Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Python vs. Object REXX?

506 views
Skip to first unread message

John Brock

unread,
Oct 26, 1996, 3:00:00 AM10/26/96
to

I noticed new book in the bookstore yesterday called "Programming
Python". I like the O'Reilly series, so I browsed through it, and it
looked pretty exciting. As background, I am a long time user of REXX
on VM/CMS and OS/2. I really like the way REXX does things, and I have
yearned for a similar language on Unix. (Yes, I know there are Unix
versions of REXX, but they don't seem to have much chance of being
adopted as a widespread Unix standard, and Perl just doesn't scratch
that itch for me). I have never used Object REXX, but I understand it
is a fairly straightforward object oriented extension of REXX, and I am
looking forward to giving it a try when I get around to installing Warp
4.0. I wonder if there is anyone out there who is familiar with both
Python and Object REXX who could compare them for me, in particular in
terms of:

General Similarity. Does Python seem in any way "REXX like" to someone
who also knows REXX, or is it a completely new experience? With REXX I
feel like I am writing pseudocode that works. Is Python like that?

Ease of Learning and Use. REXX just seemed to make sense for me, and
once I learned it it stayed learned. With Perl I was always running
back to the book. Mike Cowlishaw's REXX book is remarkably thin, while
"Programming Python" is rather fat; does this tell me something about
the learning curve I should expect with Python?

Features. Are the two languages of similar power? Are there important
features present in one that are absent in the other?

Extensibility. Programs that are called by REXX can directly set
variables in the calling REXX program. For example, under VM REXX if I
wanted to pull data from an SQL/DS database I could call a program
called RXSQL from within a REXX program and pass it a string containing
a SELECT statement (including a FROM clause containing the names of
certain variables in the calling REXX environment). Once it started
executing RXSQL would directly read the named REXX variables to
complete the FROM clause, pull the selected data from the database,
write the values into a REXX array (also named in the input string),
and then exit. This is a very convenient method for extending a
language, since it requires absolutely no changes to the language
itself. With Perl and Sybase under Unix it was necessary to compile a
new, extended version of Perl itself, called sybperl, in order to
achieve the same functionality. This is *not* convenient. So is there
a similar easy way to extend the functionality of Python without
creating new versions?

Portability. How easy is it to write programs that will run unchanged
under different operating systems?

Regularity and Orthogonality of Syntax. Which language is more
"logical", and freest from quirks and weirdities? And why does
"weirdities" get past my spell checker? :-O

An additional point I'd like to make is that one of the features of
REXX that I really like is its infinite precision arithmetic. I
noticed that Python has infinite length integers; is there any chance
of adding infinite length floats (and math functions acting on them) to
Python?
--
John Brock
jbr...@panix.com

Chris Trimble

unread,
Oct 26, 1996, 3:00:00 AM10/26/96
to

I'm not versed in REXX, but I hope this can be helpful for you...

John Brock <jbr...@panix.com> wrote:

: With REXX I feel like I am writing pseudocode that works.
: Is Python like that?

When I first started coding in Python, I kind of had that feeling as
well. There is a lot going on behind the scenes when I say stuff like "if
mykey is in list", and I really appreciate those features of Python. Now
I've come to expect those features from any language I use... :-)


: Ease of Learning and Use. REXX just seemed to make sense for me, and


: once I learned it it stayed learned. With Perl I was always running
: back to the book. Mike Cowlishaw's REXX book is remarkably thin, while
: "Programming Python" is rather fat; does this tell me something about
: the learning curve I should expect with Python?

Python stays with you. I go back a book (or HTML docs) for library
function syntax far more than anything else; sometimes there's something
weird that comes up, and I usually just try it out in the interactive
mode. Just like you said, I find myself continually looking back to the
book for Perl questions, or just copying other code because I could never
get a real grasp of what is going on in some of those hack PERL5
extensions. (I suppose I need the NEW perl book, right?). Python, in my
opinion, is very solid conceptually, and that makes it easy to hold on to.

