matplotlib-devel Mailing List for matplotlib (Page 25)

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

IPython notebooks got a write up in Nature.

http://www.nature.com/news/interactive-notebooks-sharing-the-code-1.16261

All of the examples depend on mpl.

Tom

Hi,

Thanks for your answers and advices.

2014-11-07 16:03 GMT+01:00 Benjamin Root <ben...@ou...>:

> I can't speak about trademarks, but with regards to projects that are
> released under BSD license, it is my understanding (but, IANAL) that it is
> perfectly valid to sell a product using BSD code. I also think that it is
> ok to relicense it as GPLv3, but I am not entirely clear on that.
>
> Trademarks are an entirely different issue, though, and one that I have
> zero clue about. Do get a lawyer, or at least read up on the legal FAQs
> from FSF and other organizations.
>
> Ben Root
>
> On Fri, Nov 7, 2014 at 9:49 AM, Apps Embedded <app...@gm...>
> wrote:
>
>> Hi,
>>
>> This was just in case we could get some answers about selling some
>> open-source code including Matplotlib.
>>
>> Apps Embedded Team.
>>
>> 2014-11-07 15:42 GMT+01:00 Sandro Tosi <mo...@de...>:
>>
>>> Hello,
>>> I agree with Thomas this is not the right (and only) forum to ask
>>> these questions. If you seek legal advice you should ask a lawyer.
>>>
>>> On Fri, Nov 7, 2014 at 2:28 PM, Apps Embedded <app...@gm...>
>>> wrote:
>>> > Hi,
>>> >
>>> > Could I name this app "Scipy Console Free" ?
>>>
>>> do no use any name already taken, thats a general rule of thumb to
>>> avoid legal issues
>>>
>>> >
>>> > In the BSD licence of Scipy library, it is said
>>> > "...
>>> >
>>> > Neither the name of Enthought nor the names of the SciPy Developers
>>> may be
>>> > used to endorse or promote products derived from this software without
>>> > specific prior written permission.
>>> >
>>> > ..."
>>> >
>>> > So Scipy Console shoud be ok ? Since Scipy does not seem to be
>>> forbidden by
>>> > the licence text.
>>>
>>> Nope you didnt get the licence right (at list that point), there it
>>> means you cannot say the entity Enthought or any SciPy developers (the
>>> people who wrote the scipy code) are endorsing your product as good,
>>> valid, appropriate, anything.
>>>
>>> Again, seek legal advices with legal professionels.
>>>
>>> Regards,
>>> --
>>> Sandro Tosi (aka morph, morpheus, matrixhasu)
>>> My website: http://matrixhasu.altervista.org/
>>> Me at Debian: http://wiki.debian.org/SandroTosi
>>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>
>>
>

I can't speak about trademarks, but with regards to projects that are
released under BSD license, it is my understanding (but, IANAL) that it is
perfectly valid to sell a product using BSD code. I also think that it is
ok to relicense it as GPLv3, but I am not entirely clear on that.

Trademarks are an entirely different issue, though, and one that I have
zero clue about. Do get a lawyer, or at least read up on the legal FAQs
from FSF and other organizations.

Ben Root

On Fri, Nov 7, 2014 at 9:49 AM, Apps Embedded <app...@gm...>
wrote:

> Hi,
>
> This was just in case we could get some answers about selling some
> open-source code including Matplotlib.
>
> Apps Embedded Team.
>
> 2014-11-07 15:42 GMT+01:00 Sandro Tosi <mo...@de...>:
>
>> Hello,
>> I agree with Thomas this is not the right (and only) forum to ask
>> these questions. If you seek legal advice you should ask a lawyer.
>>
>> On Fri, Nov 7, 2014 at 2:28 PM, Apps Embedded <app...@gm...>
>> wrote:
>> > Hi,
>> >
>> > Could I name this app "Scipy Console Free" ?
>>
>> do no use any name already taken, thats a general rule of thumb to
>> avoid legal issues
>>
>> >
>> > In the BSD licence of Scipy library, it is said
>> > "...
>> >
>> > Neither the name of Enthought nor the names of the SciPy Developers may
>> be
>> > used to endorse or promote products derived from this software without
>> > specific prior written permission.
>> >
>> > ..."
>> >
>> > So Scipy Console shoud be ok ? Since Scipy does not seem to be
>> forbidden by
>> > the licence text.
>>
>> Nope you didnt get the licence right (at list that point), there it
>> means you cannot say the entity Enthought or any SciPy developers (the
>> people who wrote the scipy code) are endorsing your product as good,
>> valid, appropriate, anything.
>>
>> Again, seek legal advices with legal professionels.
>>
>> Regards,
>> --
>> Sandro Tosi (aka morph, morpheus, matrixhasu)
>> My website: http://matrixhasu.altervista.org/
>> Me at Debian: http://wiki.debian.org/SandroTosi
>>
>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
>

