Granted that in the realm of scientific software, professional packages such as Maple and Mathematica (or even Matlab), which are proprietary, “cathedral-built” software, have the edge over “bazaar-built” FOSS packages such as Octave, Scilab, Python, R or Mayavi. R is a recent and very strong open source contender to commercial statistical packages, with development supported by many statisticians and open source coders; but still lags behind, as some claim, to its commercial rivals.
I also have to agree that open source programmers “... are not obliged to follow any rigid code of compatibility.” No, they are not. They are not obliged to follow prescriptions laid by proprietary software. And why so? The great Ghalib wrote a couplet that could explain:
Hain aur bhi duniya mein sukhanwar bahut achhe
Kahte hain ki Ghalib ka hai andaz-e-bayan aur
(Translated by Ralph Russell as:
The world holds other poets too who write good poetry
But Ghalib’s way of saying things, they say, is something else).
FOSS programmers too want to say things differently. They are not dictated by commands of the industry, or specific users. They may take a zameen and build upon it, or fork to an entirely different path. But, they do follow standards - it is well-known that open data standards are, most prominently, supported by the FOSS community. Open standards allow programs to be compatible and data to be exchanged freely. However, and inevitably, FOSS programmers will follow their own andaz-e-bayan .
(Rahul D é is Hewlett-Packard Chair Professor of Information Systems at IIM, Bangalore.)