keyboards need tactile feedback.
on
Keyless Keyboard
·
· Score: 2
The problem with virtual glove keyboards is that you can't *feel* the keys, which means you can't type fast with tactile feedback. This is also why those touch-screen controls featured in Star Trek Next Generation will never replace keyboards either. You have to look at the screen to type.
Re:The Best Goddamn Keyboard in the World, Ever
on
Keyless Keyboard
·
· Score: 2
I loved the old heavy keyboards on the Indy machines, but not the newer ones on the o2's. Those old heavies were awesome - all the solid durability of the IBM clickity-clack keyboards, but with a smooth feel instead of the annoying sound. There's a couple of old Indy's at work that hardly get used anymore, and I can tell you the temptation to swipe those keyboards is really strong.
That would be fine, so long as *fradulent* includeed the common abuse of language where they use the past tense to describe what are merely statistical predictions. It is a *lie* to say, "So-and-so has won in Florida", when no official
count has even begun yet in Florida and all they have are guesses based on exit polls. (And they should be slapped around for calling it when it was so close that the difference of votes is lost in the course granularity of the exit polling techniques. (The difference is so small that when scaled down to the small population of an exit poll, it amounts to a fraction of one individual.)
Note that individual communities are financially responsible for their own balloting techniques, which is why different systems are in place all around the country. Changing this requires more than just money - it requires a lot of law changes if you want to standardize, and there would be resistance from communities that want to retain their independance of federal control.
But that said, here's some standards *I* would like to see, assuming that they can even be implemented, which I have my doubts about.
MAKE IT ILLEGAL TO PUBLISH EXIT POLL RESULTS UNTIL AFTER ALL POLLS ARE CLOSED IN THE COUTNRY!!! Every other country does this, but we don't. This practice probably skews the vote more than balloting irregularities.
Require the implementation of systems that provide immediate user feedback when the ballot is improperly filled-out, so the voter is made aware of the problem right away and has the chance to try again. For example, with scanned paper ballots, if two candidates for the same race are picked, the machine could spit the ballot back out right away as the voter tries putting it in the machine, and the voter could try again with another ballot. This is implemented in some places already. Since honest mistakes will happen, it is unconsionable to have a person's change to vote rejected without their knowlege.
To aid in the above, for each question on the ballot, there should be a way to make a mark for "abstain", and then it could be made an error when a question is left blank. This prevents people from accidentally missing a question they didn't realize was there. Here in Wisconsin, some years the ballots have enough referenda that questions spill over onto the back of the ballot, but not always. It would be nice if the user got notification when questions like that are missed.
I would *not* switch to an on-line system, but I would reccomend switching to an electronic counting system with paper backup. Why? Because it's easier to break into an on-line system than a sneaker-net one, and if the sneaker-net system is broken into, it's more likely to be noticable. The paper backup is essential so that in case of fraud allegations there is a more permanent, unhackable source to go look at. (So, I envision a system where you go up to a voting "kiosk" computer, make your selections on a touchscreen, then hit "submit" when you are done. At that point, a paper record is spit out into a lockbox, and your electronic vote is recorded. If your selections are invalid, you get an error message telling you why. These kiosks would be in public view of the election volunteers (but the screens would be shrouded), so that if someone spends an inordinate amount of time (perhaps trying to break out of the kiosk interface), this gets noticed.
Don't abolish the electoral college, but *do* change it so that you are not electing *people* who do the voting, but are simply talling numbers for each state. The fact that it is perfectly legal for an elector in the college to change their vote is unacceptable. I can understand the need to force politicians to pat attention to rural areas with the weighted system we have, but I don't understand this archaic idea that the electors could in theory stage a coup and just ignore what their states voted for and pick something else.
But the EC that you are praising has the opposite problem: more say per person for those in rural areas. If you live in a less populous state, say with only one rep and two senators, for a total of 3 EC votes, and let's say your state's population is only 800,000. (let's say for the sake of argument that the ratio is currently 800,000 people per representative - I don't know what the actual number is.) Then your vote is worth 3/800,000 (0.00000375) of an EC vote. Compare this to someone living in a more populous state, with
18.3 million people ( 23 reps + 2 senators, for 25 EC votes.) Then your individual vote is only worth 25/18,300,000 (0..00000136), roughly 1/3 as much as the person living in the rural state.
