Print

Print


 From: Saašha Metsärantala
> Sent: Monday, May 23, 2011 1:31 PM


> I suggest to remove the parentheses around "later"
> 
> > (listElement",")* (later)
> (listElement",")* later

Ok.

 
> I also wonder whether there are some reasons for not having longYear in
> the lists. I suggest to add them (read below).

I'm having trouble imagining a use case here, maybe I'm not imaginitive enough, but I think of, for example, a list of publication years:
{1667,1668, 1670..1672}

but a list like:

{1667,1668, 1670..1672, 17e7}

What might that be a list of?




> 
> There are some differences between the BNF's lists and the lists as of
> EDTF spec.
> 
> The BNF accepts lists with mixed consecutives, such as
> 
> [2011-05..2012]

By mixed, you mean one is a year-month and the second is a year.  Yes, this example of a consecutive makes no sense. 

I just wonder if it is worth complicating the BNF so that such a production is disallowed.  I think it we examine the BNF closely enough we will find quite a few productions that the BNF allows but do not make sense.  A perfect BNF ("perfect" in the sense that it allows only productions that are meaningful) would probably be so complex that nobody could read it, and I don't know if anyone could write it, and if they could, it would surely have errors anyway.



>I would suggest to just discard single-element
> lists in the BNF.

Ok. I changed listContent  

from:

earlier
| (earlier “,”)? listElement ("," listElement)*
| (earlier ",")? (listElement",")* later

to:

 earlier ("," listElement)+ (,later)?
|(earlier ",")? (listElement",")+ later
| listElement ("," listElement)+

--Ray