Clean Room NetRexx: Waiting for Godot

[Robert L Hamilton]

I will point this out once again.  Could the problem be that many IBM products contain netrexx code; if they release it then any one can tinker with that code --  why would IBM ever release the foundation of IBM products and services.

Just think about it.

End the prattle about releasing NetRExx. It ani't gonna happen.

Bob Hamilton
Richardson, Texas USA

On Sun, May 1, 2011 at 11:13 PM, [hidden email] <[hidden email]> wrote:

Chip,

Many months ago I sent emails to both Rene and Virgil asking what the hold
up is.  Virgil said politely that he could not comment, and that Rene would
be best suited to answer.  The implication was that the ball was in the
RexxLA court.  I specifically asked Rene who was holding up the works, but
he did not answer my email.

Based on that, I can only assume that it is RexxLA that is dragging its
feet by not taking the initiative and working through the OSSC process,
possibly because RexxLA lacks the necessary resources, motivation or
management skills.  Could it be that each side is assuming the other will
perform various tasks?

I would certainly welcome finding out differently.

Instead of the sarcasm, as an insider why don't you privately discuss the
situation with Rene and Virgil, get to the bottom of the deadlock, and tell
us what the real story is?  Exactly who is waiting for whom to do what?

What I want Rene to do is to answer the question:  Has RexxLA done
everything in its power to comply with the IBM OSSC process?

If he can truthfully say yes and will document it, then we can legitimately
make the claim that IBM is intentionally not following through on its
promise to open the NetRexx source.

And if that is true, then we have the basis for loud complaints to IBM
Corporate, IBM legal and a publicity campaign with the open source
community.

On the other hand, if RexxLA is not performing, why not?

Frankly I see all the secrecy as just a cover up.  A little sunshine might
work wonders.

Is that a concrete enough?

Bill

PS.  By the way, other than Rene, who are the current RexxLA board members
and officials?  I'm wondering why they do not join this discussion.

On 5/1/2011 5:13 PM, Chip Davis wrote:
> You're right, George.  The time is 'way overdue.
>
> Therefore, in my capacity as a named "proper inheritor of NetRexx" (and
without any authority or permission from RexxLA or IBM) I hereby appoint
YOU to be the Lead Developer of Open NetRexx.
>
> Congratulations and Huzzahs all around!
>
> Now, by the end of the week I expect to see a Project Plan detailing
exactly what you intend to do with the IBM NetRexx source code.  Feel free
to get Bill, Thomas, and Fernando to help you.
>
> It is not that RexxLA "sees themselves as the proper inheritors of
NetRexx" as you snarkily put it.  It is the fact that IBM is simply not
going to give their IP to a loose aggregation of individuals connected by
nothing more substantial than a free listserver (that IBM is hosting).
>
> If you prefer, I suspect that at this point René would be happy to
appoint you to be the official NetRexx Transition Representative to IBM
(which will require you to be a RexxLA member of course) so we can all see
what sort of Sisyphus (or Estragon for that matter) you make.
>
> There is absolutely nothing RexxLA (or anyone else) can do to keep you
from starting on a clean-room version of NetRexx.  How hard could it be?
You have TNRL and Thomas' expertise at writing language processors, but not
the people who know the most about the language. Should be a piece of cake.
>
> What part of "It's in IBM's hands" is not clear?  If anyone has a
_concrete_ suggestion for an action that René could take that would have a
positive effect on the process, I'm sure we would all like to hear it.
>
> -Chip-

--------------------------------------------------------------------
mail2web LIVE – Free email based on Microsoft® Exchange technology -
http://link.mail2web.com/LIVE

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[Mike Cowlishaw]

In response to some questions/comments on this thread ... 

 

• Causes of delay:

I will point this out once again.  Could the problem be that many IBM products contain netrexx code; if they release it then any one can tinker with that code --  why would IBM ever release the foundation of IBM products and services. 

 

I doubt that is the problem -- there are some IBM products or parts of products written in NetRexx, but I am not aware of any other than my BigDecimal class that was distributed with NetRexx source code.

• On writing to IBM, I did so a few weeks ago. There is just one person charged with facilitating the NetRexx open source (among many other things). His reply was: "I am still working on this. We have made some progress, but it is slow going."

• On source for NetRexxC: it is all in NetRexx.  The compiler was originally written in Rexx (with very limited arithmetic, minimal control structures, etc.) then ported over to NetRexx.  Some parts of the code are probably very close to the original Rexx (and hence more 'procedural' than OO purists might prefer).
  