On 7 Nov 2014 14:44, "Apps Embedded" <app...@gm...> wrote:
>
> Hi,
>
> You are right.
> Maybe a "Science Console" name should simply do it.
>
> And for the Python trademark, is there any legal issue with charging an
Android app using the python langage ?

This is like me asking you whether I can legally call something "android"
because you are android developers and thus totally qualified to speak on
behalf of Google's legal department.

-n

Hi,

This was just in case we could get some answers about selling some
open-source code including Matplotlib.

Apps Embedded Team.

2014-11-07 15:42 GMT+01:00 Sandro Tosi <mo...@de...>:

> Hello,
> I agree with Thomas this is not the right (and only) forum to ask
> these questions. If you seek legal advice you should ask a lawyer.
>
> On Fri, Nov 7, 2014 at 2:28 PM, Apps Embedded <app...@gm...>
> wrote:
> > Hi,
> >
> > Could I name this app "Scipy Console Free" ?
>
> do no use any name already taken, thats a general rule of thumb to
> avoid legal issues
>
> >
> > In the BSD licence of Scipy library, it is said
> > "...
> >
> > Neither the name of Enthought nor the names of the SciPy Developers may
> be
> > used to endorse or promote products derived from this software without
> > specific prior written permission.
> >
> > ..."
> >
> > So Scipy Console shoud be ok ? Since Scipy does not seem to be forbidden
> by
> > the licence text.
>
> Nope you didnt get the licence right (at list that point), there it
> means you cannot say the entity Enthought or any SciPy developers (the
> people who wrote the scipy code) are endorsing your product as good,
> valid, appropriate, anything.
>
> Again, seek legal advices with legal professionels.
>
> Regards,
> --
> Sandro Tosi (aka morph, morpheus, matrixhasu)
> My website: http://matrixhasu.altervista.org/
> Me at Debian: http://wiki.debian.org/SandroTosi
>

Hi,

You are right.
Maybe a "Science Console" name should simply do it.

And for the Python trademark, is there any legal issue with charging an
Android app using the python langage ?

Apps Embedded Team.

2014-11-07 15:36 GMT+01:00 Thomas Caswell <tca...@gm...>:

> You need to ask that on the scipy list, not here.
>
> I suspect no project will be happy with you shadowing their name.
>
> I suggest you also talk to NumFocus.
>
> Again, this is written in a personal capacity.
>
> Tom
>
> On Fri, Nov 7, 2014 at 9:28 AM, Apps Embedded <app...@gm...>
> wrote:
>
>> Hi,
>>
>> Could I name this app "Scipy Console Free" ?
>>
>> In the BSD licence of Scipy library, it is said
>> "...
>>
>>    - Neither the name of Enthought nor the names of the SciPy Developers
>>    may be used to endorse or promote products derived from this software
>>    without specific prior written permission.
>>
>> ..."
>>
>> So Scipy Console shoud be ok ? Since Scipy does not seem to be forbidden
>> by the licence text.
>>
>> Thanks for your help.
>>
>> Apps Embedded Team.
>>
>> 2014-11-07 15:03 GMT+01:00 Thomas Caswell <tca...@gm...>:
>>
>>> I (personally) would prefer you not use the work 'pylab' in your app
>>> name.  There is enough confusion related to the term 'pylab' as it is with
>>> out adding another (some what orthogonal) meaning.
>>>
>>> I am speaking for my self in this email.
>>>
>>> Tom
>>>
>>> On Fri, Nov 7, 2014 at 8:21 AM, Apps Embedded <app...@gm...>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> We are about to publish an Android app released under GPL v3 with full
>>>> source access
>>>> giving the SciPy environnement to any Android devices.
>>>>
>>>> We would like to call this app "Pylab Console" on a freemium model with
>>>> a free and premium version.
>>>> This app will give graphics support in its premium version.
>>>> And thus it will use Matplotlib version 1.4.2
>>>>
>>>> From a legal point of view, are we able to use the term "Pylab" in our
>>>> Android App name ?
>>>>
>>>> Moreover, as Python is a trademark, should we need to ask the PSF the
>>>> authorisation of publishing such an app ?
>>>>
>>>> Thanks for your help.
>>>>
>>>> Apps Embedded Team.
>>>>
>>>>
>>>> ------------------------------------------------------------------------------
>>>>
>>>> _______________________________________________
>>>> Matplotlib-devel mailing list
>>>> Mat...@li...
>>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>>>
>>>>
>>>
>>>
>>> --
>>> Thomas Caswell
>>> tca...@gm...
>>>
>>
>>
>
>
> --
> Thomas Caswell
> tca...@gm...
>

Hello,
I agree with Thomas this is not the right (and only) forum to ask
these questions. If you seek legal advice you should ask a lawyer.

On Fri, Nov 7, 2014 at 2:28 PM, Apps Embedded <app...@gm...> wrote:
> Hi,
>
> Could I name this app "Scipy Console Free" ?

do no use any name already taken, thats a general rule of thumb to
avoid legal issues

>
> In the BSD licence of Scipy library, it is said
> "...
>
> Neither the name of Enthought nor the names of the SciPy Developers may be
> used to endorse or promote products derived from this software without
> specific prior written permission.
>
> ..."
>
> So Scipy Console shoud be ok ? Since Scipy does not seem to be forbidden by
> the licence text.

Nope you didnt get the licence right (at list that point), there it
means you cannot say the entity Enthought or any SciPy developers (the
people who wrote the scipy code) are endorsing your product as good,
valid, appropriate, anything.

Again, seek legal advices with legal professionels.

Regards,
-- 
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi

You need to ask that on the scipy list, not here.

I suspect no project will be happy with you shadowing their name.

I suggest you also talk to NumFocus.

Again, this is written in a personal capacity.

Tom

On Fri, Nov 7, 2014 at 9:28 AM, Apps Embedded <app...@gm...>
wrote:

> Hi,
>
> Could I name this app "Scipy Console Free" ?
>
> In the BSD licence of Scipy library, it is said
> "...
>
>    - Neither the name of Enthought nor the names of the SciPy Developers
>    may be used to endorse or promote products derived from this software
>    without specific prior written permission.
>
> ..."
>
> So Scipy Console shoud be ok ? Since Scipy does not seem to be forbidden
> by the licence text.
>
> Thanks for your help.
>
> Apps Embedded Team.
>
> 2014-11-07 15:03 GMT+01:00 Thomas Caswell <tca...@gm...>:
>
>> I (personally) would prefer you not use the work 'pylab' in your app
>> name.  There is enough confusion related to the term 'pylab' as it is with
>> out adding another (some what orthogonal) meaning.
>>
>> I am speaking for my self in this email.
>>
>> Tom
>>
>> On Fri, Nov 7, 2014 at 8:21 AM, Apps Embedded <app...@gm...>
>> wrote:
>>
>>> Hi,
>>>
>>> We are about to publish an Android app released under GPL v3 with full
>>> source access
>>> giving the SciPy environnement to any Android devices.
>>>
>>> We would like to call this app "Pylab Console" on a freemium model with
>>> a free and premium version.
>>> This app will give graphics support in its premium version.
>>> And thus it will use Matplotlib version 1.4.2
>>>
>>> From a legal point of view, are we able to use the term "Pylab" in our
>>> Android App name ?
>>>
>>> Moreover, as Python is a trademark, should we need to ask the PSF the
>>> authorisation of publishing such an app ?
>>>
>>> Thanks for your help.
>>>
>>> Apps Embedded Team.
>>>
>>>
>>> ------------------------------------------------------------------------------
>>>
>>> _______________________________________________
>>> Matplotlib-devel mailing list
>>> Mat...@li...
>>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>>
>>>
>>
>>
>> --
>> Thomas Caswell
>> tca...@gm...
>>
>
>

-- 
Thomas Caswell
tca...@gm...