As you say, lack of EC gives unfair power to urbanized areas at the expense of the needs or rural ones, but I'm saying the that EC in place has the opposite effect, giving people in rural areas more than their fair say. (I theorize that this is one of the reasons religion plays a big part in politics - religion is more popular in the rural areas where people's individual opinions have more political sway.)
This goes out to all those who say the voters should have just asked for help if the ballot was confusing:
1.
What if it wasn't obvious that they needed help? There was no indication that they had done anything wrong when they filled out the ballots, and the error was not caught until later. Where I voted (Fitchburg, WI), they have you stick your ballot card into the scanner right away on your way out of the room. That way it is counted right away, and you are the only one that touches your ballot, so someone can't alter it after you leave. This system makes for much faster counting (But it isn't used everywhere in Wisconsin, which is why it was one of the 'undecided' states for so long). Anyway, presumably if the card contained invalid input (like two choices for president), an
error indication could have been given right then and there, before I left. (Although I don't know if this is really what would happen if I'd filled out the card wrong.) Presumably in Florida this isn't what they did, so those who voted erroneously didn't have any feedback that their ballot was going to be thrown out for errors.
This should be importat to people who have been talking about this as a user interface issue. Sure, they showed the ballots to people ahead of time and got no complaints, but the real UI mistake here wasn't in the ballot itself, it was in the lack of error feedback when using it. (Something you can't test for by just showing people what the ballot is going to look like).
2.
Asking for help destroys your right to a secret ballot. (Sometimes it's inevitable, as in the case of the blind, but if you think you can avoid it, you generally don't want to tell people at the booth whom you are voting for.)
You misunderstand me. I wasn't talking about inconsistent coding style, but style that assumes a color highlighter is available - for example, only indenting 2 spaces is ugly even if you do it universally, and having no whitespace in expressions is ugly even if you do it universally.
It's inherently impossible for the software to get any better. Consider the task it's facing - it has to build a list of the whole of the world wide web, and categorize it. Even Yahoo, Google, and their brethren can't do that.
So they can use cpu-expensive algorithms to look for meaning in phrases and censor more accurately, but doing so means they don't have enough CPU time to examine more than a small part of the WWW. So if they want to filter enverything, then they have to do it clunky and badly so they can build the lists quickly enough.
Now, you might argue that hardware will get faster and thus be able to examine content faster, but that ignores the fact that the WWW itself is also going to be growing during that time too, so the problem space will be larger when the software is faster, so you probably won't gain any ground.
It is possible to run a filter that operates on
*all* traffic on certain port numbers that it
propigates. For example, here at UW-Madison,
all web traffic goes through a proxy (for cacheing
purposes, not censoring purposes), whether you like it or not. Anything going out on port 80 gets filtered through the proxy at the IT department, regardless of your own settings. I've even seen error messages from the proxy server show up in lynx. (when the remote site is down, the UW proxy server issues you an error message. Were it not for that message, I'd never have realized the proxy server existed. None of my local settings on the browser say anything about it.)
I've worked in a place where *some* of the programmers had access to syntax highlighting
editors and others did not. That made for terrible looking code. The problem is that code
that looks like it has plenty of white space in color mode doesn't really when you turn off the color syntax - things are run together and badly
indented because the color highlighting made things look like they were still easily discernable and hid this fact. This is especially bad in printouts.
Look, you DOWNLOADED a current copy of kde from
their website, but didn't DOWNLOAD the latest qt
that it uses, instead you took the out-of-date one
from the CD in your hand. Of course that doesn't work. That's the same as, say, in Windows trying to use a game that requires DirectX7, and then installing DirectX5 from your old Win install CD.
If you had also downloaded the qt that the KDE site *has a URL pointing you to*, then it wouldn't have been as messed up. I'm not denying that it's a lot of steps, and it's not easy, but your example situation seems contrived to be harder than it needs to be on purpose.
Yeah, but think of the really neato sports games
that could be done inside a spinning cylnder habitat. Imagine trying to toss a ball back and
forth with curvy coriolis-induced arcs. It could
make for a wicked game of Ping Pong or racketball.
It's like an anglophone's pronunciation of the french letter "r".
Anglophones and States-siders can't even agree on the pronounciatuion of the English "r", much less the french one. (The British "r" is often fainter than the US "r", so that to my US-bred hearing I don't even hear the "r" in a lot of words when a British person speaks them. Then there's the extra "r"s that get thrown in where they aren't in the spelling, like this word "idear" I keep hearing on imported British TV shows, I've had many "ideas", but I have no idea what an idear is.)
