Any interest in Xquery language?

  • 5
  • 22

As part of a consulting job, I have been developing a base Xquery language. The job was a data transformation project and the plan was to have some higher level dsls map down to the xquery language, and generate the scripts from it. I have about 60% of the language and generator finished, but the customer cancelled the project last week.

I'm thinking about open-sourcing the language, but I wasn't sure how much demand there would be (Xquery isn't the most popular kid on the block). But if anyone is interested, I can release what I have, although I probably won't have time to finish it in the near term. But if someone else wants to pick it up and run with it, that would be great.

Also, I'm looking for new projects, and if anyone has any MPS related projects, they would get top priority (and second, and third - this is the area I want to work in more than any other).
XQuery is interesting, though strictly speaking one might argue that XSLT won the battle, despite absolute atrocious support for both these languages in common frameworks such as .NET.

I'm working on a text editor in MPS, the idea being to make a blog text editor that supports popular mark-up options as well as advanced code gen features. Its predecessor can be found <a href="http://bitbucket.org/nesteruk/typografix">here</a>.

IMO it's a toss-up for transformation problems. The main advantage I think XSLT has over XQuery is pattern matching. But the syntax for XQuery is so much better to work with when coding by hand. Of course, when generating code that becomes less of an issue, although it's still nice to have a more concise syntax when reading the generated code (not that XQuery is particularly concise, but compared to XSLT it is).

Although the concrete syntax is very different, the abstract syntax is similar between the two. Once a language is developed for one, the other is trivial. The main complication with XQuery is that you need a textgen facet (actually I used gtext), whereas with XSLT you can just generate to the built-in XML language.

Your project sounds interesting, and I have had similar ideas with a different focus. Whereas you are focused more on the publishing side, I've been focusing on the integration and transformation side. Sounds like there might be some synergy there, and I have a lot more ideas if you want to chat and/or collaborate.

Yeah, feel free to ping me on skype (dmitri.nesteruk), I'm open for discussion. Also, by email: dn /at/ jetbrains /dot/ com.
Dmitri Nesteruk, Developer Evangelist, ReSharper Team

I think it would be a good idea to open-source your language. It's a common one and there's no reason of throwing away your work.

Unfortunately that project was cancelled before I completed the language. But I can donate what I have. If I recall, the type system was mostly complete along with the overall block/ statement structure as well as their generators. The expression system is what still needs work.


<< Newer
Older >>