Hi,

Could I name this app "Scipy Console Free" ?

In the BSD licence of Scipy library, it is said
"...

   - Neither the name of Enthought nor the names of the SciPy Developers
   may be used to endorse or promote products derived from this software
   without specific prior written permission.

..."

So Scipy Console shoud be ok ? Since Scipy does not seem to be forbidden by
the licence text.

Thanks for your help.

Apps Embedded Team.

2014-11-07 15:03 GMT+01:00 Thomas Caswell <tca...@gm...>:

> I (personally) would prefer you not use the work 'pylab' in your app
> name.  There is enough confusion related to the term 'pylab' as it is with
> out adding another (some what orthogonal) meaning.
>
> I am speaking for my self in this email.
>
> Tom
>
> On Fri, Nov 7, 2014 at 8:21 AM, Apps Embedded <app...@gm...>
> wrote:
>
>> Hi,
>>
>> We are about to publish an Android app released under GPL v3 with full
>> source access
>> giving the SciPy environnement to any Android devices.
>>
>> We would like to call this app "Pylab Console" on a freemium model with a
>> free and premium version.
>> This app will give graphics support in its premium version.
>> And thus it will use Matplotlib version 1.4.2
>>
>> From a legal point of view, are we able to use the term "Pylab" in our
>> Android App name ?
>>
>> Moreover, as Python is a trademark, should we need to ask the PSF the
>> authorisation of publishing such an app ?
>>
>> Thanks for your help.
>>
>> Apps Embedded Team.
>>
>>
>> ------------------------------------------------------------------------------
>>
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>
>>
>
>
> --
> Thomas Caswell
> tca...@gm...
>

I (personally) would prefer you not use the work 'pylab' in your app name.
There is enough confusion related to the term 'pylab' as it is with out
adding another (some what orthogonal) meaning.

I am speaking for my self in this email.

Tom

On Fri, Nov 7, 2014 at 8:21 AM, Apps Embedded <app...@gm...>
wrote:

> Hi,
>
> We are about to publish an Android app released under GPL v3 with full
> source access
> giving the SciPy environnement to any Android devices.
>
> We would like to call this app "Pylab Console" on a freemium model with a
> free and premium version.
> This app will give graphics support in its premium version.
> And thus it will use Matplotlib version 1.4.2
>
> From a legal point of view, are we able to use the term "Pylab" in our
> Android App name ?
>
> Moreover, as Python is a trademark, should we need to ask the PSF the
> authorisation of publishing such an app ?
>
> Thanks for your help.
>
> Apps Embedded Team.
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
>

-- 
Thomas Caswell
tca...@gm...

Hi,

We are about to publish an Android app released under GPL v3 with full
source access
giving the SciPy environnement to any Android devices.

We would like to call this app "Pylab Console" on a freemium model with a
free and premium version.
This app will give graphics support in its premium version.
And thus it will use Matplotlib version 1.4.2

>From a legal point of view, are we able to use the term "Pylab" in our
Android App name ?

Moreover, as Python is a trademark, should we need to ask the PSF the
authorisation of publishing such an app ?

Thanks for your help.

Apps Embedded Team.

I removed the rgb2lab_local branch now (I decided that this is not the way
to go and I have a local copy in my own remote of this). On a related note
should be consider removing tags for old release candidates? I know that
IPython does this and it does clean up the tags quite a bit since
approximately half the tags are for release candidates.