The reason some of us remember the old games fondly is because we didn't place as high a value on the irrelevancies like graphics and sound. What has improved in the last two decades is not the storylines or dynamic gameplay, but the whiz-bang effects. Big Deal. Some of those past games were really good, just as good as today's games if you ignore the whiz-bang window dressing.
Here's my list, off the top of my head, and what made the game good in my opinion: GAME....................GOOD FOR Jumpman(c64)...........Needed tactical think-ahead movement to avoid death.
.......................Variety of traps and tricks.
Impossible Mission.....Groundbreaking for complexity of the "map".
.......................Required logical thought to evade the robots.
.......................That scream sound when you fall to your death was addictive.
.......................(The first time I played I kept throwing myself to my
.......................death just to hear it again.)
Phantasie 1,2,3........Predecessor to the Ultima series, by the same guy.
.......................Fun map exploration, party-style D&D setting.
.......................Could change the item descriptions in a text file, so
.......................after a while I was fighting with "trashcan lids",
......................."Big sticks", and so on.
Atari 4-paddle games (like Warlords, Quadrapong, etc) - Instead of
.......................the impossible task of making the computer an interesting
.......................challenge, it put the players against each other, but
.......................in the same room. (Moderm net play is too impersonal.)
Civ (original).........Nice game length, and a user interface that was efficient.
.......................(unlike Civ:CTP).
Lemmings...............Hilarious premise, and a fun challenge to boot.
Does this mean all new games are bad? No, just that the extra graphics and sound don't really make a game good all by themselves. It's important to have a good, *fun* idea at the core of the game. Ask yourself the question, "Would this game still be fun if the graphics and sound weren't as good?" If the answer is "no", then there's no substance to the game.
One modern game I do like is Thief 1&2. Why? Because they had the balls to try something *different* instead of just another Duke Quakem clone. Much like Ultima Underworld (which was also by Looking Glass), this game was groundbreaking and unique without needing the best graphics of the day - they relied on other stuff to make it a good game - like a plot, and a realistic set of physics. (While the graphics engine in Thief wasn't the greatest, the realistic physics for throwing projectiles and shooting bows was awesome. Having an FPS where you have to arc your shot was a new idea, and they did a good job with it.)
I thought you were going to name some good songs. In other words, songs that don't all sound exactly the same as each other, unlike the pop pablum you listed. (The only exception being Enter Sandman - one of the last Metallica songs from back when they were still good.) There are better examples of unique work that was produced (mostly) in the '90s:
Primus
Midnight Oil (Yes, they are still going)
Dave Mattews Band
And then there's all the earlier bands that were not started in the '90s but still were producing good new work in the 90's, like Rush, Bad Religion, Iron Maiden, etc.
Anecdotes are not as good as carefully created, unbiased tests. But since such tests don't exist, it's all there is to go on if you want to avoid the lies associated with marketting.
Once I was in a hotel room where they had a similar setup, where you were supposed to only be able to get to internet explorer and nothing else. This was a bit annoying as I wanted to test a few things with "nslookup" and "ping" - a site wasn't responding and I wanted to see what was up with it. There's a trick around this - just use a file: URL in the browser to get to the PC's C: drive. From there I could click on \windows\command.exe and it ran it. I don't know how well this technique would work elsewhere, though.
Go to the preferences window and turn off Javascript support, then you can start closing off the windows and they'll stay closed. They use some Javascript hooks to implement that annoying feature. (You can register a javascript function to run on window-exit, so they set one up that opens a new window when you exit one).
It's annoying that assholes like that have to go and ruin a perfectly fun thing like Javascript and it's because of them that I usually browse with it turned off.
(Side question about slashdot - why is "submit" set up as the default option when you hit return on a form? That's annoying. The reason for the above blank post is because I hit return after typing in the subject, out of habit.)
Did anyone go far enough down to see the C code
for use with that Word/RFC thing? Blech! I am very glad I got out of Windows programming long ago. What the hell were they doing calling "_close" instead of just "close"? What ever happened to leading underscore identifiers being for *internal* things? I doubt a program that just converts a text file from one format to another really needs to be making low-level leading-underscore calls deep into the bowels of the libraries like that. What the hell is wrong with these people who program at Microsoft? That's like steering your car by going under the
hood and manhandling the front wheels, as opposed to using that round thing that sticks out of the dashboard.
