home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
ftp.cs.arizona.edu
/
ftp.cs.arizona.edu.tar
/
ftp.cs.arizona.edu
/
icon
/
newsgrp
/
group95a.txt
/
000051_icon-group-sender _Sun Feb 5 13:43:56 1995.msg
< prev
Wrap
Internet Message Format
|
1995-02-09
|
3KB
Received: by cheltenham.cs.arizona.edu; Sun, 5 Feb 1995 12:44:02 MST
Date: Sun, 5 Feb 1995 13:43:56 -0600
From: jeffery@runner.jpl.utsa.edu (Clinton L. Jeffery)
Message-Id: <9502051943.AA05130@runner.utsa.edu>
To: goer@mithra-orinst.uchicago.edu
Cc: icon-group@cs.arizona.edu
In-Reply-To: <9502051751.AA10095@mithra-orinst.uchicago.edu> (goer@mithra-orinst.uchicago.edu)
Subject: Re: sockets, loadfunc, etc.
Content-Length: 2788
Errors-To: icon-group-errors@cs.arizona.edu
[Richard Goerwitz states:]
Loadfunc, as it stands now, also appears to lean heavily on features
specific to just a few Unix platforms.
I just don't see this as being true. What is true is that dynamic loading is
in general our best hope of providing you guys with your "lean core runtime
system" and letting you extend it. I think most modern operating systems
provide a dynamic loading facility (.dll's on DOS and OS/2, .lib's on Amigas,
.so's on UNIX...). loadfunc() is not UNIX specific, but like the graphics
interface, it must be implemented on each OS before it is available there.
Implementing it on most OS'es is trivial compared with implementing graphics.
[Richard goes on to write ..."the point is that, at version 8.0, even a
dummy like me could do this sort of thing."]
I am a bit tired of hearing that "Version 8 was so easy to extend that even
I could extend it, but I can never do that with Version 9". Version 8 was
hard to extend, Version 9 is in many ways easier. In Version 8 if your
extension had any interesting use of memory, interfacing it to the garbage
collector was a nightmare. In Version 9 this is much cleaner. In
Version 8 if you wanted to extend it your only choice was to relink your
iconx, making public distribution of your extension unlikely. In Version 9,
you can add it as a built-in to iconx or write some C code for loadfunc().
[John Paolillo agrees, seconding Richard's request that we "provide
facilities for dummies like me to access facilities already present
in our OS for doing them."]
The bottom line is that for the kinds of features you've been asking for
lately, like multi-byte character support or Macintosh OSA support, there
is NO MAGIC BULLET. There is no way the features can be implemented
without a small army of programmers, even if it is supported in a given OS.
The Icon community has to either pay for it to happen, or find a sponsor,
or implement it ourselves.
Should I be organizing an RTL training seminar, or an on-line RTL tutorial
group? I'd be willing to donate money to an Icon "internationalization
endowment fund", but it will take a lot of resources. If anyone wants to
come spend a few weeks or months to work on Icon, or wants to do an
Icon-related graduate degree, I'd be happy to do what I can. :-)
I appreciate how frustrating it is to love an obscure language that happens
to be designed better than so many popular languages with heavy industry
backing. The question is: how much do you love Icon? Enough to live with
its limitations and use it when its appropriate? Or enough to do something
yourself to make Icon better or support its development, besides posting news?
Clint Jeffery
cjeffery@cs.arizona.edu, jeffery@ringer.cs.utsa.edu
The University of Texas at San Antonio