I was wondering if it was possible to use Telnet as a rudimentary web browser and apparently it is. I managed to connect to Google, get various types of errors, do searches (by knowing the URL beforehand ) and "view" the Google logo. HTTP is... interesting.

Now to try to actually use the Google search box [or some other page's text input box]...

I imagine it depends on the species. A wasp can probably sting until it gets tired as a snake can bite until it's tired, but the amount of venom it can inject would depend heavily on the precise species.

Now, the real question is when are we going to see this over calcnet?

EDIT: Playing with random people on the internet isn't as fun as playing with omni people. The other guys just die two minutes in 

Thanks. I'll just ignore the META-INF folder since no one likes it

Anyway, apparently I screwed something up in the source and the Java instructions were lost. No idea what happened. Either way, I spent a few hours today and rewrote most of them to be faster and more efficient. Also, I slowed down the interpreter loop by an outrageous 400 nanoseconds to eliminate around 4 KB of parser overhead PER THREAD. The runtime memory used by the interpreter itself should now be roughly the same whether there are 2 threads in a given language or 20. The side effect to this is that the interpreter (ironically) has to ask for almost as much memory as was saved from the system

That's not the point. The point is to get better quality than is currently available in games.

The end all of all graphics and graphical computation. Welcome to the most advanced computer graphics the world has ever known.

Runer is right. It's possible, but there are some pretty big downsides. It's nowhere near the ultimate graphics technology and I doubt it will even displace current methods in a lot of applications.

For one thing, you're dealing with so many discrete points, you can never hope to apply things globally. Sorry, but computers aren't fast enough now or in the foreseeable future to apply rules to trillions of points multiple times a second. Secondly, the "unlimited detail" method doesn't allow animations. It's too slow for that. Thirdly, there is a ton of repetition in that model. If each point were discrete, the resulting map would literally take petabytes of data. For comparison, a low to medium resolution polygon mesh over that same area takes somewhere between 120 MB and 1 GB.

So, it's probably real, but remarkably useless.

A fast routine? You're probably better off doing floating point than 64 bit integer math on a processor with 16 bit multiplication.

Assuming d0 and d2 are the high words:

128 bit result: d0*d2*2^64+(d0*d3+d1*d2)*2^32+d1*d3. This uses 32 bit multiplication, a problem I assume you've already solved.

I don't have a division routine, but it sounds positively painful.