I was helping a friend once when he couldn't get his e-mail. The situation was this: He used to be on AOL - he moved to a local ISP instead, but still had leftover time on his AOL account that had not run out yet. He couldn't get to his new mail one day because the local ISP's modems were out of service in his area (I never found out why). So, knowing that once you get on through AOL you have a working TCP/IP connection, I suggested that he log in to his old AOL account just to get a connection, then bring up his normal (non-AOL) mail client and point it at his ISP's POP server. It sort of worked. He could read mail that way, but when he tried sending it, AOL blocked it. A popup came up saying that he is only allowed to send mail through an AOL mail program.
Although they grudgingly brought themselves into the internet age by finally using a normal TCP/IP stack, AOL still tries to act like a insular glorified BBS system at times.
People are bitching about the actual ads themselves, when that's not really the terrible part of all this. The terrible part is that it prevents the students from choosing their own software on their own computers. That's a much more blatant abuse of power than the forced advertising (At least from the descriptions I've seen this is what a lot of universities are doing - I don't have firsthand experience with using this system.)
They say they want to be the "Windows" of campus "Webtops". Let's see how they're doing:
- Try to force students to have to use their software and no other.
- Get it forced on the techies by going over their heads and giving their bosses a bullshit story.
- Combine programs that already exist, and call it something new.
- Try to spread by convincing the ignorant masses and saing "screw you" to to the knowlegable minority.
Yup, they're the Windows of campus webtops alright - good job guys!
That's not the way I read what he said. It sounds like he meant "user-land" vs "kernel-land", not "desktop-land" vs "server-land". It's too high-level to be a Linux issue, since Linux is just the kernel. It's like mentioning a flaw in Netscape and mistakenly calling it a Linux flaw.
It wasn't about end-users vs programmers.
Take advantage of the fact that these people are idiots who don't realize the internet is about more than just web sites. Put DeCSS on an FTP site instead. Post it to an appropriate.binaries newsgroup. Instead of "linking" to it, just
mention the URL in plain text (not in an HREF=...) and tell people "Go to...... to see DeCSS", that way it's technically not a link.
Encrypt it and use DMCPA against them for once. Make part of the "licence agreement to download" be that you have to agree not to sue before you can decrypt it. Heck, for added measure, use CSS to do the encryption. (Hmm - on second thought, no. That defeats the purpose since presumably the downloading person doesn't have DeCSS yet, that's why he's downloading it.)
Put it in seperate pieces, to be cat'ted together,
with links to each piece.
Instead of putting it in text file form, make an obfuscated C program that prints it out as output.
Have people run the C program.
In a "newspaper cut-out" fashion like they use in the movies, find a lot of image files from out on the 'net that have pictures of letters. (I'm sure if you look long enough, you can find some from different sites). Then, make a whole
lot of image URLS that point at those letters to
make a "cutout letter" version of DeCSS (Hmm, that could be fun.).
Slashdot Mirror
The problem with virtual glove keyboards is that you can't *feel* the keys, which means you can't type fast with tactile feedback. This is also why those touch-screen controls featured in Star Trek Next Generation will never replace keyboards either. You have to look at the screen to type.
I loved the old heavy keyboards on the Indy machines, but not the newer ones on the o2's. Those old heavies were awesome - all the solid durability of the IBM clickity-clack keyboards, but with a smooth feel instead of the annoying sound. There's a couple of old Indy's at work that hardly get used anymore, and I can tell you the temptation to swipe those keyboards is really strong.
That would be fine, so long as *fradulent* includeed the common abuse of language where they use the past tense to describe what are merely statistical predictions. It is a *lie* to say, "So-and-so has won in Florida", when no official count has even begun yet in Florida and all they have are guesses based on exit polls. (And they should be slapped around for calling it when it was so close that the difference of votes is lost in the course granularity of the exit polling techniques. (The difference is so small that when scaled down to the small population of an exit poll, it amounts to a fraction of one individual.)
But that said, here's some standards *I* would like to see, assuming that they can even be implemented, which I have my doubts about.
As you say, lack of EC gives unfair power to urbanized areas at the expense of the needs or rural ones, but I'm saying the that EC in place has the opposite effect, giving people in rural areas more than their fair say. (I theorize that this is one of the reasons religion plays a big part in politics - religion is more popular in the rural areas where people's individual opinions have more political sway.)
1.