I was frightened with the size of Mark Lutz's book as well ... I thought
maybe there was a lot of the language that I didn't know about and made it
as nutso as PERL :-). Seriously, there is stuff I've learned about from
his book; a lot of that space is taken up in covering going GUIs with
Python. To check out the basic language stuff, you should look at Guido's
original documentation, which has a pretty good tutorial to start,
http://www.python.org/doc/

The next documentation project is that thin book you're talking about,
which I think is very possible (wink wink, nudge nudge, say no more)


: Features. Are the two languages of similar power? Are there important


: features present in one that are absent in the other?

I don't really know... sorry.


: Extensibility. Programs that are called by REXX can directly set


: variables in the calling REXX program. For example, under VM REXX if I
: wanted to pull data from an SQL/DS database I could call a program
: called RXSQL from within a REXX program and pass it a string containing
: a SELECT statement (including a FROM clause containing the names of
: certain variables in the calling REXX environment).

[...]
: So is there a similar easy way to extend the functionality of Python
: without creating new versions?

For the first part, to preserve a more modular design, you'd generally
import a Python module and pass an retrieve data in a functional fashion
instead of setting variables in the calling module. The key to the second
part is, most of the time that imported module will be Python code, but it
can also be a DSO (DLL, whatever). So you can import compiled C
functionality into Python pretty seemlessly, which is a convincing factor
for a lot of Python users (like me). On the Python site, there are DSO
modules like this for mSQL, Informix, Oracle, etc... check them out:
http://www.python.org/ftp/python/contrib/Database


: Portability. How easy is it to write programs that will run unchanged
: under different operating systems?

It depends how different you're talking about. There is a general
portable "OS" module, which remains consistent for most stuff, but not
all. Across UNIXes, you probably won't have to change anything. From
UNIX to Mac or UNIX to MS-DOS, there may be some things that need to be
changed. There are provisions for os-independent path separators and
such, but not everything you need is probably implemented for MS-DOS.


- Chris

Bernie Schneider

unread,
Oct 26, 1996, 3:00:00 AM10/26/96
to

jbr...@panix.com (John Brock) wrote:

>I noticed new book in the bookstore yesterday called "Programming
>Python". I like the O'Reilly series, so I browsed through it, and it
>looked pretty exciting. As background, I am a long time user of REXX
>on VM/CMS and OS/2. I really like the way REXX does things, and I have
>yearned for a similar language on Unix. (Yes, I know there are Unix
>versions of REXX, but they don't seem to have much chance of being
>adopted as a widespread Unix standard, and Perl just doesn't scratch
>that itch for me). I have never used Object REXX, but I understand it
>is a fairly straightforward object oriented extension of REXX, and I am
>looking forward to giving it a try when I get around to installing Warp
>4.0. I wonder if there is anyone out there who is familiar with both
>Python and Object REXX who could compare them for me, in particular in
>terms of:

I've been programming in Rexx, Python, and Perl for some time now, and
in general, I like Rexx the best for its simplicity, readability,
power, and ease of learning. I like Python too, but I dropped it
because I'd already learned Perl (what a struggle!!), and Python
seemed to not offer any more function than Perl, so I just didn't need
it. If I had it to do over again, I'd probably learn Python because
it's just so much easier to learn, and then not need Perl.

>General Similarity. Does Python seem in any way "REXX like" to someone

>who also knows REXX, or is it a completely new experience? With REXX I


>feel like I am writing pseudocode that works. Is Python like that?

I wouldn't call Python "Rexx like", but it does share many of Rexx's
best qualities: ease of learning, very readable code, etc. It doesn't
have variable substitution, but the dictionary data type provides for
associative arrays. Python (and Perl) both have built in functions for
managing associative arrays that Rexx lacks (Object Rexx does however,
add those missing functions).

>Ease of Learning and Use. REXX just seemed to make sense for me, and
>once I learned it it stayed learned. With Perl I was always running
>back to the book.

Me too! I will have to buy the new Perl 5 "Camel" book mainly because
I've completely worn out my version 4 manual. I seldom have to refer
to my Rexx manual, because it "stayed learned".

> Mike Cowlishaw's REXX book is remarkably thin, while
>"Programming Python" is rather fat; does this tell me something about
>the learning curve I should expect with Python?

