I didn't know languages had weight

I think what you mean is that PHP's interpreter is just slower. http://shootout.alioth.debian.org/u32q/benchmark.php?test=all&lang=php&lang2=python

CPython is faster and has smaller code size - straight up win-win right there.

Yeah, so... The last thing we worked on with this was getting password protected files working. We've got an idea that we are going to work with, I think. So now it is just a matter of time to get it to work right As of now, though, it compiles programs and gives you source that works. I think we are going to work on some default headers people can use with their source to make it easier. I forget if Souvik included a link, but if he has, it will probably change a little later.

Uh, why? And what do you even mean by "password protected files"?

We were trying to make sure people wouldn't try to overwrite other peoples' programs.

As in, it would keep your files on the server and then you could give someone a link to the file, and they would have to put the right password in to open the file. That way you could compile your projects and not have people steal your source code. Or is that a pointless idea, even for just practice?

You should have people have different accounts. Or give people a hash that uniquely identifies the files they've uploaded, or something along those lines.

Hehe, I have fond memories of you insisting that Python was superior to Java in speed.

I hope you sandboxed the assembler heavily; Brass is not designed for use in this manner at all.

(Even the worst assembler written in PHP would likely be faster than Brass, which was horribly written).

I seriously doubt that based on my experiences with your excellent software, Ben. Although I do agree that sandboxing is important with any such software being run with arbitrary input from the interwebz.

Ben's probably right - Brass has a very long start-up time by way of the .NET runtime. With a big project like Doors this is reasonably amortized, but for small projects it's absurd.

SPASM isn't a paragon of clean coding, but it is *much* faster than Brass.

As nice as it may be to blame that on .NET, it's quite clearly a problem with the shoddiness of the parser in Brass - the larger the project, the slower it becomes. (Other fun bugs include infinitely recursive macro expansion; try #define foo foo then invoke "foo"). SPASM is a far better piece of software.

Well, I'm still happy enough with it, and I have yet to find any fatal glitches that make Doors CS not build properly.