-class option, not the only option available. ForumZilla was briefly the answer for semantic foruming, but despite some quick uptake from software such as slash and scoop it faded from view rather than the "protocol" being used for other client software. I once wrote a HTML-based wrapper around ForumZilla, which is in my opinion the correct order of things: spit out the data in a machine-readable way first, and worry about making the HTML interface afterwards using that data.
Slashdot has a "light" template which was probably originally intended for lynx and ilk. It uses headings reasonably sensibly, and I read slashdot with it using my user stylesheet.
An issue at this point, not just with slashdot but with all similar sites, is that there is no decent HTML construct for marking up threaded discussions, so you either get table/css hacks or (in the rare case that the author is a fanatic) you get nested ordered lists with the markers hidden in CSS, which just makes a mess in non-CSS browsers.
One possible approach would be to put the expiry time in the clear in the cookie but also to hash it with the password so that (assuming the hashing algorithm is good) you can only edit the cookie if you know the password. Of course, the easier option would just be to have the session expire on the server, just like every other system under the sun.
You can't use the remote IP address, though, because several ISPs (AOL, for instance) use a pool of HTTP proxy servers to handle HTTP requests for users, so each request from the same host might well go through a different proxy. The HTTP headers for proxy tracking are useless in this situation because the client can just forge them to look like the legitimate user was proxying through the attacker's host.
Be careful. If you put the server in a copyright-free country but continue to reside in a country which has copyright laws, you could find yourself sued or prosecuted. Servers don't break laws, people do.
An analogue stick's data isn't much different to a mouse's data. The mouse sends the X and Y delta since the last "frame" and the stick sends the current current X and Y position relative to the centre.
Imagine if I move my mouse at a constant rate of one unit per second in the positive X axis. This creates the same data as holding the analogue stick at one unit in the positive X axis. Sure, it's not as usable as a mouse, but if you make the UI elements large enough you can hit them okay, and hopefully you'll be mostly using application that have been designed or adapted for the joypad interface so you won't have to use it very often.
I'd be weary of leaving an X-Box running permanently. I doubt it was designed with consistant usage in mind, so you'd better make sure there's good airflow around it and hope none of the cheaply-manufactured components kick the bucket.
My housemate brought in his original IBM PC out of the garage yesterday. We were quite pleased to see that it was still completely functional apart from the second floppy drive being a little knackered at the floppy with DOS on it having deteriorated a little. Old hardware is quite resiliant compared to current hardware, where manufacturers cut as many corners as possible to shave off manufacturing costs so they can compete with each other.
The IBM PC box weighs more than its monitor!
I still use a machine with a 350MHz Pentium II chip for almost everything I do, and I do a much wider variety of things than your average user would do.
Unfortunately, I don't have a word processor installed so I can't do what I wanted to do and load an "average user"'s set of software and see how fast it runs, but suffice it to say that currently I have a bunch of text editors, my web browser, a VoIP client, an IM client, my email client, an IRC client, a whole tonne of terminals running application-wise, and in the background Apache and mySQL running, occasionally serving requests from other users of my network.
The thing all of these pieces of software have in common is that they are interactive: they don't do anything unless the user is actively using them. The ones I'm not currently looking at are using a minimal amount of CPU perhaps processing the occasional packet, or whatever. The foreground application might occasionally have quite an intensive task to perform, but it's usually over within 30 seconds tops, and the scheduler ensures that the other apps get a chance to run anyway.
It's applications like games, video playback and so on that beat the CPU constantly that become a problem. Having said that, I regularly play back video files from over the network fullscreen in mplayer and with the use of some cache and the hardframedrop option there's no discernable degredation apart from the occasional sound stutter or decompression artifact where a key frame gets skipped. Realistically no action game since Quake III Arena would run on this machine, which is its only real downfall. I don't generally play computer action games, though, so it works well enough for me.
The hard part, of course, is finding a 400MHz CPU to buy new. Second hand could work (and that's how I got my CPU three years ago), but new you'd probably be looking at a lot more than something a thousand MHz faster just because there are no economies of scale attached to such a CPU anymore. If people became interested in a low-cost PC to the point where there was a demand for such CPUs, I'd assume that today with the lessons learned from the faster CPUs companies could make a much leaner, meaner "slow" CPU that runs a lot cooler and with much less power consumption than the Pentium 2 family did. The CPUs on the fanless EPIA Mini-ITX motherboards are a good example of this, but you can't buy those separately of course.
The web isn't the only thing you can browse.
Imagine that I am watching television. (It doesn't happen often, but this is hypothetical.) You and a friend sit down next to me and have a conversation which distracts me from the television. Clearly, your conversation is going to bother me. Who is in the wrong here?
My TV has an off button which isn't a toggle. To turn it back on, you simply press a channel number.
I've seen several compilers using object-oriented design. The most famous example is the Mono C# compiler. Most of them aren't entirely object-oriented, but there's not really any reason why you can't just make a small chunk of your program procedural if that design makes more sense.
