What are the current tools used to work on XML documents? I would like to focus on problematics concerning text, text mining, linguistics. Is the use of programming language (Perl : XML: Parser, etc.) mandatory, or is there any GUI-software to work on XML?
I have been Oxygen as a main tool for XML for a long time now, if you have a little time to learn XSLT there isn’t much that you can’t do with your XML data. I was able to do things with unstructured data using XSLT that developers claimed couldn’t be done in “real languages”.
There is a big archive of problems and solutions as well as a lot of people that are willing to help when you are stuck on this list.
http://www.mulberrytech.com/xsl/xsl-list/
While SAX and DOM are the base, often you need something between the two. SAX can handle enormous documents, but is totally event-driven, so finding values within a structure requires significant work. DOM makes finding values easy, but requires the whole lot to be read into memory, which makes large documents hard to handle. XSLT depends on DOM, so again its great, but cannot handle really big documents.
If you need very large documents, I've used the Perl module XML::Twig, which allows event-based handling to a level, and still provides DOM-like functions within that level. I like it, but it isn't fast. LibXML has a pull parser you can use for the same kind of purpose, it's similar to StAX in Java, and XmlReader in C#). These are invaluable if you are handling large XML files, too large to fit (as a DOM structure) in memory, and yet more complex than you can handle with SAX. LibXML's Perl interface, because it is C underneath, is actually very fast.
To generate XML, I usually use the small and nice XML::Writer in Perl. I've also used JAXB in Java. which maps objects to XML, and is a very nice approach to serializing data in XML. JAXB can also read, but I tend not to use it for that, as it's got the same problems as DOM in memory handling.
Under the topic(s) of Database(s) the SAX and DOM seem to be quite low level technologies, and more proper technology to discuss is provided by the XQuery extension of ISO SQL standard, which has been adapted from W3C and implemented for example in DB2 versions 9.5 and later, and some other DBMS products. In addition to queries, XQuery can also be used for updating XML documents. The academic/research people should get interested in the DB2 Express-C edition, which is free and available on the major platforms.
It depends on many factors, like what you want to do and do you have preferable or mandatory choices of development environments and tools etc.
For programmatic access I would strongly recommend to consider XQuery, which is very powerful and optimal for querying, analyzing and fiddling with XML data. You can find a hopefully useful introduction to XQuery at http://dx.doi.org/10.1002/spe.1140 (even though written from a peculiar point-of-view of applying it to problem solving).
XSLT is in principle equally powerful, but in practice the language of choice only for transformations, that is, for mapping entire documents (or large portions of them) into another form like HTML.
If you need to take resource constraints or efficiency into account, say, if your documents are big or HUGE, then SAX or DOM or some of their variants may be relevant to you, but as a starting point for experimentation I would strongly recommend XQuery.
About GUIs for manipulating XML I know hardly anything - sorry for that.
Pekka Kilpeläinen
DOM and SAX are two different mechanism to handle XML documents. DOM is based on tree model and SAX based on event model. XML is having a wide application areas including data representation. People from different areas are using XML and created their own markup language like MathML, CML, SBML, RSS, etc. For handling these XML files we need parsers. Java provides a rich API for his purpose. Lots of GUI based parsers are already including MSXML.
- Badges
- Science topic
- Similar topics
- Linguistics