gnuplot-beta

Re: Strange behaviour of "set xtics"

[Petr M. <mi...@ph...>]

> gnuplot> reset
> gnuplot> unset xtics          # <-- must be unset first
> gnuplot> set border 1+2+8     # remove top border
> gnuplot> set xtics nomirror   # xtics at bottom, but *not* at top
> gnuplot> p sin(x)             # voil=FF=FF- tics at top!

I have fixed it (also for yet another case undiscovered above).

But I see yet another bug:

 =09reset; unset tics; splot x

this annulates "ticslevel" in current gnuplot 4.1. I wonder how this is=20
possible?

---
PM

Re: Strange behaviour of "set xtics"

[Petr M. <mi...@ph...>]

>> But I see yet another bug:
>>
>>     reset; unset tics; splot x
>> 
>> this annulates "ticslevel" in current gnuplot 4.1. 
>
> It's not exactly annulating the ticslevel (as you can see if you 'show tics': 
> it's still 0.5).  But if there's no baseplane being drawn (x and y tics and 
> grids all off), then there's not much use for the extra space controlled by 
> 'set ticslevel', either.  The ticslevel controls the position of something 
> that's not part of this plot, so it can't have any effect.
>
> To see it for yourself, look at graph3d.c:do_3dplot(), line 566.

I considered "unset tics": tics are not drawn, but nothing else changes, so 
neither the base plane. 'help ticslevel' does not write about any 
correlation with '(un)set tics'. But maybe it does not matter...

BTW, in 'help ticslevel' there is:

        set ticslevel <level>
        set xyplane <level>
        set xyplane at <level>

but later 'frac' and 'zvalue' are used instead of <level>. I think it should 
be changed to

        set ticslevel <frac>
        set xyplane <frac>
        set xyplane at <zvalue>

---
PM

Re: Strange behaviour of "set xtics"

[<br...@ph...>]

Petr Mikulik wrote:
> I considered "unset tics": tics are not drawn, but nothing else changes, 
> so neither the base plane. 

Well, the base plane *does* change --- it vanishes.  As I see it, the 
whole reason the base plane exists is to allow the base grid to be far 
enough from the surface to not be confused with it.

> but later 'frac' and 'zvalue' are used instead of <level>. I think it 
> should be changed to
> 
>        set ticslevel <frac>
>        set xyplane <frac>
>        set xyplane at <zvalue>

Ethan: since it appears you introduced 'set xyplane', I think this up to 
you.

Re: Strange behaviour of "set xtics"

[Ethan A M. <merritt@u.washington.edu>]

On Saturday 04 February 2006 03:33 pm, Hans-Bernhard Br=F6ker wrote:
> Petr Mikulik wrote:
> > I considered "unset tics": tics are not drawn, but nothing else changes=
,=20
> > so neither the base plane.=20

When you "unset title", extra vertical space above the plot is not reserved.
When there is no xlabel, extra vertical space below the plot is not reserve=
d.
When there is no ylabel, extra space to the left of the plot is not reserve=
d.
So logically when there is no base plane, there should be no space reserved
for it.

> > but later 'frac' and 'zvalue' are used instead of <level>. I think it=20
> > should be changed to
> >=20
> >        set ticslevel <frac>
> >        set xyplane <frac>
> >        set xyplane at <zvalue>
>=20
> Ethan: since it appears you introduced 'set xyplane', I think this up to=
=20
> you.

I will change the documentation, if people think that <frac> and <zvalue>
are clearer.


=2D-=20
Ethan A Merritt
Biomolecular Structure Center
University of Washington, Seattle 98195-7742

Re: Strange behaviour of "set xtics"

[<br...@ph...>]

Ethan A Merritt wrote:
>> Petr Mikulik wrote:

>>> but later 'frac' and 'zvalue' are used instead of <level>. I think it 
>>> should be changed to
>>>
>>>        set ticslevel <frac>
>>>        set xyplane <frac>
>>>        set xyplane at <zvalue>

> I will change the documentation, if people think that <frac> and <zvalue>
> are clearer.

Clarity is not the main issue, consistency is.  The same names used for 
<arguments> in the Syntax overview are supposed to be appear in the 
longhand explanation, so people know what to put where.   That's what 
they have <names> for in the first place.

Re: Strange behaviour of "set xtics"

[<br...@ph...>]

Petr Mikulik wrote:

> But I see yet another bug:
> 
>     reset; unset tics; splot x
> 
> this annulates "ticslevel" in current gnuplot 4.1. 

It's not exactly annulating the ticslevel (as you can see if you 'show 
tics': it's still 0.5).  But if there's no baseplane being drawn (x and 
y tics and grids all off), then there's not much use for the extra space 
controlled by 'set ticslevel', either.  The ticslevel controls the 
position of something that's not part of this plot, so it can't have any 
effect.

To see it for yourself, look at graph3d.c:do_3dplot(), line 566.