Jens

On Sat, Nov 1, 2014 at 7:56 PM, Thomas Caswell <tca...@gm...> wrote:

> This is done now.  All of the branches were fully merged except for v1.1.x
> which had a single line change to contents.rst which ended up on the main
> branch through other means.
>
> I have local branches pointing to all of the removed branches so if there
> is panic about their removal and _everyone_ runs a prune command on the
> upstream repos we still have this information around.
>
> I left rgb2lab_local because there is still an open PR against it, but
> will go away when we close that PR.
>
> Tom
>
> On Sat Nov 01 2014 at 2:32:19 PM Eric Firing <ef...@ha...> wrote:
>
>> On 2014/11/01, 5:49 AM, Thomas Caswell wrote:
>> > Does anyone protest to removing all of the branches from the main repo
>> > except:
>> >
>> >   - master
>> >   - v1.4.x
>> >   - v1.4.2-doc
>> >
>> > Having old branches around can lead to confusion (see
>> > https://github.com/matplotlib/matplotlib/pull/3748#issuecomm
>> ent-61372162).
>> >
>> > Tom
>>
>> Seems to me like a good idea.
>>
>> Eric
>>
>>
>> ------------------------------------------------------------
>> ------------------
>> _______________________________________________
>> Matplotlib-devel mailing list
>> Mat...@li...
>> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>>
>
>
> ------------------------------------------------------------------------------
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>
>

Hi,

I have a piece of code that graphs streaming data in a new window. If the
cursor is within the graphing window and the cursor has been moved after the
command has been sent to start graphing data, the graphing window would
crash and the message says that "python.exe has stopped working". 
In the CMD window, the following error is displayed:
"TclStackFree: incorrect freePtr. Call out of sequence?
This application has requested the Runtime to terminate it in an unusual
way. 
Please contact the application's support team for more information." 
I've tried changing from python 2.7.7 to 2.7.8 and matplotlib-1.3.1 to
1.4.2. The problem persists in both cases.
Any help or insight would be greatly appreciated. 
Thank you. 

Python code used pertaining to graphing streaming data:

def start():
	global ekg_data
	global ekg_lock
	ekg_lock = Lock()

	ekg_data = Queue()

	conn.register_notification(Handle_EKG, Handle_EKG_Notify,
callback=ekg_data_ready)
	conn.GATT_WriteCharValue(handle=Handle_StartSweep, value=StartEKG_Value)

	ylim([0,128])
	ion()
	show()

	t = Thread(target=ekg_plotter)
	t.daemon = True
	t.start()

def ekg_plotter():
	global ekg_data
	global ekg_lock
	graph_data = zeros(2000)
	graph, = plot(graph_data)
	while (True):
		newData = []
		with ekg_lock:
			size_q = ekg_data.get()
			graph_data = np.concatenate((graph_data[(len(size_q)):], size_q))	
			ekg_data.empty()
		graph.set_ydata(graph_data)
		pause(.01) 

def ekg_data_ready(packet_dictionary):
	global ekg_data
	global ekg_lock
	packet, dictionary = packet_dictionary
	data = dictionary["values"][0]
	dataList = [data[i:i+EKG_DATA_SIZE] for i in range(0, len(data),
EKG_DATA_SIZE)]
	parsedList = [unpack("<b", x)[0] for x in dataList]
	with ekg_lock:
		ekg_data.put(parsedList)

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/TclStackFree-incorrect-freePtr-Call-out-of-sequence-tp44287.html
Sent from the matplotlib - devel mailing list archive at Nabble.com.

