Too much xml the base library?

ruby-talk
1

Hi gurus and nubys,

I understand that when XMLRPC4R or SOAP4R were written it made sense
to support as many xmlish parsers as possibile, but as they’re bundled
with REXML, now, what is the logic in still supporting all this
parsers?

Looking at XMLRPC it seem that it can use as a backend:
REXML
NQXML
XMLScan
XMLTree

IIRC SOAP4R supported nqxml too, but it seem is not there anymore,now.
IT still has support for rexml, xmlparser (?) and xmlscan.

Is there a reason to have all this xml backends?
I mean, is just that rexml is a fallback choice in case anyone of this
other parsers are not there ?

I believe, i.e., that xmlscan is faster than rexml.
But when I last heard of xmlscan the lightweight api of rexml was’nt
there yet…

2

gabriele renzi surrender_it@remove.yahoo.it wrote in message news:s7si509sn3i23dglqv5jvbqk4bp17racj1@4ax.com…

I understand that when XMLRPC4R or SOAP4R were written it made sense
to support as many xmlish parsers as possibile, but as they’re bundled
with REXML, now, what is the logic in still supporting all this
parsers?

Different strokes for different folks. If Michael and Hiroshi can
support multiple back-ends comfortably, then I think that’s great.

— SER

3

Hi,

From: “gabriele renzi” surrender_it@rc1.vip.ukl.yahoo.com
Newsgroups: comp.lang.ruby
Sent: Thursday, March 18, 2004 7:09 PM

IIRC SOAP4R supported nqxml too, but it seem is not there anymore,now.
IT still has support for rexml, xmlparser (?) and xmlscan.

Yes.

Is there a reason to have all this xml backends?
I mean, is just that rexml is a fallback choice in case anyone of this
other parsers are not there ?

As SER stated, every processors has own advantage.
Tons of features, memory footprint, speed, character encoding
support, API fravour, etc. etc. Users should choose his/her
processor.

…but to say the truth, I’m tired of supporting distinguishing
processors. Xmlscan will be the first class XML processor of
the next release of soap4r.

I wish if we had RAXP (JAXP for Ruby).

Regards,
// NaHi

4

Wrote “NAKAMURA, Hiroshi” nahi@keynauts.com, on Fri, Mar 19, 2004 at 11:37:49PM +0900:

…but to say the truth, I’m tired of supporting distinguishing
processors. Xmlscan will be the first class XML processor of
the next release of soap4r.

I wish if we had RAXP (JAXP for Ruby).

What about XOM4J? It uses an underlying stream parser to implement the
xml parsing, so should be able to be used to hide the underlying engine.

I really like that it was designed not by looking at one other API, and
thinking “I could do better”, but by looking at (and writing a book
about), a number of APIs, then thinking “I can do better”.

Cheers,
Sam

···

–
Sam Roberts sroberts@certicom.com