Re: Pretty printing OWL FSS : request for preferences and samples

On 11 Mar 2015, at 16:30, Simon Spero <sesuncedu@gmail.com> wrote:

> I've been doing a little work on improving parts of the OWLAPI rendering code in order to render things in order (this makes a huge difference for VCS).
> 

Very happy to hear that.  Can we now use the OWLAPI to get fixed ordering of functional syntax and/or Manchester Syntax files?


> Now that this seems to be behaving, I'm feeling motivated to fix something that I've meant to poke at for a long time; the way that OWLAPI FSS renderers every axiom on a single line[1].
> 
> Whenever I need to format an even moderately complicated line says, I end up copying it in to an emacs buffer, and inserting line breaks and indentation by hand. 
> 
> Sometimes this  layout is formatting  for a \verbatim or a slide. These cases are  driven by space and aesthetic concerns, and are generally  hand tweaked anyway.
> 
> Generally I'm doing this to try and figure out what the difference is between two similar looking blobs of text (like a c-diff). This is a good fit with the needs of a VCS.
> 
> There are a number of issues to consider when rendering. Some of this is a tension between minimizing the size of line diffs and minimizing the use of vertical space.
> 
> For example:
> 
> * Should there always be a line break after an axiom annotation? What about after an annotation annotation+
> 
> * For axioms like EquivalentClasses, if all the class expressions will fit on a single line should they be rendered in a single line?
> 
> * If any expressions are too long to fit the target width, should all expressions be followed by a line break?
> 
> * If an axiom or expression is split over multiple lines, should closing parentheses be on their own line, like C/Java braces, or should they be grouped together, lisp-style?
> 
> * the closing parenthesis for "Ontology(" traditionally appears on a line by itself. This seems to be a special case; are there other cases where this is idiomatic?
> 
> The examples in the OWL 2 syntax specification are not uniform.
> 
> Does anybody have strong preferences? Otherwise I am likely  to err on the side of using more line breaks sake of simplicity.
> 
> Simon
> 
> [1] Except for literals that contain newlines.
>