Dear all,

I made a pull request for this: #3753, at

https://github.com/matplotlib/matplotlib/pull/3753

Cheers,
Fabio

On 11/03/2014 01:30 PM, Fabio Zanini wrote:
> @Pierre: Yeah, my code looks 99% the same like yours. I'll make a PR
> starting from a mix of both - probably closer to yours than to mine :-P
> 
> @Jens: symlog extends to infinity, logit has a compact range, that's why
> it's not exactly the same.
> 
> Ok I'll make a pull request. Shall I start with a test that fails, then
> add the function and show it succeeds? or shall I start directly with
> the code and add a test later on?
> 
> Cheers,
> Fabio
> 
> On 11/03/2014 12:58 PM, Thomas Caswell wrote:
>> Please create a pull request.
>>
>> This sounds reasonable to me, but I have never seen a plot with that
>> scale and don't really understand it from your description. Seeing the
>> code usually clarifies things.
>>
>> Tom
>>
>> On Mon, Nov 3, 2014, 05:58 Fabio Zanini <fab...@tu...
>> <mailto:fab...@tu...>> wrote:
>>
>>     Dear all,
>>
>>     I've been using matplotlib with great satisfaction for a few years, but
>>     one feature I've been missing is a "logit" scale. This is essentially a
>>     nonlinear scale that is log both towards 0+ and log towards 1-. It is
>>     useful when one has frequencies in a population (i.e. floats between 0
>>     and 1) and both rare events and very common ones are interesting.
>>
>>     For instance, say you ask about the fraction of people with blue eyes in
>>     various world populations, you want to spot even tiny deviations from
>>     zero or one.
>>
>>     I have coded a scale according to matplotlib's documentation and it
>>     works well, so I was wondering whether you are interested into merging
>>     it into the the main repository. I think it'd be useful because lots of
>>     people have such frequency data, especially now that matplotlib is
>>     becoming popular in the biology/social sciences research communities.
>>
>>     If there is interest, I'll just start a pull request on github and try
>>     to adapt the code to your coding style. It's already PEP8 and similia.
>>
>>     Thanks. Cheers,
>>     Fabio
>>
>>     ------------------------------__------------------------------__------------------
>>     _________________________________________________
>>     Matplotlib-devel mailing list
>>     Matplotlib-devel@lists.__sourceforge.net
>>     <mailto:Mat...@li...>
>>     https://lists.sourceforge.net/__lists/listinfo/matplotlib-__devel
>>     <https://lists.sourceforge.net/lists/listinfo/matplotlib-devel>
>>
> 

@Pierre: Yeah, my code looks 99% the same like yours. I'll make a PR
starting from a mix of both - probably closer to yours than to mine :-P

@Jens: symlog extends to infinity, logit has a compact range, that's why
it's not exactly the same.

Ok I'll make a pull request. Shall I start with a test that fails, then
add the function and show it succeeds? or shall I start directly with
the code and add a test later on?

Cheers,
Fabio

On 11/03/2014 12:58 PM, Thomas Caswell wrote:
> Please create a pull request.
> 
> This sounds reasonable to me, but I have never seen a plot with that
> scale and don't really understand it from your description. Seeing the
> code usually clarifies things.
> 
> Tom
> 
> On Mon, Nov 3, 2014, 05:58 Fabio Zanini <fab...@tu...
> <mailto:fab...@tu...>> wrote:
> 
>     Dear all,
> 
>     I've been using matplotlib with great satisfaction for a few years, but
>     one feature I've been missing is a "logit" scale. This is essentially a
>     nonlinear scale that is log both towards 0+ and log towards 1-. It is
>     useful when one has frequencies in a population (i.e. floats between 0
>     and 1) and both rare events and very common ones are interesting.
> 
>     For instance, say you ask about the fraction of people with blue eyes in
>     various world populations, you want to spot even tiny deviations from
>     zero or one.
> 
>     I have coded a scale according to matplotlib's documentation and it
>     works well, so I was wondering whether you are interested into merging
>     it into the the main repository. I think it'd be useful because lots of
>     people have such frequency data, especially now that matplotlib is
>     becoming popular in the biology/social sciences research communities.
> 
>     If there is interest, I'll just start a pull request on github and try
>     to adapt the code to your coding style. It's already PEP8 and similia.
> 
>     Thanks. Cheers,
>     Fabio
> 
>     ------------------------------__------------------------------__------------------
>     _________________________________________________
>     Matplotlib-devel mailing list
>     Matplotlib-devel@lists.__sourceforge.net
>     <mailto:Mat...@li...>
>     https://lists.sourceforge.net/__lists/listinfo/matplotlib-__devel
>     <https://lists.sourceforge.net/lists/listinfo/matplotlib-devel>
> 

A pull request would be welcome. Note that we already have the symlog
scale
http://matplotlib.org/examples/pylab_examples/symlog_demo.html?highlight=symlog
which is essentially a logscale with a linear fraction in the centre. As I
understand this it is not quite the same but useful for a different purpose

