octave-maintainers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Patch to add axis position property


From: Peter Gustafson
Subject: Re: Patch to add axis position property
Date: Thu, 09 Aug 2007 18:51:37 -0400
User-agent: Thunderbird 2.0.0.6 (X11/20070806)

John W. Eaton wrote:
> On  9-Aug-2007, Shai Ayal wrote:
> 
> | Hi Peter,
> | I have a few comments:
> | 
> | On 8/9/07, Peter Gustafson <address@hidden> wrote:
> | > Hi all,
> | >
> | > Attached is a patch which adds axis position properties for a plotted
> | > line, ie corresponding to x2y2 etc for gnuplot and eventually plotyy
> | > from matlab.  Although certainly not a complete implementation (For
> | > example I've not yet determined how to set limits based on the
> | > options).  However it is usable in a very basic way and lays a
> | > foundation for further work.  I'd like to have it evaluated for inclusion.
> | >
> | > The property ends up like this:
> | >
> | > octave:1> h=plot(0); get(h)
> | > ans =
> | > {
> | >   type = line
> | >  ...
> | >   xaxislocation = bottom
> | >   yaxislocation = left
> | > }
> | 
> | This properties are not matlab compatible. The way matlab does plotyy
> | is to create two overlapping axes, and have each line be a chilld of
> | the appropriate axes.
> | I think it's OK to have octave as a superset of matlab, but I can't
> | see a way in which to reconsile these two approaches.
> 
> Right, Matlab does have {x,y}asislocation properties, but they are
> axes properties, not line properties, so I don't think we can include
> this patch.
> 
> jwe


OK, so I've learned a few things today.  Good.

I see a problem (you may see many problems)... If two simultaneous axes
are created with outer position [0 0 1 1], when plots are made to those
axes they do not preserve the boundary positions.  Meaning the box in
which the plot resides changes slightly depending on the labels etc that
surround it.  So although it would be possible to define two full frame
axes (one with x1y1, one with x1y2), it would still look bad.  Could
this be overcome?  Is this why nobody else has done this yet.

What else would impair me from defining these as axes properties, and
working it out that way?

Thanks,

Pete


reply via email to

[Prev in Thread] Current Thread [Next in Thread]