Having said all that, I'm thinking of object-oriented in the more "static" sense that people are used to from Java, C++ and C#. Writing a compiler in a prototype-based OO language like JavaScript or Lua would perhaps be more "interesting"... there are some things which would get easier, and others which would get harder. That's true when you switch between procedural and OO as well, though.
I'm not disagreeing with you that OO isn't the answer to everything, though. Clearly there are problems where the "object" paradigm isn't the best approach.
In my humble opinion Photoshop has a terrible user interface. I used it for a short time on a friend's computer because it was all that was available and the interface was very non-intuitive and often required me to dig around menus (which were very oddly set-out -- presumably because they were too lazy to do a real Windows version) to find options with really bizarre names.
In fact, I'd say this is true of most Adobe software. Illustrator and InDesign, which I also dealt with for a short time, are similarly confusing. Perhaps a little confusion comes with having so many different features, but I think that partly the reason why people are attached to Photoshop is because, despite its odd UI, they've already learnt it. It would be nicer to try to create a better interface than Adobe's floating palettes and wacky menus.
(I always found that bits of my document got lost under the floating palettes. Why are these things considered to be so great? I guess it's just because of the MDI interface in Windows... it's probably better on a Mac.)
I'm currently running an automated information line for a small non-profit in the UK using VoIP. The only cost is running the server, which hasn't made a major impact on outgoings since the server was already there. In the UK, we have the special 0845 and 0870 dialling codes which are local and national rate respectively, regardless of where the caller is located.
The benefit of these numbers is that when they are used a slice of the call charge goes to the callee, meaning that there are companies willing to provide free services related to these numbers. One such service is for mapping an 0870 number to a SIP endpoint, although frustratingly I can't remember their name nor find them again right now. Asterisk is listening for incoming SIP connections from the incoming PSTN provider and then running an AGI script to deal with the incoming calls.
One nice thing about this is that this organisation has many PSTN lines and the non-profit has none, (well, except the one they make phone calls on!) so many calls can be handled simultaneously with the available PSTN lines shared between several SIP endpoints. The peak number of concurrent users has been five, though, so this is not a major thing I'm talking about. If you're going to be dealing with hundreds of concurrent calls this solution is probably not for you.
There are two obvious attacks on this system. The first is automatable, and just involves searching for common words like "the", "and" and "but" and scraping the results to work out which pages you're getting. Keep on at it until you have every page (or at least every page you are interested in) and you can rebuild an entire chapter or even an entire book locally. Since the highlighting is predictable it would be trivial to have a mapping from the highlight colors to the non-highlight colors.
The second attack is an interactive one, useful if you just want to read a chapter onscreen. What you do is find some way of finding the first, second or third page of the chapter (an exercise for the reader) and then use excerpts from page+2 to shift your "viewing margin" along each time you hit the limit.
Google can't get away with this, surely?
...and configure his normal user account to not have "Administrator access" and show him how to use "Run As..." to install software.
The Perl bit is already written. You just need to write a set of Template Toolkit templates. I seem to remember from looking before that the way they are used in Slash is pretty obvious once you find the template files in the source distribution.
Write a little script to just feed junk data directly into one of their SMTP servers. Upload a significant amount (say, 250MB) and then just use the service to forward the message to yourself over and over. The local-to-local delivery should hopefully be relatively fast compared to your upload speed, so you'll be able to make progress quickly.
The articles on A List Apart create a static HTML mockup of the front page. However, there's more to Slashdot than the front page, and it's not just a matter of copying that mockup onto the site and having done with it. The Template Toolkit templates have to be rewritten to use the new code, and similar new markup and CSS must be written for things like comments, the comment form, the nutty little boxes and so on.
CowboyNeal has said repeatedly that if someone was to submit a complete, working template he'd consider making use of it. Also, more recently it was claimed that someone was working on one. The software that powers Slashdot is an open source project, and Template Toolkit is not specific to Slashdot and pretty well documented. If it really bothers you, scratch your own itch and submit a patch.
If a majority of files on Freenet were being illegally distributed, it could be argued that users of Freenet know that by using Freenet they are knowingly assisting with copyright infringment by providing part of the work at hand as there is a high probability that part of the data you are hosting is unlicenced. Fortunately, Freenet isn't widely used for copyright infringement right now, so no-one's really paying much attention to it.
I'd much rather that UK employers employ me once I'm done with my masters degree than try to encourage people from abroad to take all of the jobs. I was actually offered a job in the US before I started my masters degree but I was denied a VISA because the US (quite rightly, I'd say) gives preferential treatment to US citizens. Until there is a plethora of available jobs, I'd obviously prefer that they employ me rather than US citizens jumping ship to get away from stifling laws.
Microsoft will offer both versions, and OEMs will probably just take the version with media player attached, since otherwise they'd have to research other players to bundle instead and that's probably a lot more effort than they want to put in.
In this day and age, a PC vendor can hardly sell a PC without the ability to play music and video. That is can be added later is irrelevant to the masses who are used to just double-clicking the icon for some track in explorer and having it start playing in a program that doesn't even appear to have a name if you're running it in Skin Mode. It's just "the music thingy".
What made you think it was a good idea to put an apostrophe in "Gates", by the way?