What if it wasn't obvious that they needed help? There was no indication that they had done anything wrong when they filled out the ballots, and the error was not caught until later. Where I voted (Fitchburg, WI), they have you stick your ballot card into the scanner right away on your way out of the room. That way it is counted right away, and you are the only one that touches your ballot, so someone can't alter it after you leave. This system makes for much faster counting (But it isn't used everywhere in Wisconsin, which is why it was one of the 'undecided' states for so long). Anyway, presumably if the card contained invalid input (like two choices for president), an error indication could have been given right then and there, before I left. (Although I don't know if this is really what would happen if I'd filled out the card wrong.) Presumably in Florida this isn't what they did, so those who voted erroneously didn't have any feedback that their ballot was going to be thrown out for errors. This should be importat to people who have been talking about this as a user interface issue. Sure, they showed the ballots to people ahead of time and got no complaints, but the real UI mistake here wasn't in the ballot itself, it was in the lack of error feedback when using it. (Something you can't test for by just showing people what the ballot is going to look like).
2.
Asking for help destroys your right to a secret ballot. (Sometimes it's inevitable, as in the case of the blind, but if you think you can avoid it, you generally don't want to tell people at the booth whom you are voting for.)
You misunderstand me. I wasn't talking about inconsistent coding style, but style that assumes a color highlighter is available - for example, only indenting 2 spaces is ugly even if you do it universally, and having no whitespace in expressions is ugly even if you do it universally.
So they can use cpu-expensive algorithms to look for meaning in phrases and censor more accurately, but doing so means they don't have enough CPU time to examine more than a small part of the WWW. So if they want to filter enverything, then they have to do it clunky and badly so they can build the lists quickly enough.
Now, you might argue that hardware will get faster and thus be able to examine content faster, but that ignores the fact that the WWW itself is also going to be growing during that time too, so the problem space will be larger when the software is faster, so you probably won't gain any ground.
It is possible to run a filter that operates on *all* traffic on certain port numbers that it propigates. For example, here at UW-Madison, all web traffic goes through a proxy (for cacheing purposes, not censoring purposes), whether you like it or not. Anything going out on port 80 gets filtered through the proxy at the IT department, regardless of your own settings. I've even seen error messages from the proxy server show up in lynx. (when the remote site is down, the UW proxy server issues you an error message. Were it not for that message, I'd never have realized the proxy server existed. None of my local settings on the browser say anything about it.)
I've worked in a place where *some* of the programmers had access to syntax highlighting editors and others did not. That made for terrible looking code. The problem is that code that looks like it has plenty of white space in color mode doesn't really when you turn off the color syntax - things are run together and badly indented because the color highlighting made things look like they were still easily discernable and hid this fact. This is especially bad in printouts.
Real programmers use:
vi /dev/sda
Look, you DOWNLOADED a current copy of kde from their website, but didn't DOWNLOAD the latest qt that it uses, instead you took the out-of-date one from the CD in your hand. Of course that doesn't work. That's the same as, say, in Windows trying to use a game that requires DirectX7, and then installing DirectX5 from your old Win install CD. If you had also downloaded the qt that the KDE site *has a URL pointing you to*, then it wouldn't have been as messed up. I'm not denying that it's a lot of steps, and it's not easy, but your example situation seems contrived to be harder than it needs to be on purpose.
Yeah, but think of the really neato sports games that could be done inside a spinning cylnder habitat. Imagine trying to toss a ball back and forth with curvy coriolis-induced arcs. It could make for a wicked game of Ping Pong or racketball.
Here's my list, off the top of my head, and what made the game good in my opinion:
GAME....................GOOD FOR
Jumpman(c64)...........Needed tactical think-ahead movement to avoid death.
.......................Variety of traps and tricks.
Impossible Mission.....Groundbreaking for complexity of the "map".
.......................Required logical thought to evade the robots.
.......................That scream sound when you fall to your death was addictive.
.......................(The first time I played I kept throwing myself to my
.......................death just to hear it again.)
Phantasie 1,2,3........Predecessor to the Ultima series, by the same guy.
.......................Fun map exploration, party-style D&D setting.
.......................Could change the item descriptions in a text file, so
.......................after a while I was fighting with "trashcan lids",
......................."Big sticks", and so on.
Atari 4-paddle games (like Warlords, Quadrapong, etc) - Instead of
.......................the impossible task of making the computer an interesting
.......................challenge, it put the players against each other, but
.......................in the same room. (Moderm net play is too impersonal.)
Civ (original).........Nice game length, and a user interface that was efficient.
.......................(unlike Civ:CTP).
