Agreed - it doesn't seem harmful to include aliases such that people *can* call kibibyte and get the correct result, or possibly even move the actual implementation into those new methods and redefine the kilo/mega methods as aliases*...
However, changing the default behaviour of kilobyte and megabyte is going to a) drop-kick the principle of least surprise b) break stuff in potentially non-obvious ways By all means start the kibi revolution, but begin it with gentle re-education rather than sudden Numeric disruption. I suspect it may be a while before the world is ready to deal with tibibytes.... - james * or... write a plugin that defines the correct behaviour, and persuade people to love it. On 6/13/06, Jamis Buck <[EMAIL PROTECTED]> wrote:
My vote: changing Numeric#kilobyte at this point will result in breakage. Thus, we can't cause the existing kilo/mega/etc. methods to change. However, I think the unitless approach is worth investigating: Numeric#kilo #-> n*1_000 Numeric#mega #-> n*1_000_000 etc. - Jamis On Jun 13, 2006, at 12:43 PM, Johan Sorensen wrote: > > On 13 jun 2006, at 18.08, Dr. Ernie Prabhakar wrote: > >> Much as I admire the kibibyte concept, I have never seen *anyone* >> using this in real life. Ultimately, language is determined by >> usage, not standards, and I think we need to bow to public opinion >> in this case. > > Strongspace, a rails application, does: http://www.strongspace.com/ > faqs#what-is-a-gib > > I don't think we've ever heard anyone really complain about it, > other than "it's great you're using the proper units". > > JS > (my ~0.341 bytes) > > _______________________________________________ > Rails-core mailing list > Rails-core@lists.rubyonrails.org > http://lists.rubyonrails.org/mailman/listinfo/rails-core _______________________________________________ Rails-core mailing list Rails-core@lists.rubyonrails.org http://lists.rubyonrails.org/mailman/listinfo/rails-core
-- * J * ~ _______________________________________________ Rails-core mailing list Rails-core@lists.rubyonrails.org http://lists.rubyonrails.org/mailman/listinfo/rails-core