• Cleanroom should indeed be possible based on the published NetRexx2 (NRL+supplement) document at http://speleotrove.com/misc/NetRexx2.pdf ... it would be a fairly large piece of work, however: the current compiler is 26,500 LOC, runtime (Rexx class etc.) 4,200 LOC, testcases 5,500 LOC.  The arithmetic in the Rexx class could probably be back-modified from the BigDecimal classes.
  

Hope that helps.

 

Mike

 

 

 

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[George Hovey-2]

Mike,
This is the first piece of real news we have had in ages, and it seems (to me) rather positive: the problem is organizational and within IBM, but they haven't abandoned the effort; there are no signs of sinister forces at work.  Thanks for clearing the air.
George

On Mon, May 2, 2011 at 3:56 AM, Mike Cowlishaw <[hidden email]> wrote:

In response to some questions/comments on this thread ... 

 

• Causes of delay:

I will point this out once again.  Could the problem be that many IBM products contain netrexx code; if they release it then any one can tinker with that code --  why would IBM ever release the foundation of IBM products and services. 

 

I doubt that is the problem -- there are some IBM products or parts of products written in NetRexx, but I am not aware of any other than my BigDecimal class that was distributed with NetRexx source code.

• On writing to IBM, I did so a few weeks ago. There is just one person charged with facilitating the NetRexx open source (among many other things). His reply was: "I am still working on this. We have made some progress, but it is slow going."

• On source for NetRexxC: it is all in NetRexx.  The compiler was originally written in Rexx (with very limited arithmetic, minimal control structures, etc.) then ported over to NetRexx.  Some parts of the code are probably very close to the original Rexx (and hence more 'procedural' than OO purists might prefer).
  
• Cleanroom should indeed be possible based on the published NetRexx2 (NRL+supplement) document at http://speleotrove.com/misc/NetRexx2.pdf ... it would be a fairly large piece of work, however: the current compiler is 26,500 LOC, runtime (Rexx class etc.) 4,200 LOC, testcases 5,500 LOC.  The arithmetic in the Rexx class could probably be back-modified from the BigDecimal classes.
  

Hope that helps.

 

Mike

 

 

 

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[David Requena]

Many thanks, Mike, for a very informative reply! :-)

-
Saludos / Kind regards,
David Requena

NOTE: The opinions expressed here represent the opinions
of the authors and do not necessarily represent the opinions
of those who hold other opinions.


-----Original Message-----
From: "Mike Cowlishaw" <[hidden email]>
Sender: [hidden email]
Date: Mon, 2 May 2011 08:56:54
To: 'IBM Netrexx'<[hidden email]>
Reply-To: IBM Netrexx <[hidden email]>
Subject: RE: [Ibm-netrexx] Clean Room NetRexx: Waiting for Godot

_______________________________________________
Ibm-netrexx mailing list
[hidden email]



_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[Davis, James (NIH/CIT) [E]]

Mike or any,

 

A long shot – is any part of the TestCases public? (even a list of them)

 

Thanks,

 

James Davis
National Institutes of Health

[hidden email]

 

From: Mike Cowlishaw [mailto:[hidden email]]
Sent: Monday, May 02, 2011 3:57 AM
To: 'IBM Netrexx'
Subject: RE: [Ibm-netrexx] Clean Room NetRexx: Waiting for Godot

 

In response to some questions/comments on this thread ... 

 

• Causes of delay:

I will point this out once again.  Could the problem be that many IBM products contain netrexx code; if they release it then any one can tinker with that code --  why would IBM ever release the foundation of IBM products and services. 

 

I doubt that is the problem -- there are some IBM products or parts of products written in NetRexx, but I am not aware of any other than my BigDecimal class that was distributed with NetRexx source code.

• On writing to IBM, I did so a few weeks ago. There is just one person charged with facilitating the NetRexx open source (among many other things). His reply was: "I am still working on this. We have made some progress, but it is slow going."

• On source for NetRexxC: it is all in NetRexx.  The compiler was originally written in Rexx (with very limited arithmetic, minimal control structures, etc.) then ported over to NetRexx.  Some parts of the code are probably very close to the original Rexx (and hence more 'procedural' than OO purists might prefer).
• Cleanroom should indeed be possible based on the published NetRexx2 (NRL+supplement) document at http://speleotrove.com/misc/NetRexx2.pdf ... it would be a fairly large piece of work, however: the current compiler is 26,500 LOC, runtime (Rexx class etc.) 4,200 LOC, testcases 5,500 LOC.  The arithmetic in the Rexx class could probably be back-modified from the BigDecimal classes.

Hope that helps.

 

Mike

 

 

 

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[Robert L Hamilton]

If the Open Source netrexx project is short handed, why not see if some ex-IBMers might be willing to volunteer to help.????

