Help in picking a parser

Sam Gentile (samg@fundtech.com)
Wed, 5 Aug 1998 10:46:24 -0400

I am working on a E-Commerce Banking system and we have finally convinced
people to use XML to transport our data from the browser front-end to the
Web Server and out through our TUXEDO application servers. There is a need
for XM parsers in several locations. On the browser side, we have DHTML and
Microsoft's Visual J++ WFC "Code Behind DHTML" Java classes. I thought there
was a WFC class to do parsing but it turns out that there is a
com.ms.xml.parser in the Java SDK. So that side is taken care of.

Our problem is more on the Web Server side. On the client, we are streaming
the Java classes into an XML buffer to go to the Web server. The Web server
will need to parse this XML and use a factory to dynamically instantiate and
re-create the Java class hierarchy.
Then the data will be put in TUXEDO FML format for shipment to the Tuxedo
servers. On the way back, I will have to take an FML buffer and convert to
an XML buffer.
They say they don't want to use DTDs and validating parsers and that we may
need the parser to be cross-platform. So I have been looking at
Nonvalidating parsers like Lark and XP. I am new to XML. It seems like these
parsers are a maze of source code. Will I be better off having my code
"spawn" an existing parser instead of integrating with this source code? Can
Validating parsers be used even when DTDs are not used (like Microsoft's and
IBM's)?

Any suggestions and thoughts would be greatly appreciated.

Sam Gentile
Senior Software Engineer
Fundtech Corporation
samg@fundtech.com <mailto:samg@fundtech.com>
Ê