Re: can we have last() having a consistent value ?

/ Innovimax SARL <innovimax@gmail.com> was heard to say:
| Jeni,
|
| You just make the point, I was fearing
|
| Since we have position()....why not having last() !!
|
| And that would be the question every user would do
|
| And I don't want to pave the way to our own destruction...

I think all of the following proposals are entirely rational and none
of them lead to destruction:

1. Set last() = position() and explain to users that that's the way
   it works in V1.

2. Set last() = MAXINT and explain to users that that's the way
   it works in V1.

3. If the pipeline author uses last() then the step has to buffer
   and give the right answer. If that causes a resource error in the
   implementation (out of memory, for example), so be it.

I would prefer, in this order, 1, 2, or 3. I feel a little uneasy
about 2 for no concrete reason and, speaking as an implementor relying
on a third party library to evaluate XPath expressions, I'd rather not
have to parse them myself to implement 3. But I will if that's what
the WG decides, of course.

| Here is my proposal
|
| Since making context having more sense than just (size=1/position=1)
| imply most immediately that last() has a sense

I just don't understand how making size=1/position=1 is any more
sensible than any of 1, 2, or 3 above.

| I propose
|
| to make a special function p:position() with the semantic we have said
| and make for V.next the fixing of the context if we think it will be sensible

I don't think there's any conceptual complexity in fixing the context.
It's just computation expensive in terms of time and and space to do.
Hence my prefernece for 1.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | If someone tells you he is going to
http://nwalsh.com/            | make 'a realistic decision', you
                              | immediately understand that he has
                              | resolved to do something bad.--Mary
                              | McCarthy

Received on Saturday, 26 May 2007 13:33:50 UTC