Re: non-matched output port

/ Alessandro Vernet <avernet@orbeon.com> was heard to say:
| On 5/1/07, Alex Milowski <alex@milowski.org> wrote:
|> I suppose we could allow text descendants to be replaced (or any child) but
|> we couldn't allow things that aren't children.
|
| Yes, I see no reason not allow text nodes to be replaced or deleted.

Yes, I agree.

| More generally, the micro-components give me a bit of an uneasy
| feeling. Especially those that perform some "InfoSet CRUD operation".
| They seem to duplicate work that has been already done before.
| Consider for instance the XQuery Update Facility [1] or XForms actions
| like <xforms:delete> and <xforms:insert> [2].

Well, neither of those are applicable in this context. The whole
XQuery exercise was a complete duplication of XSLT with a funky
non-XML syntax, so there's certainly precedent.

| I am more familiar with the latter,  and can say that a lot of
| thinking has been going into those operations.

Can you bring some of that experience to bear here?

| I am afraid that not
| only we are duplicating what has already been done by other working
| groups, but that we lack the time to do this as thoroughly as other
| working groups did.

We can make the time if that's what it takes. On the other hand, most
of these steps hardly seem like rocket science. If you think we're
overlooking significant aspects of any of them, please tell us what
they are.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | Debugging is 99% complete most of the
http://nwalsh.com/            | time--Fred Brooks, jr.