I am gleaning that from personal testing. That test script is very system-dependent, unfortunately, so not meaningfully distributed.<div><br></div><div>But to clarify a bit, I tested with the ExpatXS module and using the M::R::new_from_xml() method. Creating the new object from an XML record with 1,000 embedded 952 tags took half the amount of time in ExpatXS as compared to LibXML::Parser.<div>

<br></div><div>Clay</div><div><br><br><div class="gmail_quote">On Thu, Nov 11, 2010 at 4:17 PM, Thomas Krichel <span dir="ltr">&lt;<a href="mailto:krichel@openlib.org">krichel@openlib.org</a>&gt;</span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">

  Clay Fouts writes<br>
<div class="im"><br>
&gt; Expat is considerably faster than the LibXML implementation,<br>
<br>
</div>  What is your source for this information? A quick Google<br>
  search seems to suggest that Expat is a bit faster, but<br>
  not much.<br>
<br>
  Cheers,<br>
<font color="#888888"><br>
  Thomas Krichel                    <a href="http://openlib.org/home/krichel" target="_blank">http://openlib.org/home/krichel</a><br>
                                <a href="http://authorclaim.org/profile/pkr1" target="_blank">http://authorclaim.org/profile/pkr1</a><br>
                                               skype: thomaskrichel<br>
</font></blockquote></div><br></div></div>