Re: Serialization Analysis and Proposal

On 5/16/07, Norman Walsh <ndw@nwalsh.com> wrote:
> / Alessandro Vernet <avernet@orbeon.com> was heard to say:
> | On 5/16/07, Norman Walsh <ndw@nwalsh.com> wrote:
> |> [...]
> |> And the p:store, p:http-request, etc. components would all treat an
> |> input document with a root element of c:serialized-form in a special
> |> way.
> |>
> |> Is that right?
> |
> | Yes, exactly. And the same system can be extended to support binaries
> | (XSL-FO producing PDF, SVG transformed in a PNG, processor packing a
> | number of XML documents into a zip, etc).
>
> It's an interesting trade-off. I like the fact that it puts the whole
> serialization package into a single step. But I'm a little bit
> uncomfortable with the fact that it means every step (and every
> extension step) has to examine the root element of the document(s) it
> receives and potentially behave differently for the c:serialized-form
> document(s).
>
> What do others think?

I am uncomfortable with the idea of passing binaries and pseudo-serialized
XML around in a pipeline.  Escaped XML is still a sequence of unicode
characters and there is no expectation by users that large chunks of a
document are going to handled efficiently.  That might not be true for
"serialization".

Besides, if I want to use something like p:http-request to send an XML
document to a web service, I'm better off having the serialization
happen in that component as I'll be writing to the open connection.  The
same is true of p:store.


-- 
--Alex Milowski
"The excellence of grammar as a guide is proportional to the paucity of the
inflexions, i.e. to the degree of analysis effected by the language
considered."

Bertrand Russell in a footnote of Principles of Mathematics

Received on Wednesday, 16 May 2007 17:30:10 UTC