It probably does. Although Python's learning curve shouldn't be as
steep as Perl's, I have yet to find a language that is as easy to
learn and master as Rexx. I find MFC's book remarkable for its brevity
and completeness. The book sort of mirrors the language in that
regard. By the way, you should see his newest language, NETREXX. It's
a dialect of Rexx that emits Java source code, which can then be
compiled and run on any platform that has a JDK. And since NETREXX is
itself written in NETREXX, it too can be used on any of those same
platforms from one standard set of binaries! You can mix NETREXX
classes with Java classes, including the standard Java API, in any way
you like to produce some pretty interesting programs. Yet NETREXX
implements Cowlishaw's ability to boil a language down to its
essentials, simplify it, and still give you tremendous power, just as
Rexx does. I'm just now getting into it, but its learning curve
appears to be to Java's as Java's is to C++'s. Sorry for the
digression...

>Features. Are the two languages of similar power? Are there important
>features present in one that are absent in the other?

Python has object oriented features, which Rexx lacks, but which
Object Rexx now has. One of the nicest features of Rexx, the PARSE
instruction, has been implemented in Python as regular expression
pattern matching, and oh, how I wish Rexx had regular expression
pattern matching as part of the PARSE syntax. Python is much more
modular than Rexx, you can import separately compiled modules into a
Python program. Managing program modules is quite a chore in Rexx, but
once again, Object Rexx provides some relief in this area. Both
languages have crude, but effective debuggers built in.

>Extensibility. Programs that are called by REXX can directly set
>variables in the calling REXX program. For example, under VM REXX if I
>wanted to pull data from an SQL/DS database I could call a program
>called RXSQL from within a REXX program and pass it a string containing
>a SELECT statement (including a FROM clause containing the names of

>certain variables in the calling REXX environment). Once it started
>executing RXSQL would directly read the named REXX variables to
>complete the FROM clause, pull the selected data from the database,
>write the values into a REXX array (also named in the input string),
>and then exit. This is a very convenient method for extending a
>language, since it requires absolutely no changes to the language
>itself. With Perl and Sybase under Unix it was necessary to compile a
>new, extended version of Perl itself, called sybperl, in order to

>achieve the same functionality. This is *not* convenient. So is there


>a similar easy way to extend the functionality of Python without
>creating new versions?

With Python, you can imbed the language in an application as you can
in Rexx, and although I haven't tried it myself, Python seems to have
a bit better facilities for handling external programs than Rexx does.
I'd like to see better features put into Rexx for accessing DLLs
without the messiness of writing C interface layers.

>Portability. How easy is it to write programs that will run unchanged
>under different operating systems?

Just as in Rexx, Python is portable across many platforms, but of
course, you do have to watch out for OS specific stuff. In Python,
there are a number of standard libraries provided which, if used
consistently, can transparently handle many OS differences, and there
is a POSIX library to provide further assistance here too.

>Regularity and Orthogonality of Syntax. Which language is more
>"logical", and freest from quirks and weirdities? And why does
>"weirdities" get past my spell checker? :-O

I find Rexx to be a bit more "logical", intuitive, and more free of
surprises than Python, but Python's syntax is still a breath of fresh
air compared to Perl's. One of Python's "wierdities" is the
syntactical _requirement_ to use indenting to indicate control
structure nesting. This isn't bad, but it's different, and it can lead
to some interesting debugging sessions. An editor that self-indents
can be a big help with Python.

>An additional point I'd like to make is that one of the features of
>REXX that I really like is its infinite precision arithmetic. I
>noticed that Python has infinite length integers; is there any chance
>of adding infinite length floats (and math functions acting on them) to
>Python?
>--

Don't know about that one.

>John Brock
>jbr...@panix.com

One other point, calling REXX "archaic", as another poster just did,
is meaningless in this context other than just to bash any non-Perl
language. What difference does the age of a language make if it's
still doing the job for you? Python has some advances and advantages
over Rexx, that's true enough. But for my money, no language gives you
the ease of learning, ease of use, and power with simplicity that Rexx
does. Python is an excellent language, that makes a lot of sense if
you really need to switch from Rexx. But what's wrong with Rexx in a
Unix environment? Just because it's little used there, doesn't
necessarily mean that it can't be effective. And there's no learning
curve to a language you've already learned.

