[plug] Bitrot

[Shayne O'Neill]

Its a bizare argument. If its FOSS, make sure a copy of the sourcecode for 
the original software is saved.

I actually see the reverse of this problem quite a bit with accounting 
software. Companies that had MS-DOS based accounting systems, STILL using 
MS-DOS and the old accountng system, because the latest XP based ones 
won't read the old data. If they where FOSS, and granted folks didn't 
really know as much about FOSS back then, they could simply recompile the 
accounting software for the new permutation of the OSS.

Fortunately most of the old stuff still works on freedos (And under 
linux!) but it'd be nice to have some of that old discontinued stuff FOSS 
so coders could code a fancy new gui and update it to the times.

Regardless, FOSS or not, scientific data just needs to be stored in a 
documented format (well documented XML is plenty fine for smaller 
datasets. For bigger datasets, just use either an open spec, or *make* an 
open spec if your data type is unique and document it well. But more to 
the point, keep that source code.

Shayne