Jens
On Mon, Nov 3, 2014 at 12:04 PM, Pierre Haessig <pie...@cr...>
wrote:

> Hi Fabio,
>
> Le 03/11/2014 11:41, Fabio Zanini a écrit :
> > I've been using matplotlib with great satisfaction for a few years, but
> > one feature I've been missing is a "logit" scale. This is essentially a
> > nonlinear scale that is log both towards 0+ and log towards 1-.
>
> Nice coincidence, I was also playing with this idea back in April.
> http://nbviewer.ipython.org/gist/pierre-haessig/7e3e6a818edeb6819708
>
> I was actually wondering whether this could interest other people in
> other fields...
>
> best,
> Pierre
>
> >   It is
> > useful when one has frequencies in a population (i.e. floats between 0
> > and 1) and both rare events and very common ones are interesting.
> >
> > For instance, say you ask about the fraction of people with blue eyes in
> > various world populations, you want to spot even tiny deviations from
> > zero or one.
> >
> > I have coded a scale according to matplotlib's documentation and it
> > works well, so I was wondering whether you are interested into merging
> > it into the the main repository. I think it'd be useful because lots of
> > people have such frequency data, especially now that matplotlib is
> > becoming popular in the biology/social sciences research communities.
> >
> > If there is interest, I'll just start a pull request on github and try
> > to adapt the code to your coding style. It's already PEP8 and similia.
> >
> > Thanks. Cheers,
> > Fabio
> >
> >
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

Hi Fabio,

Le 03/11/2014 11:41, Fabio Zanini a écrit :
> I've been using matplotlib with great satisfaction for a few years, but
> one feature I've been missing is a "logit" scale. This is essentially a
> nonlinear scale that is log both towards 0+ and log towards 1-.

Nice coincidence, I was also playing with this idea back in April.
http://nbviewer.ipython.org/gist/pierre-haessig/7e3e6a818edeb6819708

I was actually wondering whether this could interest other people in 
other fields...

best,
Pierre

>   It is
> useful when one has frequencies in a population (i.e. floats between 0
> and 1) and both rare events and very common ones are interesting.
>
> For instance, say you ask about the fraction of people with blue eyes in
> various world populations, you want to spot even tiny deviations from
> zero or one.
>
> I have coded a scale according to matplotlib's documentation and it
> works well, so I was wondering whether you are interested into merging
> it into the the main repository. I think it'd be useful because lots of
> people have such frequency data, especially now that matplotlib is
> becoming popular in the biology/social sciences research communities.
>
> If there is interest, I'll just start a pull request on github and try
> to adapt the code to your coding style. It's already PEP8 and similia.
>
> Thanks. Cheers,
> Fabio
>
>

Please create a pull request.

This sounds reasonable to me, but I have never seen a plot with that scale
and don't really understand it from your description. Seeing the code
usually clarifies things.

Tom

On Mon, Nov 3, 2014, 05:58 Fabio Zanini <fab...@tu...>
wrote:

> Dear all,
>
> I've been using matplotlib with great satisfaction for a few years, but
> one feature I've been missing is a "logit" scale. This is essentially a
> nonlinear scale that is log both towards 0+ and log towards 1-. It is
> useful when one has frequencies in a population (i.e. floats between 0
> and 1) and both rare events and very common ones are interesting.
>
> For instance, say you ask about the fraction of people with blue eyes in
> various world populations, you want to spot even tiny deviations from
> zero or one.
>
> I have coded a scale according to matplotlib's documentation and it
> works well, so I was wondering whether you are interested into merging
> it into the the main repository. I think it'd be useful because lots of
> people have such frequency data, especially now that matplotlib is
> becoming popular in the biology/social sciences research communities.
>
> If there is interest, I'll just start a pull request on github and try
> to adapt the code to your coding style. It's already PEP8 and similia.
>
> Thanks. Cheers,
> Fabio
>
> ------------------------------------------------------------
> ------------------
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

Dear all,

I've been using matplotlib with great satisfaction for a few years, but
one feature I've been missing is a "logit" scale. This is essentially a
nonlinear scale that is log both towards 0+ and log towards 1-. It is
useful when one has frequencies in a population (i.e. floats between 0
and 1) and both rare events and very common ones are interesting.