Hope this helps you. Good luck.

-- Bernie Schneider --


Tom Christiansen

unread,
Oct 26, 1996, 3:00:00 AM10/26/96
to

[courtesy cc of this posting sent to cited author via email]

In comp.lang.python,
jbr...@panix.com (John Brock) writes:
:With Perl and Sybase under Unix it was necessary to compile a


:new, extended version of Perl itself, called sybperl, in order to
:achieve the same functionality. This is *not* convenient.

This is ancient history. Perl is happy to link into run-time DLLs without
a problem. No need for new versinos. Sounds like your only experience
with Perl is the old perl4 stuff -- both the program and the book. Each
has been updated now into something rather quite better.

Python is a much better language for many purposes than perl4 was.
Perl5, however, is a different story, since we stole all the cool python
stuff and made it unix-friendly. :-) [I'm just teasing; it's much more
elaborate than that.] As for REXX, it's a pretty archaic language whose
principle advantage is that it's not JCL. Some have said that it might
as well be interpreted FORTRAN, but it's better than a shell script.
But then so is a boot to the head.

--tom
--
Tom Christiansen tch...@alumni.cs.colorado.edu


The secretaries don't understand me. --Rob Pike

Fredrik Lundh

unread,
Oct 28, 1996, 3:00:00 AM10/28/96
to

> > Mike Cowlishaw's REXX book is remarkably thin, while
> >"Programming Python" is rather fat; does this tell me something about
> >the learning curve I should expect with Python?

> It probably does. Although Python's learning curve shouldn't be as
> steep as Perl's, I have yet to find a language that is as easy to
> learn and master as Rexx.

IMO, "Programming Python" is more of an encyclopedia than a
tutorial, so you'll definitely fool yourself if you start counting
pages. Some 70% of the book are dedicated to modular/OO design (does
REXX support OO at all?), GUI techniques, object persistency, C/C++
extending/embedding issues (150 pages), debugging, application design,
etc. You can use Python in many ways, and PP covers with most of
them.

But if you read the first 300 something pages, you'll learn almost
everything about the language itself that you'll ever need. And if
that's too much, there's an even more compact tutorial in an appendix.
And don't forget Guido's 70-page Python tutorial that's included in
the standard distribution.

Cheers /F

Mark Lutz

unread,
Oct 28, 1996, 3:00:00 AM10/28/96
to

> Mike Cowlishaw's REXX book is remarkably thin, while
>"Programming Python" is rather fat; does this tell me something about
>the learning curve I should expect with Python?

Well, size isn't everything ;-). IMHO, Python is a remarkably
simple language, and you can't really learn much by measuring
the size of a book.

The O'Reilly Python book is big because it covers lots of territory:
both introductory and advanced language material, common extensions,
etc. And it's geared more towards showing how Python can be applied
in realistic systems; it's not just a language definition.

As with any technical book, you can pick the parts you need. For
instance, part-2 covers the language itself, and part-3 presents
more advanced topics like GUIs, text processing, and C integration.
It's as comprehensive as you want it to be.

Mark L.
[okay, so I'm a little biased :-)]

Mark Lutz

unread,
Oct 29, 1996, 3:00:00 AM10/29/96
to

> Back to the book, though. I for one would have preferred two or
> three thinner volumes to a single behemoth. Easier to carry around.

That's a valid point, and FWIW we considered it too. But then
you'd have to pay two or three times to get a complete picture,
and I wasn't crazy about that idea.

Think of it as "2 for the price of 1" :-),
Mark L.

John Turner

unread,
Oct 29, 1996, 3:00:00 AM10/29/96
to

lu...@rmi.net (Mark Lutz) writes:

> The O'Reilly Python book is big because it covers lots of territory:
> both introductory and advanced language material, common extensions,
> etc. And it's geared more towards showing how Python can be applied
> in realistic systems; it's not just a language definition.
>
> As with any technical book, you can pick the parts you need. For
> instance, part-2 covers the language itself, and part-3 presents
> more advanced topics like GUIs, text processing, and C integration.
> It's as comprehensive as you want it to be.

