On Wednesday, 10 December 2014 at 15:53:59 UTC, H. S. Teoh via
Digitalmars-d wrote:
I find the obsession with small integers (aka version numbers)
rather
petty. We should start with some random number, like 49183029,
What about assigning a prime number to each semantic concept in
the language, then calculate the product to get the unique
version number? If all languages used the same scheme then you
could just factorize the version number to figure out what
features a given language-version supports.
Another alternatives is just use something more descriptive for
versioning, like obscure movie titles:
D:
http://www.imdb.com/title/tt0287205/?ref_=fn_al_tt_2
D-Train:
http://www.imdb.com/title/tt3534602/?ref_=fn_al_tt_5
Tenacious D in The Pick of Destiny:
http://www.imdb.com/title/tt0365830/?ref_=fn_al_tt_3
Or perhaps the initial version should be Graham's number, and
every
following version is obtained by calling the Ackermann function
on the
previous version number. :-P That'll beat all competitors, for
sure.
Yes, it is well known that a product does not qualify as mature
until the storage requirements for the version number exceeds the
requirements for the executable. High resolution versioning is a
tremendous benefit for anyone doing modern agile iterative
development.
