I agree with all of Robert’s comments.
From: <email@example.com> on behalf of Robert D Anderson <firstname.lastname@example.org>
Date: Friday, June 9, 2017 at 12:05 PM
To: OASIS DITA TC List <email@example.com>, <firstname.lastname@example.org>
Subject: [dita] Misc grammar file items
Finally, here is a collection of other items noticed while scanning the grammar files, and I think this has to be my last one for the day...
In base DITA, the content model for <dlentry> allows one-or-more <dt> and one-or-more <dd>, but in LwDITA it has one required <dt> and <dd> (as I'd expect for LwDITA). With that in mind - I'm not sure it makes sense to have the filter attributes on <dt> and <dd>, because filtering either of those cannot result in a valid model.
In LwDITA topics, @outputclass is nearly universal, as I expected. The only place it's not available is on the highlight elements b, i, sub, sup, and u. I'm not sure why it's removed from those, and not sure if it's intentional. I interpret that as an effort to discourage their use for semantic purposes -- "Don't use <b> and then call it something else with @outputclass" -- just because I've heard that advice in full DITA. But that doesn't seem to fit with the "simple" approach of LwDITA, and those elements all have @outputclass in full DITA, so it feels like they should be here.
Similarly ... in LwDITA topics, I don't see @outputclass on anything at all. It should at least be available on elements shared with topics. I also expected to find it on map elements to help with specialization mock-ups. Full DITA allows it on map, topicref, keydef, and linktext (but not topicmeta), so it would seem like a good idea to have it on those 4 of the 5 LwDITA map elements.