I picked the book up today, and can't wait to try some Python.

I've fought with Fortran 77 for many years, but have loved doing
object-based prog. with F90 over the past few. I've also done
enough Tcl/Tk to be a bit frustrated with Tcl. Never learned much
perl because I could do what I needed with awk. Also use octave
for various quick linear algebra things. Anyway, I'm excited
about what I've read so far about Python. Sounds like it may fill
many of these needs.

Back to the book, though. I for one would have preferred two or
three thinner volumes to a single behemoth. Easier to carry around.

But who asked me?

(I definitely like all the Monty Python references, though.)

--
John A. Turner
Los Alamos National Laboratory, MS B226, Los Alamos, NM 87545
Group: XTM (Radiation Transport Methods)
Location: TA-3, Bldg. 43, Rm. D150
Phone: 505-665-1303 e-mail: tur...@lanl.gov

Joseph Skinner

unread,
Oct 30, 1996, 3:00:00 AM10/30/96
to

Having used REXX under OS/2 and under unix I find that python is in many
ways a much nicer language to work with.

My reasons for saying this are that python has many of the features that
I always found missing from REXX.

For an example of this take the case of returning a list or agregate
type from a function under REXX. For all pratical purposes you can't do
this without a real hack. Under python you cannot only return a list you
can also return multiple values which is a real advantage in many cases.

ie

def some_function():
...
return an_error_code, mylist


error_code, my_list = some_function()

These and many other factors make python great once you get into larger
program sizes.

Also when compared to Object REXX I found that python is the sort of
language that I thought Object REXX should have been.

Joe.

Andre Doff

unread,
Oct 30, 1996, 3:00:00 AM10/30/96
to

In article <3275FDB3...@earthlight.co.nz>, Joseph Skinner
<j...@earthlight.co.nz> wrote:

JS> Under python you cannot only return a list

With REXX, I think you can use a stem to do this, but I don't know
how close that comes to a Python list (don't know Python).

JS> you can also return multiple values which is a real advantage
JS> in many cases.

Also possible with REXX, e.g. RETURN 1 2 3. With PARSE, the
individual variables can be extracted from that.

Ave,

Andre Doff
ad...@ibm.net

Mike Cowlishaw

unread,
Oct 31, 1996, 3:00:00 AM10/31/96
to

In <yFzdyAIB...@ibm.net>, ad...@ibm.net (Andre Doff) writes:
>In article <3275FDB3...@earthlight.co.nz>, Joseph Skinner
><j...@earthlight.co.nz> wrote:
>
> JS> Under python you cannot only return a list
>
> etc...

Further, this thread is comparing Object REXX with Python; in Object
REXX, as in NetRexx, any object may be returned from a routine. This
includes collections such as lists and vectors.

--
Mike Cowlishaw, IBM Fellow
http://www2.hursley.ibm.com


Ian Collier

unread,
Oct 31, 1996, 3:00:00 AM10/31/96
to

In article <54t5rn$bmf$1...@csnews.cs.colorado.edu>, tch...@mox.perl.com (Tom Christiansen) wrote:
> As for REXX, it's a pretty archaic language whose
>principle advantage is that it's not JCL. Some have said that it might
>as well be interpreted FORTRAN, but it's better than a shell script.
>But then so is a boot to the head.

Now now, that's just flame bait. What does the age of a language (and I
believe we're talking no earlier than 1978 here) have to do with its power
or useability? You yourself are using a derivative of a language which has
been around in more or less its present form for many centuries.

You are presumably aware that you posted this to a newsgroup whose readers
are well aware of the advantages of REXX over and above the fact that it is
not JCL (or EXEC, for that matter). Incidentally, how many languages can
you name that have ANSI standards?...

