My name is Rudolf de Grijs and I do currently work for Software AG Netherlands.
In my day to day work I deal a lot with XML. XML2Sequential (or Seq2XML, I do not care) has been developed to parse sequential files with a fixed lay-out. I already have used this package to convert mainframe workfiles (VSAM, Uncompressed ADABAS files, workfiles..) to XML. Afterwards the XML file is transformed (XSLT) and stored into Tamino (XML database of Software AG. The current release, version 4.1, is a real leap forward ;-).
At the moment I'm working on a XML Pull Parser. I do hope that I can develop a very fast (non-validating) XML parser. Up till now I barely manage to get the same outcome as piccolo SAX parser. If the file gets bigger (> 100K), the result gets worse. I did notice that the other pull parsers suffer from this same phenomena. Perhaps it's a good idea to use JFlex and BYACC/J (or ANTLR) to build the XML pull parser, just like has been done for Piccolo.. Now I have written the parser myself.