return to first page linux journal archive
keywordscontents

The Last Word

e-Market e-Madness, e-Nough.

by Stan Kelly-Bootle

There are, they say, hard and soft sciences, not to be confused with difficult vs. easy, but more related to claims of objective, measurable, repeatable precision (we stout hardies) versus subjective, hand-waving, proofs-by-assertion (you big softies). The spectrum of fuzziness typically ranges from pure mathematics and physics at the diamond-tipped top, then descends via chemistry and biology (almost converging to synonyms), ending with a series of ``life'' and ``social'' sciences. The latter at least co-opt the intentions and vocabulary of the scientific method and rightly escape the ``beyond-the-soggy-pale'' category of pseudo-science (astrology, UFOlogy, pyramidiotology, hidden-Bible cryptology ... ad astra, ad nauseam). Honest ``theology'' wriggles through the colander by rejecting the ``scientific'' model.

Of course, science, life and society being what they are (send me a $1,000 check for this month's definitions), the hard/soft debate is doomed to waffle on. The paradox is that comparing the hardnesses of any two disciplines requires a valid metric--and that metric will depend on the hardness of the hardest involved domain. You can hear the magic predicate meta creeping into the equation. It reminds us that the very foundations of the purest of pure mathematics (formal set theory) were softened (nay, Osterized and Cuisinarted) by Gödel's meta-mathematical shocks in 1931.

It's still difficult to accept that the hardest queen of the sciences proved so brittle after millennia of complacency. Harder still to note that everyday mathematics and all the dependent sciences rumbled on, regardless. However, there was a clear dent in the traditional hierarchy.

We may avoid an arithmetical operator>> with a sort of ``diamond-scratches-glass'' ordering, but that moves us to unprovable, contentious, domain-dependent, metaphorical comparisons (e.g., in Political Science, we readily declare that Lincoln was a better president than Clinton).

There are other parameters available for comparing sciences. Some, such as ``usefulness'', are possibly more determinable and worthy of discussion. Thus, cosmology (already disputed in the hard/soft science ratings since we can't, as yet or ever, repeat the ``big bang'' experiment, do a tachyon glide into the local wormhole, or contact our nearest parallel universe) seems to offer useless theories and predictions: whether the cosmos expands forever or collapses after 10 billion years is, to most readers of Angela's Ashes, hardly even worth mentioning.

What of our provably most useful sciences: economics and computer science? Demi-soft economics has been dubbed the ``dismal'' science, while semi-hard computer science must surely be the most boring (most of its formal results, dated 1935, are proofs that ``this is impossible--don't bother!'') Yet hand-in-hand, both sciences have triggered the most undismal, unboring IPO stock-market episodes since the South Sea bubble. What is anything really worth? Forget the Marxian cosource-sweat-value-added axioms. If dazed bidders offer $1 million for a single Yahoo Japan share, then that is, by definition, its current ``worth''. Next question? Red Hat may or may not fairly distribute its IPO gains, but they missed out 500% by not calling themselves e-Red e-Hat e-Linux. And just wait until I launch e-skb-gnu-e-free-hardware.com.

Stan Kelly-Bootle (skb@crl.com) has been computing on and off since his EDSAC I (Cambridge University, UK) days in the 1950s. He has commented on the unchanging DP scene in many columns (``More than the effin' Parthenon''--Meilir Page-Jones) and books, including The Computer Contradictionary (MIT Press) and UNIX Complete (Sybex).