Ian Collier - i...@comlab.ox.ac.uk - WWW Home Page (including REXX section):
http://www.comlab.ox.ac.uk/oucl/users/ian.collier/index.html

Rony.Fl...@wu-wien.ac.at

unread,
Nov 1, 1996, 3:00:00 AM11/1/96
to

In <3275FDB3...@earthlight.co.nz>, Joseph Skinner <j...@earthlight.co.nz> writes:
-Having used REXX under OS/2 and under unix I find that python is in many
-ways a much nicer language to work with.
-
-My reasons for saying this are that python has many of the features that
-I always found missing from REXX.
-
-For an example of this take the case of returning a list or agregate
-type from a function under REXX. For all pratical purposes you can't do
-this without a real hack. Under python you cannot only return a list you
-can also return multiple values which is a real advantage in many cases.
-
-ie
-
-def some_function():
- ...
- return an_error_code, mylist
-
-
-error_code, my_list = some_function()
-
-These and many other factors make python great once you get into larger
-program sizes.

As many pointed out this is possible with Object Rexx too (and in form of a
parsable string even in classic Rexx).

-Also when compared to Object REXX I found that python is the sort of
-language that I thought Object REXX should have been.

Now, that you seem to know both Python and Object Rexx (at least that is how I
read the above paragraph) I would like you to point out the differences between
the two and to supply examples where you feel that Python is superiour to Object
Rexx, and I would like yout to point out examples where you feel that Object
Rexx is superiour to Python.

Please note, that I do not want to regress to a flame-war, but would like to
sorberly discuss factual issues.

If you are not able to come up with examples than I suggest we stop this thread.

---rony

a...@icm.com

unread,
Nov 1, 1996, 3:00:00 AM11/1/96
to tur...@branagh.lanl.gov

In article ,
John Turner wrote:

>
> lu...@rmi.net (Mark Lutz) writes:
> Back to the book, though. I for one would have preferred two or
> three thinner volumes to a single behemoth. Easier to carry around.
> But who asked me?
> (I definitely like all the Monty Python references, though.)

There is a thinner book available. One of the goals was to give the
reader (assumed to have a bit of programming experience) a good
feel for the language in chapters 1+2 (71pp) and to give basically
all you need to know about Python in chapters 3..6 (172pp). The
rest of the book is devoted to example applications and fancy stuff
like embedding and extending.

Oh yes: it's called Internet Programming with Python, Watters, Van Rossum
and Ahlstrom, M&T books, isbn 1-55851-484-8 ($34.95US).

-- Aaron Watters (also aaron_...@msn.com)
===
Howard: Do I shock you?
Mom: Howard, I have 3 kids: nothing shocks me.
(excerpt from Howard Stern)

-----------------------------------------------------------------------
This article was posted to Usenet via the Posting Service at Deja News:
http://www.dejanews.com/ [Search, Post, and Read Usenet News!]

Robert McDermid

unread,
Nov 1, 1996, 3:00:00 AM11/1/96
to

In message <559u3j$p...@zen.hursley.ibm.com> - m...@vnet.ibm.com (Mike

Cowlishaw)31 Oct 1996 10:13:07 GMT writes:
>
>Further, this thread is comparing Object REXX with Python; in Object
>REXX, as in NetRexx, any object may be returned from a routine. This
>includes collections such as lists and vectors.
>
>--
>Mike Cowlishaw, IBM Fellow
>http://www2.hursley.ibm.com
>


Hey, it's cool to see a post from you here (for people who don't
know, Mike, among other things, is the creator of both Rexx and
NetRexx). I'd just like to pay my compliments on an excellent
job with NetRexx; I just finished reading the language specification,
and am hacking my way through the awt classes. I'm just itching
to start writing some code with it. All in all, it seems very
well designed. I'm a big C++ fan, but even so, I find NetRexx to
be much cleaner and a more elegant design by far. Do you think
there's any chance of getting VisualAge Java to also support
NetRexx? That'd be slick.

-- Rob

==============================================================
Rob McDermid Hummingbird Communications Ltd.
mcde...@hcl.com All opinions expressed are my own.
==============================================================


Joseph Skinner

unread,
Nov 5, 1996, 3:00:00 AM11/5/96
to

Mike Cowlishaw wrote:
>
> In <yFzdyAIB...@ibm.net>, ad...@ibm.net (Andre Doff) writes:
> >In article <3275FDB3...@earthlight.co.nz>, Joseph Skinner
> ><j...@earthlight.co.nz> wrote:
> >
> > JS> Under python you cannot only return a list
> >
> > etc...