For instance, say you ask about the fraction of people with blue eyes in
various world populations, you want to spot even tiny deviations from
zero or one.

I have coded a scale according to matplotlib's documentation and it
works well, so I was wondering whether you are interested into merging
it into the the main repository. I think it'd be useful because lots of
people have such frequency data, especially now that matplotlib is
becoming popular in the biology/social sciences research communities.

If there is interest, I'll just start a pull request on github and try
to adapt the code to your coding style. It's already PEP8 and similia.

Thanks. Cheers,
Fabio

This is done now.  All of the branches were fully merged except for v1.1.x
which had a single line change to contents.rst which ended up on the main
branch through other means.

I have local branches pointing to all of the removed branches so if there
is panic about their removal and _everyone_ runs a prune command on the
upstream repos we still have this information around.

I left rgb2lab_local because there is still an open PR against it, but will
go away when we close that PR.

Tom

On Sat Nov 01 2014 at 2:32:19 PM Eric Firing <ef...@ha...> wrote:

> On 2014/11/01, 5:49 AM, Thomas Caswell wrote:
> > Does anyone protest to removing all of the branches from the main repo
> > except:
> >
> >   - master
> >   - v1.4.x
> >   - v1.4.2-doc
> >
> > Having old branches around can lead to confusion (see
> > https://github.com/matplotlib/matplotlib/pull/3748#issuecomment-61372162
> ).
> >
> > Tom
>
> Seems to me like a good idea.
>
> Eric
>
>
> ------------------------------------------------------------
> ------------------
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

On 2014/11/01, 5:49 AM, Thomas Caswell wrote:
> Does anyone protest to removing all of the branches from the main repo
> except:
>
>   - master
>   - v1.4.x
>   - v1.4.2-doc
>
> Having old branches around can lead to confusion (see
> https://github.com/matplotlib/matplotlib/pull/3748#issuecomment-61372162).
>
> Tom

Seems to me like a good idea.

Eric

I think this is reasonable. If anything, I *might* want to keep N+1
branches, but that might be too much.

On Sat, Nov 1, 2014 at 12:31 PM, Joel B. Mohler <jo...@ki...>
wrote:

> On 11/01/2014 11:49 AM, Thomas Caswell wrote:
> > Does anyone protest to removing all of the branches from the main repo
> > except:
> >
> >  - master
> >  - v1.4.x
> >  - v1.4.2-doc
> >
>
> As a person who was distracted by too many branches in that list, I
> fully support this.
>
> Joel
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

On 11/01/2014 11:49 AM, Thomas Caswell wrote:
> Does anyone protest to removing all of the branches from the main repo 
> except:
>
>  - master
>  - v1.4.x
>  - v1.4.2-doc
>

As a person who was distracted by too many branches in that list, I 
fully support this.

Joel

Todd,

The tags would not go anywhere, just the branches would be removed.

The tags are for releases, branches are for active work and there should be
no active work on anything but those three branches.

Tom

On Sat, Nov 1, 2014, 12:14 Todd <tod...@gm...> wrote:

>
> On Nov 1, 2014 4:49 PM, "Thomas Caswell" <tca...@gm...> wrote:
> >
> > Does anyone protest to removing all of the branches from the main repo
> except:
> >
> >  - master
> >  - v1.4.x
> >  - v1.4.2-doc
> >
> > Having old branches around can lead to confusion (see
> https://github.com/matplotlib/matplotlib/pull/3748#issuecomment-61372162).
> >
> > Tom
>
> What about having a "matplotlib/matplotlib-archive" fork where old
> branches and tags can be kept?  That way someone can always go back and
> install an old version if they really need to.
> ------------------------------------------------------------
> ------------------
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>

On Nov 1, 2014 4:49 PM, "Thomas Caswell" <tca...@gm...> wrote:
>
> Does anyone protest to removing all of the branches from the main repo
except:
>
>  - master
>  - v1.4.x
>  - v1.4.2-doc
>
> