XML Modules Comparisons

[toadi]

Hello,

I'm looking for some advice on differen XML modules out there.

What do I need to do?
I have to parse XML and write XML.
I looked into XML::Simple and XML::Parser. But then my boss came with another one to use: Xerces. This is some apache XML thingie...

Now my question is: Has some experience with this apache module. Is this better then the others? Are there advantages or disadvantages I need to know about it? Most documented modules are the Simple and the parser...




--
My opinions may have changed,
but not the fact that I am right

Update - Masem 2001-07-09 Edited title from "XML Modules an"

[Masem]

In both the case of XML::Simple and XML::Parser, these only deal with the input of XML into your program; you'd still need another module that deals with the output such as XML::AutoWriter; thus your input and output methods may seem very disjointed. While I've not used Xerces, it easy to see that it handles both sides of XML (parsing and generating). The only thing that is of concern is that it is a perl module that sits on top of a C/C++ library; thus, some of the function calling seems to get hairy, and there might be some performance penalties. Xerces appears to be very well documented by the Apache team (I've yet to see anything Apache puts out that's poorly documented), so you'll find just as much info about Xerces as you would with XML::Simple or ::Parser.

I'd say the best thing to do is to try to write some quick sample application code using both, and decide which one feels 'best' to work with, with all other factors being equal; consider expandibility, code maintainence, and how others may seem the code as part of your decision.

---

Dr. Michael K. Neylon - [e-mail] || "You've left the lens cap of your mind on again, Pinky" - The Brain

[clintp]

Small correction: XML::Simple does have an output method (XMLout) that will take a perl structure and produce XML for it.

[pmas]

Masem said it all.

I will consider also to compare support for arr considered modules (mail lists, if any), try to post a question if any arises during pilot test, and compare response.

If module has strong following, it is likely it will stay around for a while. SO if you invest time to develop any module, you want the module with staying power.

pmas

To make errors is human. But to make million errors per second, you need a computer.

[Coyote]

In addition to the modules suggested by Masem, you might want to check out mirod's XML::Twig module which is well documented at http://www.xmltwig.com, Matt Sargent's XML::XPath and the new XML::LibXML module. One to stay away from (IMHO) is XML::DOM which is slow, awkward to use, and doesn't seem to be actively maintained anymore. You also may want to check out the archives of the Perl-XML lists at http://aspn.activestate.com/ASPN/Mail/Browse/Threaded/perl-xml, Kip Hampton's articles on Perl and XML at http://www.xml.com, and the Reviews section here at the monestary. If you are interested in Perl-XML application servers, check out these nodes.

The articles at xml.com may be of particular interest to you as they recently ran a series of intro articles on the perl xml modules.

----
Coyote

[thraxil]

if close integration with Apache is important, don't overlook axkit or cocoon (warning: cocoon isn't perl.) The built in caching that they both do is nice for performance.

anders pearson

[aardvark]

Xerces / Xalan / FOP all have very nice command line features, which is great for somebody who hates to write java code but wants to take advantage of all the great work of the xml.apache.org project.

My most used feature of Xerces is a quick command to check the syntax and validate XML. Very handy during development, especially when your writing XML with vi.

java sax.SAXCount -v file_to_validate.xml

Documentation for Xerces Command-line utility
Documentation for Xalan Command-line utility
Documentation for FOP Command-line utility

Get Strong Together!!

[dingbat]

I'd definitely recommend one of the "mainstream" XML tools (for DOM/SAX/XSLT) over another Perl module, even if this involves a Horrible Hack to switch from the Perl environment to the Java world (a "used coffee filter", or whatever they call it) that the DOM was originally coded for. Personally I'm writing XML/XSLT code for M$oft; IE5 and ASP, so I choose to use the MXSML DOM via Win32::OLE even when I'm coding Perl.

DOMs aren't standardised; there's a lot of space around the W3C spec that is implementation dependent (especially with namespaces). Using something commonplace is good, and Xerces would be a good choice.