Menu
▾
▴
gnuplot-info
|
From: Bo J. <li...@f7...> - 2014-06-06 14:58:00
|
Dear all I am trying to understand how gnuplot sets minor tics on a logaritmic scale and if that is "correct" running the following minimal example: set logscale x set xrang [10:100] set yrang [1:10] set mxtics 4 plot "<echo '50 1'" shows that minor tics are sat at 25,50, and 75 The first interval represents a smaller linear range (10-25) whereas the later 3 represents a range of 25. It that what is expected for minor tics on a log scale. setting mxtics to 8 is even worse I now gets minor tics at: 12.5, 50, 37.5 ... in which the first interval only represents 2.5 on a linear scale, and the minor tics seems totally out of place Does any one here know what is the correct way to put minor tics? Best Bo |
|
From: BBands <bb...@gm...> - 2014-06-06 21:11:07
|
Does 'set mxtics 10' seem correct to you?
John
On Fri, Jun 6, 2014 at 7:57 AM, Bo Jakobsen <li...@f7...> wrote:
> Dear all
>
> I am trying to understand how gnuplot sets minor tics on a logaritmic
> scale and if that is "correct"
>
> running the following minimal example:
>
> set logscale x
> set xrang [10:100]
> set yrang [1:10]
> set mxtics 4
> plot "<echo '50 1'"
>
> shows that minor tics are sat at 25,50, and 75
>
> The first interval represents a smaller linear range (10-25) whereas the
> later 3 represents a range of 25.
> It that what is expected for minor tics on a log scale.
>
> setting mxtics to 8 is even worse I now gets minor tics at: 12.5, 50,
> 37.5 ...
> in which the first interval only represents 2.5 on a linear scale, and
> the minor tics seems totally out of place
>
> Does any one here know what is the correct way to put minor tics?
|
|
From: BBands <bb...@gm...> - 2014-06-06 21:19:34
|
This seems correct to me.
set logscale x
set logscale y
set xrang [1:100]
set yrang [1:100]
set mxtics 10
set mytics 10
set grid mxtics mytics
plot "<echo '50 50'"
John
On Fri, Jun 6, 2014 at 7:57 AM, Bo Jakobsen <li...@f7...> wrote:
> Dear all
>
> I am trying to understand how gnuplot sets minor tics on a logaritmic
> scale and if that is "correct"
>
> running the following minimal example:
>
> set logscale x
> set xrang [10:100]
> set yrang [1:10]
> set mxtics 4
> plot "<echo '50 1'"
>
> shows that minor tics are sat at 25,50, and 75
>
> The first interval represents a smaller linear range (10-25) whereas the
> later 3 represents a range of 25.
> It that what is expected for minor tics on a log scale.
>
> setting mxtics to 8 is even worse I now gets minor tics at: 12.5, 50,
> 37.5 ...
> in which the first interval only represents 2.5 on a linear scale, and
> the minor tics seems totally out of place
>
> Does any one here know what is the correct way to put minor tics?
