Admittedly yes, it is a weak argument (unless you're a basic user trying to put your TEI to work :)) 
On the other hand, I've already heard the difficulty of processing block-level <app>s used as a counter-argument, against the need for block-level <app/>... 

I would argue that the point is not to loosen the TEI model, but to correct it. 
As it is, the TEI model is assuming that variants do not occur at div or paragraph levels. This is just plain wrong. 
So why is the TEI assuming this very wrong thing? Either because it thinks variants affecting whole divs or whole paragraphs are not variants, and it just needs to be corrected because it is just a wrong assumption. Or because it thinks that allowing <app> to contain <p>s or even <div>s will make the user run into horrible issues of overlapping hierarchies; this is a groundless assumption, and here the TEI is creating very real problems for the users by trying to protect them from imagined problems. 
Just let the users use <app> to contain a <div> or <p>, when it makes sense and is the right thing to do philologically. If overlapping hierarchy issues arise, then standoff is the right thing to do, but there really is no point in strong-arming people to use standoff when their usual parallel-segmentation method is obviously suitable. 
We're really creating a problem that has no business even existing, by constraining the content model of <app>

Best wishes, 

----- Mail original -----
De: "Sebastian Rahtz" <[log in to unmask]>
À: "Burghart Marjorie" <[log in to unmask]>
Cc: [log in to unmask]
Envoyé: Mardi 19 Mai 2015 22:53:39
Objet: Re: Encoding divergent ending of a story (or: the <app> strikes back)

> On 19 May 2015, at 21:49, Burghart Marjorie <[log in to unmask]> wrote:
> If I encode my edition in standoff, I won't be able to use 3rd party tools like the Versioning Machine, or the TEI Critical Apparatus Toolbox, etc. etc. 
> So I stand by my point b). In the example I gave in the beginning of this thread, the logical thing is to use <app/> at block-level. This way the user can still benefit from the generic tools and stylesheets provided for the parallel segmentation method. 

This is a pretty weak type of argument. Loosen  the TEI model just because it makes things easier with one generation of tools?

Not but what you can't make a perfectly good intellectual argument for block-level <app>.

Sebastian Rahtz
Chief Data Architect, IT Services