Tables

[John Gruber]

Michel Fortin <michel.fortin at michelf.com> wrote on 11/14/05 at 6:52 PM:


> My way to think about it is that by relying on spacing for alignement  

> you increase the risk of unexpected results. It's a nice idea, but we  

> must be careful with it. For example, this overspaced table could be  

> a problem:

> 

>      |  First     |   Second    |

>      | ---------- | ----------- |

>      |    1.0     |      1.2    |

>      |   23.2     |     24.1    |

> 

> I'm not sure you could find a *reliable* algorithm to find the  

> expected alignement here (assuming it could have been written in a  

> variable-width font) because you can't just count the spaces on the  

> left and on the right and say we should align cells on the side where  

> there is more space.


I'm thinking that rather than counting any arbitrary number of
spaces, and looking for one side of a column that uses the same
number of spaces for each cell, that the rule would be that you'd
have to use 0 or 1 spaces on the right side if you want
right-alignment.

-J.G.