|
|
From: Ethan M. <eam...@gm...> - 2014-06-06 21:38:06
|
The issue of minor tics on a log scale axis has been problematic since the origin of gnuplot, lost in the dim mists of time. The oldest bug in the tracker complains about this. But over the last 10 years I have never seen a coherent proposal about how the program could better handle a "set mxtics" command in log-scale mode, let alone a patch that would implement it. IMHO the best thing is to treat "set/unset mxtics" in log scale mode as an on/off operation only. "on" means you get autogenerated minor tics, "off" means you don't get any. As John points out, you can generally set mxtics to 10. That fixes some odd corner cases when the total range on x is less than one decade, but can introduce strangeness in other corner cases. On Fri, Jun 6, 2014 at 2:19 PM, BBands <bb...@gm...> wrote: > This seems correct to me. > > set logscale x > set logscale y > set xrang [1:100] > set yrang [1:100] > set mxtics 10 > set mytics 10 > set grid mxtics mytics > plot "<echo '50 50'" > > John > > On Fri, Jun 6, 2014 at 7:57 AM, Bo Jakobsen <li...@f7...> wrote: > > Dear all > > > > I am trying to understand how gnuplot sets minor tics on a logaritmic > > scale and if that is "correct" > > > > running the following minimal example: > > > > set logscale x > > set xrang [10:100] > > set yrang [1:10] > > set mxtics 4 > > plot "<echo '50 1'" > > > > shows that minor tics are sat at 25,50, and 75 > > > > The first interval represents a smaller linear range (10-25) whereas the > > later 3 represents a range of 25. > > It that what is expected for minor tics on a log scale. > > > > setting mxtics to 8 is even worse I now gets minor tics at: 12.5, 50, > > 37.5 ... > > in which the first interval only represents 2.5 on a linear scale, and > > the minor tics seems totally out of place > > > > Does any one here know what is the correct way to put minor tics? > > > ------------------------------------------------------------------------------ > Learn Graph Databases - Download FREE O'Reilly Book > "Graph Databases" is the definitive new guide to graph databases and their > applications. Written by three acclaimed leaders in the field, > this first edition is now available. Download your free book today! > http://p.sf.net/sfu/NeoTech > _______________________________________________ > gnuplot-info mailing list > gnu...@li... > Membership management via: > https://lists.sourceforge.net/lists/listinfo/gnuplot-info > |
|
From: BBands <bb...@gm...> - 2014-06-06 22:13:14
|
In coding our own stock-market analysis graphics program we stumbled
across this mine field. We chose to give the user three options: log
tics, linear tics and decades. 'Log tics' produces equidistnat tics on
the chosen axis, 'linear grid' produces tics every n points apart and
'decades' is the gnuplot and perhaps conventional default. Explaining
this to users is hell. We just say, "Try all three and use the one you
like best."
Thinking about this makes my head hurt,
John
On Fri, Jun 6, 2014 at 2:37 PM, Ethan Merritt <eam...@gm...> wrote:
> The issue of minor tics on a log scale axis has been problematic since the
> origin of gnuplot, lost in the dim mists of time. The oldest bug in the
> tracker complains about this. But over the last 10 years I have never seen
> a coherent proposal about how the program could better handle a "set mxtics"
> command in log-scale mode, let alone a patch that would implement it.
>
> IMHO the best thing is to treat "set/unset mxtics" in log scale mode as an
> on/off operation only. "on" means you get autogenerated minor tics, "off"
> means you don't get any.
>
> As John points out, you can generally set mxtics to 10. That fixes some odd
> corner cases when the total range on x is less than one decade, but can
> introduce strangeness in other corner cases.
|
|
From: Bo J. <li...@f7...> - 2014-06-10 07:14:48
|
Thanks for the input.
I guess no optimal solution exists, it is hard to ensure the "reader" of
the
graph knows what the minor ticks means unless the "normal" 8 minor
ticks is used.
I will try to figure out if any convention exists in the mathematical
literature, and
return if I find any.
Best
Bo
On 2014-06-07 00:13, BBands wrote:
> In coding our own stock-market analysis graphics program we stumbled
> across this mine field. We chose to give the user three options: log
> tics, linear tics and decades. 'Log tics' produces equidistnat tics on
> the chosen axis, 'linear grid' produces tics every n points apart and
> 'decades' is the gnuplot and perhaps conventional default. Explaining
> this to users is hell. We just say, "Try all three and use the one you
> like best."
>
> Thinking about this makes my head hurt,
>
> John
>
>
> On Fri, Jun 6, 2014 at 2:37 PM, Ethan Merritt <eam...@gm...> wrote:
>> The issue of minor tics on a log scale axis has been problematic since the
>> origin of gnuplot, lost in the dim mists of time. The oldest bug in the
>> tracker complains about this. But over the last 10 years I have never seen
>> a coherent proposal about how the program could better handle a "set mxtics"
>> command in log-scale mode, let alone a patch that would implement it.
>>
>> IMHO the best thing is to treat "set/unset mxtics" in log scale mode as an
>> on/off operation only. "on" means you get autogenerated minor tics, "off"
>> means you don't get any.
>>
>> As John points out, you can generally set mxtics to 10. That fixes some odd
>> corner cases when the total range on x is less than one decade, but can
>> introduce strangeness in other corner cases.
|
