[lug] Entropy and OSS.

[Rob Nagler]

<pontification>
We often overrate survivors.  Like Jeffrey says, the large tail on a
peacock may or may not have anything to do with survival of peacocks
as a species.  Explanations are not facts.  We have no idea why OSS
survives, or, for that matter, if it does as a "species".  We know
that emacs has been around for decades, but so as APL, MS-DOS, and so
on.  Is an Atari 2600 so cute and fuzzy that makes people save them
and write emulators for them?

Suvivorship bias applies heavily in human phenomena such as software.
DOS is a survivor, for sure, but it's pretty tough to say why exactly
it beat out CPM, and, did that have anything to do with DOS's
survivability, or more to do with Microsoft's survivability.  I don't
think the answer is so easy. Is Warren Buffet the world's greatest
investor, or was he the least likely to go bankrupt?  Madoff may not
have been an investor at all, according to recent reports, but he
certainly survived.  Indeed, if he really wasn't trading at all, and
instead he was holding money in the bank these last couple of years,
his strategy was better than Warren Buffet's, in the short term.

It's really tough to be unbiased when looking at survivors.  They are
the ones left standing, but there aren't enough of them to say that a
"species" is surviving.  Evolution is the law of large numbers.  Is
Google really the best search engine out there, or was a quirk of
Google's particular existence that made it into the behemoth it is
today?

There particular idioms that are survivors as a species in software,
for example, the hash (associative array, lookup table, dictionaries,
etc.).  It comes up over and over again in software.  Other
evolutionary survivors are integers, floating point numbers, and
lists.  BCD is not a survivor.  It was good in its ecological niche,
but it could not survive between bignums and integers as memory and
processors got faster.

I think a lot about survivors, but I try to make sure I'm in the realm
of evolutionary space and time.  It's not a lot of fun, which is why
we hear about stuff like OSS survivors.  Humans love stories, and
individual survivors make great ones.
</pontification>

I highly recommend The Black Swan by Nassim Nicholas Taleb.  He
explains how survivorship affects our perception of risk.  For more
info, you can read my review here:
http://www.viarob.com/my/page/BookReview_The_Black_Swan

Rob