gnuplot-beta

Three small bugs with splot

[Philipp K. J. <ja...@ie...>]

1) The inline nohidden3d seems to have no effect

	set isosamples 30; set hidden3d
	splot exp(-(x**2+y**2)), 0.5 nohidden3d

The plane should be transparent; it isn't.


2) Line types in key off by one in contour plots with hidden3d

	set isosamples 30; set hidden3d
	set contour both
	splot [-2:2][-2:2] exp(-(x**2+y**2))

If you look closely, you realize that the colors of the
linesamples in the key are off by one, because the color
that's used for the underside of the surface is not taken
into account. If you do

	set hidden3d offset 0

then everything is fine (but the underside of the surface
is the same color as the top).


3) set ticscale vs set xyplane with save command

The "set ticscale" option is deprecated in favor of
"set xyplane". However, upon "save", gnuplot saves
only a "set ticscale" entry to the command file, not
a "set xyscale" entry. It works, but it is confusing.

Best,

		Ph.

Re: Three small bugs with splot

[sfeam <sf...@us...>]

On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:

> 3) set ticscale vs set xyplane with save command
> 
> The "set ticscale" option is deprecated in favor of
> "set xyplane". However, upon "save", gnuplot saves
> only a "set ticscale" entry to the command file, not
> a "set xyscale" entry. It works, but it is confusing.

"set ticscale" controls the length of the axis tic marks.
It has no connection to "set xyplane".

	Ethan

Re: Three small bugs with splot

[Philipp K. J. <ja...@ie...>]

On Sun, 25 Jan 2015 21:46:11 -0800
sfeam <sf...@us...> wrote:

> On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
> 
> > 3) set ticscale vs set xyplane with save command
> > 
> > The "set ticscale" option is deprecated in favor of
> > "set xyplane". However, upon "save", gnuplot saves
> > only a "set ticscale" entry to the command file, not
> > a "set xyscale" entry. It works, but it is confusing.
> 
> "set ticscale" controls the length of the axis tic marks.
> It has no connection to "set xyplane".

My bad. I meant: "set ticslevel" (not "ticscale").

"save" does not write "set xyplane" to file,
it write out "set ticslevel". Which is deprecated.

> 
> 	Ethan
> 
> 

Re: Three small bugs with splot

[Ethan A M. <sf...@us...>]

On Sunday, 25 January, 2015 22:38:53 Philipp K. Janert wrote:
> On Sun, 25 Jan 2015 21:46:11 -0800
> sfeam <sf...@us...> wrote:
> 
> > On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
> > 
> > > 3) set ticscale vs set xyplane with save command
> > > 
> > > The "set ticscale" option is deprecated in favor of
> > > "set xyplane". However, upon "save", gnuplot saves
> > > only a "set ticscale" entry to the command file, not
> > > a "set xyscale" entry. It works, but it is confusing.
> > 
> > "set ticscale" controls the length of the axis tic marks.
> > It has no connection to "set xyplane".
> 
> My bad. I meant: "set ticslevel" (not "ticscale").
> 
> "save" does not write "set xyplane" to file,
> it write out "set ticslevel". Which is deprecated.

Are you sure?

That's not what I see here:

gnuplot> set xyplane at z = -.123
gnuplot> save 'foo'
gnuplot> !grep xyplane foo
set xyplane at -0.123
gnuplot> !grep ticslevel foo
gnuplot>

	Ethan

Re: Three small bugs with splot

[Daniel J S. <dan...@ie...>]