Bob Hamilton
Richardson Texas USA

 

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[Mike Cowlishaw]

A long shot – is any part of the TestCases public? (even a list of them)

They are code, so the short answer is no.  Actually the most important testcase is the compiler itself -- uses almost everything.  But here's a list of the 'explicit' ones (ignore 'oldcode.nrx').

 

Mike

 

27/02/2000  19:22             1,910 DiagAbstract.nrx
22/02/2000  20:57             7,817 DiagAll.nrx
13/03/2000  22:13             3,419 DiagArrays.nrx
13/02/1997  23:11             2,556 DiagBean.nrx
27/02/2000  18:01             5,623 DiagComment.nrx
27/02/2000  15:58             2,292 DiagConcat.nrx
27/02/2000  17:58             5,478 DiagConstant.nrx
27/02/2000  18:00             1,931 DiagContinue.nrx
23/07/1999  17:36            59,136 DiagDecimal.nrx
31/05/2000  20:15            21,280 DiagExpression.nrx
26/04/2000  11:48             3,860 DiagForward.nrx
25/02/2000  13:18             5,344 DiagIf.nrx
06/04/2000  18:05             3,404 DiagIndir.nrx
07/04/2000  17:19             1,395 DiagInterface.nrx
01/01/1997  14:21             1,041 DiagLex.nrx
22/05/2001  09:49             7,933 DiagLoop.nrx
25/05/2000  20:32             6,425 DiagMinor.nrx
09/03/2000  20:42             2,028 DiagMutual.nrx
23/11/1997  14:26             1,644 DiagMutual2.nrx
25/07/1997  19:58             1,074 DiagMutual3.nrx
25/07/1997  19:59             1,125 DiagMutual4.nrx
27/02/2000  17:58             1,455 DiagNop.nrx
05/03/2000  14:18            10,037 DiagParse.nrx
15/03/2000  13:31            37,005 DiagRexx.nrx
24/02/2000  11:46             1,936 DiagSay.nrx
08/03/2000  21:36             7,918 DiagSelect.nrx
01/01/1997  14:21             1,079 diagskel.nrx
27/02/2000  17:53             3,617 DiagStems.nrx
27/02/2000  19:23             3,616 DiagUTF8.nrx
01/01/1997  14:21             1,049 DiagX.nrx
02/05/1998  12:34             5,267 oldcode.nrx
              31 File(s)        219,694 bytes

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

[ThSITC]

Hello Robert,

NetRexxC, as far as I know, is written in NetRexx!
Thomas.
=============================================================
Am 01.05.2011 19:40, schrieb Robert Hamilton:

Interesting, George.  Further speculation could be along these lines.  NetRexx/JAVA programs permeate IBM's business.  If they publish it They still have to main it internally so there would soon be two versions: IBM's and THAT OTHER COPY. So IBM will still have to maintain one copy no matter what.  They surely must have figured this out and it still doesn't explain any delay in publishing it... Just a real puzzle.

Another question: is NetRexxC written in C?

BobH
Richardson Texas  USA

On Sun, May 1, 2011 at 10:21 AM, George Hovey <[hidden email]> wrote:

Hello fellow sufferers,

In Samuel Beckett's Play "Waiting for Godot" two characters "wait endlessly and in vain for someone named Godot to arrive" (http://en.wikipedia.org/wiki/Waiting_for_Godot).  No doubt the reason for bringing this up is painfully clear: the protagonists talk endlessly, but never act; they just wait.

Is it fair to say that there is no objective evidence that IBM has made any progress toward open-sourcing NetRexx?  Even if they are moving at some glacial pace, NetRexx's prospects diminish with each passing year.

But there is a possible out: develop a "clean room" implementation of NetRexx.  This was the means by which IBM's control of the PC was broken many years ago.  Critical to such an effort is isolation of the developers from any knowledge of the internals of the program being cloned; I think that might mean that MFC and Rene would have to stay at arm's length.  The existence of a thorough published specification (TNL) would immeasurably aid the effort.

According to Wikipedia ("clean room design") software copyrights are not an obstacle, but patents are.  But even if IBM opposed such an effort it would be a positive development, as they would have some explaining to do.

RexxLA (as represented by Rene, Chip) sees themselves as the proper inheritors of NetRexx and have made the point that they have the talented manpower necessary to manage its development.  Further, they have said that that manpower will be made available at an appropriate time.  This is that time, in fact it's way overdue.  RexxLA do your stuff!

George Hovey


_______________________________________________
Ibm-netrexx mailing list
[hidden email]

_______________________________________________
Ibm-netrexx mailing list
[hidden email]

_______________________________________________
Ibm-netrexx mailing list
[hidden email]