My article “XML ohne spitze Klammern” about Vex has appeared in the German Eclipse Magazin. It will also be published at JAXenter.de at a later date. The article includes a getting started tutorial and short interviews with John Krasnay and David Carver. Read below the original interviews taken at the end last year.
John Krasnay, initial author of Vex
Eclipse Magazin: Hi John, you are the main author of Vex. What was the reason to start Vex in the year 2002 and to publish it as Open Source?
By day I’m a contract Java developer, and most of my clients prefer their documentation do be done in Microsoft Word. I find it takes great discipline to create good looking documents in Word, and it’s almost impossible to publish a decent Web site from Word files, so I was looking around for some better way to produce documentation. XML (particularly DocBook) plus XSLT looked like a good alternative. For example, given a DocBook XML file I could use one set of XSLT files to generate XSL:FO for printed documentation and another to produce clean HTML. But of course it’s very difficult to write raw XML, and the XML editors that were available all seemed to treat XML as a hierarchical data structure rather than a marked-up document. So I set out to write an XML editor that worked like a word processor, but that preserved XML semantics.
I released Vex as open-source partly with the hope I could get some development help, but mostly because I wanted lots of people to use it.
Eclipse Magazin: Are you still using Vex?
Sadly, no. I found it too difficult to keep on top of something like DocBook for the sort of general documentation I do. I tend to prefer wikis these days.
Eclipse Magazin: Which projects are you currently working on?
My spare time is consumed with starting a business, so I haven’t had time to produce any open-source work lately. I had an idea to factor the CSS rendering engine out of Vex, then use it as the basis for a tool to generate PDFs from HTML, so I may begin work on that shortly.
Eclipse Magazin: If you had to implement a visual XML editor again what you would do differently today?
I’m excited about modern browser technologies like <canvas> and SVG, so I think I might have done a browser-based editor. It would be cool to do something along the lines of Bespin but for documentation instead of code.
Eclipse Magazin: Thank you for the interview!
David Carver, migrated Vex from SourceForge to Eclipse
Eclipse Magazin: Hi Dave, it was you who took the initiative to migrate Vex from Sourceforge to Eclipse. What was the reason for it?
The main reason was to give VEX greater exposure. It had been sitting on SourceForge for years and active development on it had stopped. There was also a need for a good base in the XML community to have a good open source Visual editor for XML. There are good commercial visual editors, but none of them except for OxygenXML’s Author mode, worked natively on eclipse. I also wanted something that would take advantage of eclipse’s Web Tools Platform project’s XML support. VEX already being written on top of the eclipse RCP seemed to be a natural fit.
Eclipse Magazin: You are working for the STAR (Standards for Technology in Automotive Retail) Organization. What is your function and how free are you to work for Eclipse? [Outdated: currently, Dave works for IntalioWorks]
I’m the lead XML Data Architect at STAR. We develop business to business XML standards for the automotive, marine, powersports, and heavy duty truck retail industries. We also develop an XML IDE that is based on eclipse that is free for our members to use. STAR’s standard is used world wide, and helps reduce the data integration costs and maintenance that can occur. All of my eclipse project related work is done in my own spare time. STAR does not pay me to work on eclipse related projects. So, I work on it when I can.
Eclipse Magazin: Vex has now been an Eclipse project for more than one year. What improvements do you expect for 2010, what do you wish to be realised?
Ultimately, I want to grow the committer base for VEX. It’s just myself at the moment, but there is growing interest from others now that VEX has been at eclipse for a while. I’d like to mainly stabilize the base of VEX and get the main bugs worked out. In addition I would like to get the underlying EMF model that is backing editor to handle all of the XML related functionality. VEX’s EMF model is not reversed engineered from a schemas, but it is a straight ecore model. VEX is designed to handle any XML file you toss at it, and represent it visually with a CSS stylesheet. However, it does not support items like processing instructions, or comments. It looses these items when you serialize the XML file back to disk. So getting full round-tripping of the file, and having a source tab that leverages the WTP XML editor are both items I’d like to see done in 2010. In addition, it would be nice to have image rendering support as well.
Eclipse Magazin: The XPath processor PsycoPath is another project you have taken over from Sourceforge. Which project is the next on your list?
The PsychoPath XPath 2.0 processor was another hidden gem that sat on SourceForge for years. That is now part of the eclipse XSL Tools project in WTP Source Editing. That has taken up most of my time, and we now have it at a 99.6% pass rating of the W3C test suite. Next on my list of projects to get going is a RelaxNG set of Tools for eclipse. We have brought over Martin Schmied’s research project he developed that allows content assistance and validation with relaxng grammars. I have developed an XText based editor for the RelaxNG compact syntax language, which will allow editing and creation of RelaxNG Compact Syntax grammars. This is available in the WTP Incubator. Beyond that, I hope to spend some time on the WTP DOM to bring that into full compliance with the W3C test suite as well.
Eclipse Magazin: Thank you for the interview!