[Beowulf] g77 limits...

Robert G. Brown rgb at phy.duke.edu
Fri Feb 24 00:38:53 PST 2006


On Thu, 23 Feb 2006, Greg Lindahl wrote:

> On Fri, Feb 24, 2006 at 01:06:49AM -0500, Robert G. Brown wrote:
>
>> Physicists at least are really, really slow to port their legacy
>> sources.  You must have encountered cernlib by this point in time -- one
>> of my few recent brushes with fortran at all was trying to get cernlib
>> to actually build into an rpm.  If you've looked through it and haven't
>> shuddered, well, you're just shudder-proof.
>
> I think cernlib qualifies as the most, erm, inventive code I've ever
> seen. Invective and insult were also involved. Eventually, after a
> liberal application of beer, I washed away the memories.

I haven't yet consumed enough beer to accomplish this (he says looking
up from the gutter:-).

> And "they" say open source code is good code. "They" didn't look at
> the counter-examples.

Some open source code IS good code, especially if it undergoes a truly
open development process with some degree of evolutionary feedback.

For perfectly understandable reasons (generations of untrained physics
graduate student slave labor writing it, the original code base written
in a perfectly awful language, nobody willing to touch it to maintain it
ever after) a lot of physics code, especially legacy fortran code
untouched by whole generations of users except to recompile and link to
it, is not.

cernlib just needs to be put out of its misery, and probably be replaced
by the GSL and code derived from the GSL.  From the looks I took under
the hood, that wouldn't, actually, be THAT hard.  Although it would be a
lot easier to just rewrite the code that uses it to use something else.

But not by me, not unless somebody pays me a LOT of money.  Enough to
keep me likkered up enough to stand the dual pain of cernlib and fortran
in general;-)

And alas, cernlib is far from the only such package out there.

    rgb

-- 
Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu





More information about the Beowulf mailing list