Re: get rid of minimum-[XY]-extents (issue 6294086)

[david . nalesnik]

On 2012/06/19 10:37:28, janek wrote:
> Hi Keith,
> David (i'd like to know your opinion about something here)
> and all,

> > The minimum-*-extent is convenient when you want extra
> > space on one side, and don't want to think about what
> > extent the object should have on the other side.

> Indeed they can be handy in situations like that.
> However, i think that we should approach this problem differently,
> in a more flexible and less hard-coded way.

> Here's the downside of minimum-*-extent solution: what if i need to
> specify extent's upper bound instead of lower one?  There's no
> "maximum-[XY]-extent" property.  What if i want to restrict
[XY]-offset?
> Again, [minimum|maximum]-[XY]-offset properties don't exist.
> So, should we create them for all similar properties?
> Surely not, that would be a hell lot of boilerplate code!

> I think that we should use scheme functions for things like that.
> I'm pretty sure that one can easily write a scheme function
> which would set [XY]-extent to at least something.
> Or a funtion that would modify only one element in a pair.
> David, could you confirm that it's possible to easily define
> such funtions?

It's no problem to create higher-order functions, and you can easily
manipulate the elements of pairs and lists in any number of ways.  What
I can't visualize here is how a procedure which, say, limits the
elements of a pair to a maximum, should be called (if not as a music
function, as I did in the offsetting function you mention below).

> Keith, for an illustration of such "modifying functions" see this
> email from David Nalesnik:
> http://lists.gnu.org/archive/html/lilypond-devel/2012-06/msg00102.html



http://codereview.appspot.com/6294086/