W3C home > Mailing lists > Public > public-rif-wg@w3.org > April 2009

editorial questions/thoughts about DTB

From: Sandro Hawke <sandro@w3.org>
Date: Wed, 22 Apr 2009 11:05:28 -0400
To: Axel Polleres <axel.polleres@deri.org>
Cc: public-rif-wg@w3.org
Message-ID: <32270.1240412728@ubehebe>

I'm starting to work on the list built-ins (ACTION-760), so I'm looking
at DTB more closely.  None of these comments are terribly important, but
I wanted to voice them in case you or others feel similarly, and/or so I
can learn why things are the way they are.

   * names like pred:isNotXMLLiteral will turn into is-not-XMLLiteral,
     right?  (I think we agreed to use hypens, except within datatype
     names, although I can't find a record of this decision.)

   * Isn't the form of schemas annoyingly redundant?  Instead of

       ( ?foo ?bar; func ( ?foo ?bar ) )

     why not just write

       func(?foo ?bar)

     Actually, I'd probably drop the question marks and use italics.
    
     I understand this form comes from FLD, but I expect many readers of
     DTB will never have any reason to read FLD, so it seems like an
     inappropriate way to burden them.

   * Why does pred:numeric-equal have a long definition and the other
     numeric predicates don't?  Is there some subtlety to how it's
     different from xpath?  Also the definition uses s1,s2 in one place
     and a1,a2, in another.  In general, I don't really understand why
     in some cases a formal mapping is given and in others it is not.
     The formal mappings also always seem trivial -- it's hard for me to
     see what value they provide.

   * Typo:  pred:yearMonthDuration--greater-than

   * Typo:  Subtracts two xs:times. Returns an xs:xs:dayTimeDuration. 
     (there are several other xs:xs: items)

   * Is there any way to get a string-join using a rif:List ?  Having
     lists, the current string-join is kind of silly.

   * Are you using any sort of tool to help with the formatting?  I'm
     inclined to use a wiki template (macro).  (I'm not a huge fan of
     the current <li> formatting, and it'd be easy to experiment, given
     templates.)  

     Actually, I'm inclined to generate this whole document out of a
     database, given its great regularity.....

   * I wonder, if we have test cases for each of these some day, whether
     to include them in-line here as examples (possibly with a hide-show
     javacript control).  It might make sense to borrow and adapt on the
     examples in xpath-functions.

That's it...

     -- Sandro
Received on Wednesday, 22 April 2009 15:05:36 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:34:05 GMT