XProc and Calabash progress

Volume 11, Issue 63; 28 Sep 2008; last modified 08 Oct 2010

The second “Last Call” for XProc ended on Friday. Things are looking good for progress to Candidate Recommendation. That means we need implementations!

Nothing is as simple as we hope it will be.

Jim Horning

The XProc specification started a second “Last Call” period back in August. That period ended on Friday, with a scant 33 issues raised. Many of those have already been closed and I fully expect we'll close the rest in short order.

The only significant unresolved issue, I think, is the question of the p:encrypt/p:decrypt steps for performing XML encryption and decryption. With my chair's hat on, I'm trying to engage with the XML Security Working Group to come to closure on those steps. Various constraints on my schedule have so far stymied my efforts.

As things stand, those steps are not part of the V1.0 specification. And I think maybe we should leave it that way. The earliest XProc drafts did have steps for that purpose, and there's at least one use case that needs them, but we removed them some time ago when it became clear that we (the members of the XML Processing Model Working Group) didn't have the relevant expertise to define them.

I'd like to set a precedent for small, separate specifications to define groups of steps related to one particular area or another. I'd like to see the RDF community define a set of RDF-related steps, and I think I'd like to see the security community define a set of security-related steps.

XProc is extensible, not everything has to be in the core language in V1.0. But there aren't any obvious, hard-and-fast rules; we'll just have to see how it works out.

XML Calabash LogoWith Candidate Recommendation in our sights, it's time to start gathering implementation experience. To do that, we need implementations. I know of a few in the works, including my own, Calabash.

Calabash is still “alpha” quality, but I think “beta” may be visible in the not too distant future. I've finally flattened all the issues that were preventing pipelines from calling user-defined pipeline steps. That's been my highest priority bug for a while now. There are a few more things that need to be fixed before beta, but if you feel like taking it for a whirl, please feel free.

And remember the xproc-dev mailing list for discussing your adventures!


Hi Norm,

I really like the idea of a core XProc recommendation, supplemented with a few smaller recommendations, targeted at more dedicated scopes. If those only define additional steps, that sounds like defining an extension or library, like any other (besides the fact that they would be designed by an authoritative organization, and that's a lot!)

Even if they are just published as notes, that would be a good place to centralize commonly used steps behind the scope of XProc core, and to provide info on how common extensions are designed, for further work on XProc itself...

Hope you'll be able to convince RDF and Security WGs!

-- Florent Georges

—Posted by Florent Georges on 01 Oct 2008 @ 02:03 UTC #