On 01/26/2015 12:50 PM, Ethan A Merritt wrote:
> On Sunday, 25 January, 2015 22:38:53 Philipp K. Janert wrote:
>
>>  On Sun, 25 Jan 2015 21:46:11 -0800
>
>>  sfeam <sf...@us...> wrote:
>
>>
>
>>  > On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
>
>>  >
>
>>  > > 3) set ticscale vs set xyplane with save command
>
>>  > >
>
>>  > > The "set ticscale" option is deprecated in favor of
>
>>  > > "set xyplane". However, upon "save", gnuplot saves
>
>>  > > only a "set ticscale" entry to the command file, not
>
>>  > > a "set xyscale" entry. It works, but it is confusing.
>
>>  >
>
>>  > "set ticscale" controls the length of the axis tic marks.
>
>>  > It has no connection to "set xyplane".
>
>>
>
>>  My bad. I meant: "set ticslevel" (not "ticscale").
>
>>
>
>>  "save" does not write "set xyplane" to file,
>
>>  it write out "set ticslevel". Which is deprecated.
>
> Are you sure?
>
> That's not what I see here:
>
> gnuplot> set xyplane at z = -.123
>
> gnuplot> save 'foo'
>
> gnuplot> !grep xyplane foo
>
> set xyplane at -0.123
>
> gnuplot> !grep ticslevel foo

Same here.  However, "ticslevel" is still saved.  E.g.,

gnuplot> set ticslevel -1
gnuplot> save 'foo'
gnuplot> !grep ticslevel foo
set ticslevel -1

If ticslevel is deprecated, then probably the equivalent "set xyplane 
relative <frac>" should be saved.

I suppose the easiest way to do this is to have a preprocess-type of 
step that replaces "set ticslevel <frac>" by the equivalent.  That way 
the command goes into the history queue with the right format.  Easier 
said than done, I guess, considering that gnuplot has word completion, e.g.,

gnuplot> se ticsl -1

is acceptable.

Dan

Re: Three small bugs with splot

[Philipp K. J. <ja...@ie...>]

On Mon, 26 Jan 2015 10:50 -0800
Ethan A Merritt <sf...@us...> wrote:

> On Sunday, 25 January, 2015 22:38:53 Philipp K. Janert wrote:
> > On Sun, 25 Jan 2015 21:46:11 -0800
> > sfeam <sf...@us...> wrote:
> > 
> > > On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
> > > 
> > > > 3) set ticscale vs set xyplane with save command
> > > > 
> > > > The "set ticscale" option is deprecated in favor of
> > > > "set xyplane". However, upon "save", gnuplot saves
> > > > only a "set ticscale" entry to the command file, not
> > > > a "set xyscale" entry. It works, but it is confusing.
> > > 
> > > "set ticscale" controls the length of the axis tic marks.
> > > It has no connection to "set xyplane".
> > 
> > My bad. I meant: "set ticslevel" (not "ticscale").
> > 
> > "save" does not write "set xyplane" to file,
> > it write out "set ticslevel". Which is deprecated.
> 
> Are you sure?

Yes, positive. That's with the gp5.0 release.

set xyplane 10
plot sin(x)
save "ticken"

fgrep "ticsl" ticken
set ticslevel 10

fgrep "xypl" ticken
(not found)

BUT:

set xyplane at 10
plot sin(x)
save "tocken"

fgrep "xypl" tocken
set xyplane at 10

fgrep "ticsl" tocken
(not found)

So, in other words, it works with 
	set xyplane at ...
but not with 
	set xyplane ...

Moreover, the default setting is set xyplane ...,
which means that, by default, gnuplot writes out
set ticslevel, not set xyplane.

> 
> That's not what I see here:
> 
> gnuplot> set xyplane at z = -.123
> gnuplot> save 'foo'
> gnuplot> !grep xyplane foo
> set xyplane at -0.123
> gnuplot> !grep ticslevel foo
> gnuplot>
> 
> 	Ethan

Re: Three small bugs with splot

[Daniel J S. <dan...@ie...>]

