Re: BLD: two issues with the BNF

>>>> I uncovered two issues with the BNF in BLD.  For me they are not 
>>>> critical for last call, but for some they might be.
>>> Jos,
>>> thanks for uncovering these inconsistencies.
>>>
>>>> The first issue is an error: Profile is not a Unicode string in the 
>>>> presentation syntax, but it is a term (see section 2.4).
>>> Yes, this is how it is defined in the math syntax, but EBNF is not consistent
>>> with this. But looking at your SWC, your profiles are not terms either. They
>>> are not even constants! 
>> Actually, they are.  But I'm using the shortcut syntax for absolute IRIs 
>> using the <> delimiters.
> 
> In the table http://www.w3.org/2005/rules/wiki/SWC#Profiles_of_Imports
> you are not using constants, but unicode strings.
> We'll change the ebnf to make it consistent, but you should also fix your table.

I think the table is okay.  It just lists a bunch of IRIs, but does not 
tell you how to write them down.
Section 5.2 talks about how they're written down, and there you can see 
that the limiters are used.

By the way, another thing I noticed with the grammar is the definition 
of IRICONST: it allows writing IRIs in just one way, mandating the use 
of a complex URI for the symbol space with one particular prefix.  I 
think it should be more liberal.  I guess it should be okay to just 
refer to constants there, since the grammar represents a superset of the 
presentation syntax anyway.

Best, Jos

> 
> michael

-- 
                          debruijn@inf.unibz.it

Jos de Bruijn,        http://www.debruijn.net/
----------------------------------------------
One man that has a mind and knows it can
always beat ten men who haven't and don't.
   -- George Bernard Shaw