Using PostScript as a General Purpose Computing Language

In article , Paul Hovnanian P.E. wrote: [...]

You are right. Its C I was showing the C error that got copied.

No, I'm on a Linux box. I tried the same program on 3 different JREs on linux and a couple on Windows. I could make it work in one but not the other and back and forth. I gave up. The whole reason I learned it was to have a portable language.

Now that there is a "bash" shell for windows, I think I may have finally found a portable language. :>

The ability to pass a pointer-to-a-function into a function can let you shrink some cases down hugely. It can get a bit hard to follow when the functions recure too.

At least they give you (1) the choice and (2) don't default to maybe a reference.

I do a lot of 8051 code. 256^65536 is more programs than I ever plan on needing to write.

The more important point is that it probably lacks microstepping

Have you ever seen any software that will load into a totaly mechanical keyboard/printer?

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

By changing the default program that opens that type of file when you double click on it.

In Windows Explorer (not IE) => Tools => Options => File Types => scroll down the list (or type "pd" to have it jump there) till you get to PDF and use the "change" button at the bottom to associate it with whatever program you want to open those files.

Or right clink on any file in Windows Explorer and choose "Open with" to choose a program to open it in but then you have to do that every time. Good for opening occasional programs in v6 or v7 and have v5 associated with the PDF filetype.

Robert

One could load some code into the old lineprinter used with the IBM

You could try Ghostscript:

gs -sDEVICE=pdfwrite -sOutputFile=result.pdf inputfile.pdf

but I'm not sure it reads everything. It is free and command line though

- not much chance it gets in the way of things.

Thomas

And 99.999% of them will crash.

John

That would be an interesting question to try to prove. We would have to define "crash" in this context through. Someone may actually want a program that rattles one bit of some port in what appears to be a random pattern etc.

Some modest fraction of all possible random programs would be the equivalent of branch-self or branch-to-start, with possibly some nonsense opcodes inside the loop. And some fraction of them might do something that's arguably useful. And of course, we'd have to define "useful" and "crash."

Academics periodically rediscover the concept of programming (and circuit design) by random mutation and natural selection. So far, I'm not worrying about my livelihood.

John

Simply measure their execution time on an infinite loop.

In article , John Larkin wrote: [....]

Doesn't Microsoft do that?

I've written programs that did this. It takes a truely huge amount of CPU time to make even a small improvement in a design. If it is going to all happen inside a processor without actually making and trying the circuits, there is an added problem. You have to specifiy a method for rating better vs worse to get the evolving circuits to trend towards the desired goal.

If you aren't careful, your system that minimizes the distortion in an amplifier will evolve to simply shorting the output. Coming up with the right way to specify the rating on a complex system isn't easy. It may even be about as hard as doing the design. I don't think anyone has yet suggested that the ratings systems be varied at random and the process run with each rating system to cause the rating systems to evolve. I could get a patent on this but have decided that in the public interest, I will not.

BTW: You still need a ratings system to rate the ratings system. I suggest that this can also be made to evolve ... etc.

In article , Don Lancaster wrote: [....]

I don't think that will work. The MTBF on our clocks isn't good enough.

Except the selection is unnatural: make your customers test it and pay you for the privilige.

Sounds like the classic Taguchi optimization of a voltage regulator circuit. When he was done with it, it had resistor values from milliohms to teraohms, near perfect load regulation, and no line regulation.

Just close the outer feedback loop on your bank balance.

John

"pdf2txt" works remarkably well on many complex documents.

There were the tines on the answerback drum....

Tim.

Relays are not "totally mechanical" are they?

Their companion GhostView allows one to do conversions via WinDoz Gooie.

What drove it? An electric motor?

that was my point :)

Ken