EditInPlace testing fears
Document Actions
Submitted by
jenn.
on 2005-05-20 12:35.
The presence of some major addition/deletion bugs in the new XPathified EIP has spared me temporarily from having to tackle full formal test-plan writing for EIP...but Brian and Max appear to be closing in on them pretty quickly.
I'm quaking in my boots, frankly, because EIP is so darned
complex, and I don't have a really good handle on the sorts of things
that can go wrong in the content arena.In addition to the obvious functional testing of the controls and user interface, I would somehow like to get at the possibility of EIP generating bad CNXML. Alas, I don't really know what "bad CNXML" is. The validators should catch most of those kind of issues, but I seem to recall that there are at least a couple of anti-best-practices that can cause formatting or other problems without actually being invalid. The old bug of insanely-complete namespace expansion is an example, where every tag with a math prefix got the entire MathML namespace declaration added to it by EIP.
This isn't any more than a suspicion, really, but I'll be taking a good look at EIP's autogenerated tags and ids to see if I can find any ways in which they're inferior to what a reasonably competent human would generate. For example, it seems that when you add a note, you don't get the option to give it a type or an id. If it's inside another element, you can edit its parent node and get to the note tag that way; but if it's a direct child of content, you can't edit the tag itself through EIP. This may just be a consequence of EIP being a partial editing solution, but I hate to see our default editor encouraging behaviors for which we would have admonished our content students in days past.
