130 likes | 232 Views
XLIFF 2.0, How to put a date on it. Developing a modular, next-generation standard Bryan Schnabel and David Filip. XLIFF 2.0 might be smaller, but . . . . What is “modular”? What is core vs. module? Modules can be added nimbly Because the TC DID allow custom namespace based extensibility
E N D
XLIFF 2.0, How to put a date on it Developing a modular, next-generation standard Bryan Schnabel and David Filip
XLIFF 2.0 might be smaller, but . . . • What is “modular”? • What is core vs. module? • Modules can be added nimbly • Because the TC DID allow custom namespace based extensibility • However, at strictly predefined points • Processing Requirements discussion still live
Why 20% is better than 1.2 ? • Core is the lowest common denominator • A study by Micah Bly (1st XLIFF Symposium 2010) showed that only <source> and <target> were supported by all tools • of course including the necessary parent elements • XLIFF 2.0 brings a core that is bigger than that but still significantly smaller than 1.2 • We heard the customer voice • Stakeholders said that the standard is so complex that they do not know what to do with all the features.. • Also it DOES have PROCESSING REQUIREMENTS and a formal conformance clause.
After Core is a standard – adding modules will be easier • Lifecycle of a good idea • extension module core • Demotion of unused features from core and modules • core module extension • This should make use of the empirical framework designed by Asanka • IFF the framework receives community support and will be sourced with a REPRESENTATIVE corpus as result
XLIFF 2.0 = core + some modules • Declared XML Namespaces: • urn:oasis:names:tc:xliff:document:2.0 • urn:oasis:names:tc:xliff:matches:2.0 • urn:oasis:names:tc:xliff:glossary:2.0 • urn:oasis:names:tc:xliff:metadata:2.0 • MultilingualWeb-LT WG at W3C is working on the canonical ITS 2.0 mapping onto XLIFF and back • Extension Module • Profile should be published as a joint Note
You can check on us yourself • The calendar date depends on how quickly we address each of the features tracked on our wiki page https://wiki.oasis-open.org/xliff/XLIFF2.0/FeatureTracking • This is a work in process, here’s the editor’s draft https://tools.oasis-open.org/version-control/browse/wsvn/xliff/trunk/xliff-20/xliff-core.pdf
An frank look at our timeline • February 2008, XLIFF 1.2 is published • TC begins observing its reception • Late 2008, early 2009, wiki tracking page marked the honest true start of technical requirements gathering • XLIFF Symposium, September 2010, TC gets its first formal “voice of customer” • TC decides XLIFF 2.0 cannot be backward compatible to address issues and bring business benefits • Modularity • Minimalism • Metadata
In that light . . . • XLIFF is in fact a normal 2-3 years standardization project from inception to OASIS stamp.. • Changes from 1.0 to 1.1 to 1.2 were incremental without much customer feedback. Only in 2010 the TC decided what 2.0 should be and started working on it seriously. • Momentum from Corporate players joining in 2011 IBM and Oracle, MSFT in February 2012
Drum roll please . . . • Bryan says January 2013 • Rodolfo says December 2012 • David says Current specification can become Committee Draft by the end of year, however it is critical to add test suit before 60 days public review is over.
Conformance test suite • Looking for volunteers to help with creating conformance testing input and output files for all Processing Requirements and normative statements in the spec
Continued work on “voice of customer” • 2nd XLIFF State of the Art report • Changing the report methodology to enhance self reporting with sample files • Make passing tests a condition for toolmakers to state conformance • No certification! Just peer monitored set of conformance tests.. • Proceedings of 3rd Symposium as part of FEISGILTT • Including customer feedback analysis • Asanka is recording
Discussion • Yves is in the building • And other TC members including of a handful of MSFT representatives Open discussion now!