Oops! Sorry!

Post #114,251

by gdaustin

8/18/03 11:00:32 PM

8/18/03 11:02:45 PM

Oops! Sorry!

But APL is interpreted!

So what is he complaining about? Should I revise the post to remove my blather about Delphi?

Edited by gdaustin Aug. 18, 2003, 11:02:45 PM EDT

Expand All History

Post #114,265

by deSitter

8/19/03 1:03:42 AM

Re: Oops! Sorry!

APL is not P-code, nor is Smalltalk. Smalltalk is byte-code, which is much different and like FORTH, "compiled interpretation". BASIC is none of the above. P-code is stupid and inefficient, was rightly rejected in 1980 (UCSD Pascal), and is barely a step above BASIC.

APL, FORTH, and Smalltalk all have a REAL VM with direct, immediate interpretation, not a stupid fucked-up P-code interpreter.

-drl

Post #114,292

by static

8/19/03 7:41:03 AM

I think it depends on your definition of "p-code".

Nowadays, to most people the term "p-code" means rather more than what UCSD Pascal used to create. It's true that Smalltalk create a intermediate form often called "byte-code", but it is virtual assembly language for the specific virtual machine. Quite a lot of modern languages work like this - so much that the stage in front normally called a compiler has a common name: translater. Even BASIC interpreters of past actually ran virtual code in a virtual machine: almost all of the home computer BASICs translated the entered text into tokens for storage - a pseudo-code, if you like. Of course, the single biggest thing you've overlooked is that all these pseudo-codes are extremely tailored for their language and their environment.

Wade.

Is it enough to love
Is it enough to breathe
Somebody rip my heart out
And leave me here to bleed

Is it enough to die
Somebody save my life
I'd rather be Anything but Ordinary
Please

-- "Anything but Ordinary" by Avril Lavigne.

Post #114,349

by jake123

8/19/03 3:44:47 PM

Rexx on warp does something similar

It creates tokenized code. That code is stored in the command file's extended attribute, leading to a much quicker load time on subsequent runs of the code. It works extremely well and really helps performance.

--\n-------------------------------------------------------------------\n* Jack Troughton                            jake at consultron.ca *\n* [link|http://consultron.ca|http://consultron.ca]                   [link|irc://irc.ecomstation.ca|irc://irc.ecomstation.ca] *\n* Kingston Ontario Canada               [link|news://news.consultron.ca|news://news.consultron.ca] *\n-------------------------------------------------------------------

Post #115,033

by static

8/23/03 3:00:48 AM

Good example.

Other variations would be how Perl and PHP do it - translation on invocation, p-code is thrown away after execution - and how Icon and Java do it - explicit translation to a second file. It's even obvious with Icon: the program you run is called icont - meaning Icon Translator.

Wade.

Is it enough to love
Is it enough to breathe
Somebody rip my heart out
And leave me here to bleed

Is it enough to die
Somebody save my life
I'd rather be Anything but Ordinary
Please

-- "Anything but Ordinary" by Avril Lavigne.

Post #115,247

by jake123

8/25/03 12:06:59 AM

I've actually been thinking about this for a while

I'm thinking that perhaps for my honour project (if I can afford a fourth year of school) I'd like to take a shot at writing an object rexx interpreter, to be licensed under the GPL. For one thing, there are some serious problems with the object rexx interpreter under warp when you want to interface it to programs written in C, since the RexxStart call doesn't account for the fact that object rexx programs can be threaded, which means that one can end up leaving open semaphores until process termination... which can terminally suck for daemons. The problem is that IBM won't fix it unless one wants to dole out some serious money. The other solution is a wholesale replacement of the subsystem... and if it was written properly, it should end up being easily portable to other systems: a dll in warp, a program in unix, and a plugin for wsh in windows.

I don't think that I'd be able to do a complete reimplementation or anything, but if I could get something that covered most of the basics, opening it up to the developer community by putting it under the GPL could make it something that would be viable in the long run.

This is all at least a year away, of course; I think I might be in a position to start looking at it seriously towards the end of this coming school year (second year).

--\n-------------------------------------------------------------------\n* Jack Troughton                            jake at consultron.ca *\n* [link|http://consultron.ca|http://consultron.ca]                   [link|irc://irc.ecomstation.ca|irc://irc.ecomstation.ca] *\n* Kingston Ontario Canada               [link|news://news.consultron.ca|news://news.consultron.ca] *\n-------------------------------------------------------------------

Welcome to IWETHEY!