Re: [RFC] Document level property drawer

[Sebastian Miele <sebastian.miele@gmail.com>]

Marco Wahl <marcowahlsoft@gmail.com> writes:

> [..]
>
> I think the distinction between Org file and Org subtree should be kept
> to a minimum.  Wouldn't it be nice if Org files can be considered as Org
> subtrees?

Yes, this would be very nice.

I have a related question or proposal.

Up until now, I basically do not use property drawers except absolutely
necessary, because properties are invisible by default. Properties "need
to be inserted into a [..] drawer", and "in order to look inside the
drawer, you need to move point to the drawer line and press ‘<TAB>’
there."

A place that comes to mind where I really would like to use properties
is the header-args property in order to specify parameters related to
tangling for all src blocks in certain subtrees.

But for such properties to satisfactorily work for me, they would have
to be visible by default. E.g. I would want the header-args to be
immediately visible just like they are when they are written after
#+BEGIN_SRC or #+HEADER. Otherwise I would find myself constantly
wondering whether this or that property drawer contains something
essential and every TAB on a collapsed headline would have be followed
by an accompanying move to the property drawer and a TAB there.

On the other hand, there are properties that are very good candidates
for remaining hidden by default, like ID.

I would like to be able to make a clear distinction between properties
that are visible by default and properties that are not. Maybe it would
be possible to allow some #+.. syntax following headings for subtree
properties that are visible by default. A requirement could be made that
such property specifications always have to be followed by a property
drawer, even if that is empty. Then everything #+.. that is before the
property drawer would belong to the heading/subtree, and everything #+..
that follows the drawer would be treated as it is until now.

Please tell me if I missed something and Org is already capable of
something like that. If not, are there others who would like
visible-by-default property specifications for headings/subtrees in
addition to invisible-by-default property specifications in drawers,
too?

Finally, I would like to state an opinion: If there is
visible-by-default (by #+..) and invisible-by-default (by drawers)
syntax for headings/subtrees, including level 0, it may be viable to
require them to be disjoint for each heading/subtree. Most probably it
would be good practice, anyway. And the precedence question raised
previously in this thread would be eliminated.