Re: commit -- fixes for endian-ness w/r/t ispell hash files.

michael@surfnetcity.com.au
Sat, 10 Jul 1999 14:49:25 +1000


On Fri, Jul 09, 1999 at 03:52:22PM -0500, Jeff Hostetler wrote:

> i've added big- and little-endian versions of hash files for
> ispell for american.hash and updated the 'canonical' step of
> abi/src/Makefile to expand the proper one.

I know this has been talked about exaustively, but I honestly think that the
ispell hash should be endian-indepenant when stored on disk.

What I'm thinking, is the possibility of:

a) A networked environment, where everyone shares a dictionary. (This is
especially so, if everyone gets sick of some in-house terms being a spelling
mistake, and wishes for that word to be added to all dictionaries)

b) The possibility of adding/modifying words in the dictionary, and
distributing the changes to friends/workmates.

c) To avoid confusion, when somebody gets a corrupt dictionary, and copies
their friends one to fix it

d) To save space on the distribution CD

Now, I'd say we should have it in big endian byte order, and x86 should have to
pay the penalty for being so backwards ;-)

(Let the flamefest begin :)

-- 
-- Michael Samuel <michael@surfnetcity.com.au>


This archive was generated by hypermail 1.03b2.