RE: Nested sequences in XPath

I think that very few use cases have come up over the last couple of years
that require nested sequences, which justifies the original decision not to
provide them. I also think it would be quite confusing to have two separate
mechanisms for managing trees in XPath, one based on XML and the other based
on nested sequences, but supporting a completely different set of
constraints and operators.

If we do ever want to add nested sequences, I don't think it's difficult; we
would just add a new kind of item called a sequence-reference.

It's certainly too late to contemplate this kind of change at this stage of
the proceedings, without a much better justification than an unwarm heart.

Michael Kay

> -----Original Message-----
> From: Paul Prescod [mailto:paul@prescod.net] 
> Sent: 07 April 2003 01:38
> To: public-qt-comments@w3.org
> Cc: evan@evanlenz.net; michael.h.kay@ntlworld.com
> Subject: Nested sequences in XPath
> 
> 
> 
> XPath 2 does not allow nested sequences. I believe that this is a 
> mistake and one which will be difficult to correct in the future. I 
> don't know how widely this has been discussed, but I found this:
> 
> http://www.biglist.com/lists/xsl-list/archives/200201/msg01723.html
> 
> It didn't exactly warm my heart. Nobody could justify the current 
> situation as anything more than "the way things evolved". "I 
> accept that 
> the above isn't a conclusive argument that we've done the 
> right thing; I 
> present it more as a post-hoc explanation of how the language 
> came to be 
> like it is."
> 
> It seems to me that this is the same sort of "simplifying" arbitrary 
> limitation that left real sequences out of XPath 1. Once you've said 
> that nested sequences are flattened it will be VERY hard to 
> fix in XPath 
> 3! I can't off the top of my head think of a relatively clean way to 
> evolve from the current model and syntax to one that allows 
> true nested 
> sequences. So let's get this right the first time!
> 
> The specification itself has the following example:
> 
> "The following example transforms the input document into a list in 
> which each author's name appears only once, followed by a 
> list of titles 
> of books written by that author."
> 
> for $a in distinct-values(//author)
> return ($a,
>          for $b in //book[author = $a]
>          return $b/title)
> 
> It seems much more likely that you'd like a sequence of pairs. 
> Otherwise, how do you differentiate the authors' names fro the book 
> titles in the sequence that is returned? There are very few 
> circumstances under which you would just not care about the 
> distinction. 
>   I can already imagine myself gritting my teeth working around these 
> sorts of situations.
> 
> Furhermore, it makes the language harder to learn (and perhaps 
> implement!), not easier, because people are used to nested sequences 
> from every other sequence manipulation language going back to the 
> fifties (with the possible exception of Perl 4).
> 
>   Paul Prescod
> 
>

Received on Monday, 7 April 2003 15:35:11 UTC