Re: [Factor-talk] restricting ratio literals

2016-10-26 Thread Jon Harper
I think there is value in having such in important word (string>number) have a documented behavior. Right now it's hard to document all the things that are correctly parsed by this word... Jon On Wed, Oct 26, 2016 at 9:40 AM, Björn Lindqvist wrote: > Imho, choose based on

Re: [Factor-talk] restricting ratio literals

2016-10-26 Thread Björn Lindqvist
Imho, choose based on what makes the implementation better. If the parser can be written in a simpler (and probably faster) way by dropping support for ratios in uncommon bases, then let's do it. Tbh, I don't think any ratios other than those on the format int1/int2 needs to be supported.

Re: [Factor-talk] restricting ratio literals

2016-10-25 Thread Jon Harper
I would say no prefix at all for ratios, only base 10. Also do people in the US consider 1+13/2 a number? Or is it really only written as 7+1/2? I think the number literals should only express what people consider as numbers. They should not replace arbitrary mathematical operations (like your

Re: [Factor-talk] restricting ratio literals

2016-10-24 Thread John Benediktsson
Those examples do look ugly -- maybe base 10 ratios only is not a bad idea. But would it support the 0d prefix? e.g. ``0d1+1/3``? On Mon, Oct 24, 2016 at 10:23 AM, Jon Harper wrote: > Hi list, > I'm reading the questions at