>
> Further, this thread is comparing Object REXX with Python; in Object
> REXX, as in NetRexx, any object may be returned from a routine. This
> includes collections such as lists and vectors.

Yes I agree and have used this in object REXX under OS/2 and found it
rather nice.

Joe.

ps. Is there any word on Object REXX being released for linux. This was
mentioned last year but I never heard anything more on it.

a...@icm.com

unread,
Nov 6, 1996, 3:00:00 AM11/6/96
to Robert McDermid

In article ,
mcde...@hcl.com (Robert McDermid) wrote:
>
> In message - m...@vnet.ibm.com (Mike

> Hey, it's cool to see a post from you here (for people who don't
> know, Mike, among other things, is the creator of both Rexx and
> NetRexx). I'd just like to pay my compliments on an excellent
> job with NetRexx; I just finished reading the language specification,
> and am hacking my way through the awt classes. I'm just itching
> to start writing some code with it. All in all, it seems very
> well designed. I'm a big C++ fan, but even so, I find NetRexx to
> be much cleaner and a more elegant design by far. Do you think
> there's any chance of getting VisualAge Java to also support
> NetRexx? That'd be slick.
> -- Rob

It'd also be slick to see an implementation of diff/patch in netrexx,
etc. Let's see some code examples and enough of the vagueries.
-- Aaron Watters
ps: I posted a diff/patch in Python 2 days ago. ;-)
===
All this chitter chatter chitter chatter chitter chatter
about shmatter shmatter shmatter shmatter shmatter
I can't give-it-away on 7th Avenue! - Mick Jagger

(hoop sheyoobee). -- Mick Jagger.

Mike Cowlishaw

unread,
Nov 9, 1996, 3:00:00 AM11/9/96
to

In <847295...@dejanews.com>, a...@icm.com writes:
> Let's see some code examples and enough of the vagueries [sic].

The best collection of NetRexx code examples can be found at:

http://www.multitask.com.au/netrexx/fac/

(Dion Gillard's collection, including all the examples from 'Java in a Nutshell', in
NetRexx.)



--
Mike Cowlishaw, IBM Fellow

http://www2.hursley.ibm.com/netrexx/


aaron_...@msn.com

unread,
Nov 10, 1996, 3:00:00 AM11/10/96
to Mike Cowlishaw

In article ,
mc...@hursley.ibm.com (Mike Cowlishaw) wrote:

>
> In , a...@icm.com writes:
> > Let's see some code examples and enough of the vagueries [sic].
> The best collection of NetRexx code examples can be found at:
> http://www.multitask.com.au/netrexx/fac/
> (Dion Gillard's collection, including all the examples from 'Java in a Nutshell', in
> NetRexx.)

Well since there are hardly any examples in that book that aren't trivial
that's not much of a trick. I'd like to see the examples from our book in
NetRexx... I suspect you'd see an average 3x code explosion, but I've
been wrong before. (Internet Programming with Python, ISBN1-55851-484-8,
M&T Books).

...of course that'd still be less than half the size of java translations...
-- Aaron Watters.
===
How many times in the past has the computer industry been reinvented
from scratch? Lessee, there was 5th generation computing, lisp machines,
the Ncube and other massively parallel machines -- all of them extremely
commercially successful, as we know today. Now there's java chips...

-----------------------------------------------------------------------
This article was posted to Usenet via the Posting Service at Deja News:

http://www.dejanews.com/ [Search, Post, and Read Usenet News]

Mike Cowlishaw

unread,
Nov 11, 1996, 3:00:00 AM11/11/96
to

In <8476330...@dejanews.com>, aaron_...@msn.com writes:
> I'd like to see the examples from our book in
> NetRexx...

I'll try and pick up a copy of your book next time I'm in the USA (next week, in fact).
Doubt I'll recode any examples though -- there are too many new programs that
need to be written for time to be spent in recoding old ones!

0 new messages