[Insight-developers] VXL : Toms library : ACM Copyright : non-commercial license

[Luis Ibanez]

On Sun, Jan 31, 2010 at 12:29 PM, Steve M. Robbins <steve at sumost.ca> wrote:
> On Sat, Jan 30, 2010 at 05:35:26PM -0500, Luis Ibanez wrote:
>
>> On Sat, Jan 30, 2010 at 4:36 PM, Steve M. Robbins <steve at sumost.ca> wrote:
>
>> > However, I'm curious about this remark:
>> >
>> >> The fact that we are still carrying around
>> >> FORTRAN code from the 80's is an obvious
>> >> symptom of technological decadence.
>> >
>> > Without checking, I would speculate that it may be that the problem
>> > was well enough understood in the 80's that it isn't worth rewriting
>> > the code each decade.
>> >
>> > What is the "obvious technological decadence" you perceive?
>> >
>> ------------------------------------------------------------------
>>
>>
>> I wish I could say that I came up with this, but the truth
>> is that I'm just quoting & copying from Isaac Asimov's
>> remarks in The Foundation Trilogy.
>>
>>
>> Two symptoms of technological decadence are:
>>
>>
>>     1)   When the technicians of today no longer know
>>           how is that devices of years past work.
>>
>>
>>     2)   When we no longer know how to build those
>>          devices or not even know how to repair them.
>>
>>
>>   and... we still use those devices,
>>             and rely on them.
>
> [ ... elaboration deleted ... ]
>
> So you're argument is that we should not rely on code that is
> unmaintained and no longer understood.  Clearly true.
>
> Your line of reasoning wasn't apparent to me reading your initial
> remark.
>
>
> Whether the code in question is properly maintained, understood, or
> the subject of Religious Zealotry, I have no real idea.  But I would
> observe that:
>
>  1. ITK itself provides a CVS repository, bugtracker, mailing list, etc.
>  2. The parts used in ITK *are* tested (to the extent ITK itself is tested).
>
> so the use of this code in ITK may not be as bleak as you portray it.
> Or it may be; I haven't looked at the code, so I don't really know.
>
--------------------------------------------------------------------

Granted,
ITK tests some fraction of that code....

Note however that we were able to fully remove
the "toms" code with zero impact to ITK.

This clearly indicates that ITK has never exercised
that code,... and therefore, that part of the code was
never tested as part of ITK.

Note that even in the cases where we use pieces
of the code, (i.e. lsqr), we do not maintain that code.


     ITK shoudn't be the Dashboard of Netlib.


nor the ITK mailing list is going to provide the necessary
support for the sparse solvers that are distributed with
in netlib.


    Software that doesn't have a community around it
    is DEAD software.


--------------------------------------------------------
>
> Have a nice day,  :-)
> -Steve
>
> P.S. Thanks for the reminder of The Foundation series; it's been a
> while since I read it.  I just may dig it out again ...
>
>
--------

Great books indeed...


    Regards,


           Luis