Every year or so, I glance at patents related to Schematron and XML.  There are a lot of bad ones, with no inventive step: it is wrong to think of them all as just scams on investors or land grabs, I think they are mainly taken up defensively: the patent system is design to discourage innovation by penalizing independent invention and rewarding the burying of disruptive technology. I have previously written a few articles here on this:  European Patent Crisis needs to be resolved and Schematron-Related Patents.

But here is one that takes the cake: Universal XML Validator (UXV) Tool: patent application  US20150278386A1  Basically, it wants to patent any kind of XML validator, that can validate one or more portions of a document against one or more grammatic or syntactic rules (including Schematron) generate some kind of report.  I am not sure why this does not directly conflict with IBMs junk patent Validating an XML Document US20150356069A1  which also talks of validating at least parts of the data objects with at least parts of the rules including Schematron rules.

It feeds my suspicion that, after the i4i patent result, the most ludicrous and waffly mention of obvious implementation techniques or uses of XML may be fair game for patenters.

But the fact remains that it completely obvious to anyone not living under a rock (and probably to many of those too) that you could apply part of a schema to part of a document.  That is the entire operation of IDEs for example; to take only part of a document and validate just what is needed in context.  All programming language parsers and linters and static analysis tools have options to let you select the rules to validate the program or text or image; now it is true that the first generation of tools using CTAGs format results would just process the whole document and then filter the results to the region of interest, but by the 1980s this was being replaced by object systems that worked on selections and regions (which is the basis of the Language Serving Protocol.)  Look at Microsoft Intellisense for a ubiquitous and venerable example of validating just part of a document. And, for SGML, you only need to look at the first generation of editors (such as Author/Editor) in the 1990s to see contextual validation, where the rules of a content model would be selected and applied, not the whole thing. Going further back, there was FrameMakers pre-SGML rules languages, for example.

So the idea of validating parts of documents with a subset of rules is not only obvious by analogy with programming languages but actual ubiquitous in practice, in structured editors.

Like jokes and porn and art, I think we know an innovative idea when we see it. Some patents are clearly major steps in intuition and embody hard work.   But not this one, unless I have missed some step.

Oh and while I am at it: if any patent examiners are reading this, could they please note that a paragraph saying “And the document could be validated with Schematron.”  says absolutely nothing innovative: any and every XML document document can be validated by Schematron. Similarly with “And a report generated“: validation/verification implies a result coming back.