On 01/26/2015 01:24 PM, Philipp K. Janert wrote:
> On Mon, 26 Jan 2015 10:50 -0800
> Ethan A Merritt<sf...@us...>  wrote:
>
>> On Sunday, 25 January, 2015 22:38:53 Philipp K. Janert wrote:
>>> On Sun, 25 Jan 2015 21:46:11 -0800
>>> sfeam<sf...@us...>  wrote:
>>>
>>>> On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
>>>>
>>>>> 3) set ticscale vs set xyplane with save command
>>>>>
>>>>> The "set ticscale" option is deprecated in favor of
>>>>> "set xyplane". However, upon "save", gnuplot saves
>>>>> only a "set ticscale" entry to the command file, not
>>>>> a "set xyscale" entry. It works, but it is confusing.
>>>>
>>>> "set ticscale" controls the length of the axis tic marks.
>>>> It has no connection to "set xyplane".
>>>
>>> My bad. I meant: "set ticslevel" (not "ticscale").
>>>
>>> "save" does not write "set xyplane" to file,
>>> it write out "set ticslevel". Which is deprecated.
>>
>> Are you sure?
>
> Yes, positive. That's with the gp5.0 release.
>
> set xyplane 10
> plot sin(x)
> save "ticken"
>
> fgrep "ticsl" ticken
> set ticslevel 10
>
> fgrep "xypl" ticken
> (not found)
>
> BUT:
>
> set xyplane at 10
> plot sin(x)
> save "tocken"
>
> fgrep "xypl" tocken
> set xyplane at 10
>
> fgrep "ticsl" tocken
> (not found)
>
> So, in other words, it works with
> 	set xyplane at ...
> but not with
> 	set xyplane ...
>
> Moreover, the default setting is set xyplane ...,
> which means that, by default, gnuplot writes out
> set ticslevel, not set xyplane.

Oh, I see.  gnuplot is treating "xyplane" as if it is the deprecated 
syntax.  That is,

set ticslevel #

is being substituted for non-"at" variation

set xyplane #

when it should be the other way around.

Dan

Re: Three small bugs with splot

[sfeam <sf...@us...>]

On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
> 
> 1) The inline nohidden3d seems to have no effect
> 
> 	set isosamples 30; set hidden3d
> 	splot exp(-(x**2+y**2)), 0.5 nohidden3d
> 
> The plane should be transparent; it isn't.

The documentation says:

  As of gnuplot version 4.6, hidden3d also affects 3D plotting styles `points`,
  `labels`, `vectors`, and `impulses` even if no surface is present in the graph.
  Unobscured portions of each vector are drawn as line segments (no arrowheads).
  Individual plots within the graph may be explicitly excluded from this
  processing by appending the extra option `nohidden3d` to the `with` specifier.

This may be insufficiently explicit, but the intended meaning is that the
plot styles {points|labels|vectors|impulses} are now subject to hidden3d
by default but can be exempted using the "nohidden3d" keyword.
This keyword does not affect plot styles other than these four, and in
particular it makes no sense to try to exclude actual surfaces from the surface
processing.

Transparency is a whole other question.  

	Ethan

Re: Three small bugs with splot

[Philipp K. J. <ja...@ie...>]

On Mon, 26 Jan 2015 17:51:49 -0800
sfeam <sf...@us...> wrote:

> On Sunday, 25 January 2015 07:57:39 PM Philipp K. Janert wrote:
> > 
> > 1) The inline nohidden3d seems to have no effect
> > 
> > 	set isosamples 30; set hidden3d
> > 	splot exp(-(x**2+y**2)), 0.5 nohidden3d
> > 
> > The plane should be transparent; it isn't.
> 
> The documentation says:
> 
>   As of gnuplot version 4.6, hidden3d also affects 3D plotting styles
> `points`, `labels`, `vectors`, and `impulses` even if no surface is
> present in the graph. Unobscured portions of each vector are drawn as
> line segments (no arrowheads). Individual plots within the graph may
> be explicitly excluded from this processing by appending the extra
> option `nohidden3d` to the `with` specifier.
> 
> This may be insufficiently explicit, but the intended meaning is that
> the plot styles {points|labels|vectors|impulses} are now subject to
> hidden3d by default but can be exempted using the "nohidden3d"
> keyword. This keyword does not affect plot styles other than these
> four, and in particular it makes no sense to try to exclude actual
> surfaces from the surface processing.

Ok, that was indeed not clear to me. 
Might be worth restating in the doc that
inline "nohidden3d" only applies to those 
styles.