Lemmings...............Hilarious premise, and a fun challenge to boot.
Does this mean all new games are bad? No, just that the extra graphics and sound don't really make a game good all by themselves. It's important to have a good, *fun* idea at the core of the game. Ask yourself the question, "Would this game still be fun if the graphics and sound weren't as good?" If the answer is "no", then there's no substance to the game.
One modern game I do like is Thief 1&2. Why? Because they had the balls to try something *different* instead of just another Duke Quakem clone. Much like Ultima Underworld (which was also by Looking Glass), this game was groundbreaking and unique without needing the best graphics of the day - they relied on other stuff to make it a good game - like a plot, and a realistic set of physics. (While the graphics engine in Thief wasn't the greatest, the realistic physics for throwing projectiles and shooting bows was awesome. Having an FPS where you have to arc your shot was a new idea, and they did a good job with it.)
- Primus
- Midnight Oil (Yes, they are still going)
- Dave Mattews Band
And then there's all the earlier bands that were not started in the '90s but still were producing good new work in the 90's, like Rush, Bad Religion, Iron Maiden, etc.Anecdotes are not as good as carefully created, unbiased tests. But since such tests don't exist, it's all there is to go on if you want to avoid the lies associated with marketting.
Once I was in a hotel room where they had a similar setup, where you were supposed to only be able to get to internet explorer and nothing else. This was a bit annoying as I wanted to test a few things with "nslookup" and "ping" - a site wasn't responding and I wanted to see what was up with it. There's a trick around this - just use a file: URL in the browser to get to the PC's C: drive. From there I could click on \windows\command.exe and it ran it. I don't know how well this technique would work elsewhere, though.
It's annoying that assholes like that have to go and ruin a perfectly fun thing like Javascript and it's because of them that I usually browse with it turned off.
(Side question about slashdot - why is "submit" set up as the default option when you hit return on a form? That's annoying. The reason for the above blank post is because I hit return after typing in the subject, out of habit.)
Did anyone go far enough down to see the C code for use with that Word/RFC thing? Blech! I am very glad I got out of Windows programming long ago. What the hell were they doing calling "_close" instead of just "close"? What ever happened to leading underscore identifiers being for *internal* things? I doubt a program that just converts a text file from one format to another really needs to be making low-level leading-underscore calls deep into the bowels of the libraries like that. What the hell is wrong with these people who program at Microsoft? That's like steering your car by going under the hood and manhandling the front wheels, as opposed to using that round thing that sticks out of the dashboard.
Although they grudgingly brought themselves into the internet age by finally using a normal TCP/IP stack, AOL still tries to act like a insular glorified BBS system at times.
People are bitching about the actual ads themselves, when that's not really the terrible part of all this. The terrible part is that it prevents the students from choosing their own software on their own computers. That's a much more blatant abuse of power than the forced advertising (At least from the descriptions I've seen this is what a lot of universities are doing - I don't have firsthand experience with using this system.)
They say they want to be the "Windows" of campus "Webtops". Let's see how they're doing: - Try to force students to have to use their software and no other.
- Get it forced on the techies by going over their heads and giving their bosses a bullshit story.
- Combine programs that already exist, and call it something new.
- Try to spread by convincing the ignorant masses and saing "screw you" to to the knowlegable minority. Yup, they're the Windows of campus webtops alright - good job guys!
That's not the way I read what he said. It sounds like he meant "user-land" vs "kernel-land", not "desktop-land" vs "server-land". It's too high-level to be a Linux issue, since Linux is just the kernel. It's like mentioning a flaw in Netscape and mistakenly calling it a Linux flaw. It wasn't about end-users vs programmers.
Encrypt it and use DMCPA against them for once. Make part of the "licence agreement to download" be that you have to agree not to sue before you can decrypt it. Heck, for added measure, use CSS to do the encryption. (Hmm - on second thought, no. That defeats the purpose since presumably the downloading person doesn't have DeCSS yet, that's why he's downloading it.)
Put it in seperate pieces, to be cat'ted together, with links to each piece.
Instead of putting it in text file form, make an obfuscated C program that prints it out as output. Have people run the C program.
In a "newspaper cut-out" fashion like they use in the movies, find a lot of image files from out on the 'net that have pictures of letters. (I'm sure if you look long enough, you can find some from different sites). Then, make a whole lot of image URLS that point at those letters to make a "cutout letter" version of DeCSS (Hmm, that could be fun.).
Make the link say, "This is not a link to DeCSS".