W3C home > Mailing lists > Public > public-webapps@w3.org > April to June 2009

Re: [WebIDL] double

From: Ian Hickson <ian@hixie.ch>
Date: Mon, 15 Jun 2009 00:08:10 +0000 (UTC)
To: Cameron McCormack <cam@mcc.id.au>
Cc: Anne van Kesteren <annevk@opera.com>, Marcin Hanclik <Marcin.Hanclik@access-company.com>, Andrei Popescu <andreip@google.com>, public-webapps@w3.org, public-geolocation@w3.org
Message-ID: <Pine.LNX.4.62.0906150006450.16244@hixie.dreamhostps.com>
On Mon, 15 Jun 2009, Cameron McCormack wrote:
> Ian Hickson:
> > How do I know when to use float or double, for types that are supposed to 
> > represent infinite range? Should I just make everything double?
> 
> If you want more range, sure, use double.  Web IDL doesn’t have an 
> arbitrary precision numeric type.  But double specifically means 64-bit 
> IEEE 754 double precision floating point, here, which is just what JS 
> Number is (except for the NaNs).

Oh. Wouldn't it make more sense for JS compatiblility to have float be 
64-bit? It seems bad to have a type that is 32-bit but then have its most 
common implementation not throw an exception or anything if setting a 
number that is 64-bit into it. (Or do UAs treat 'float's as 32-bit and 
trim the numbers when storing them?)

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 15 June 2009 00:08:46 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:49:31 GMT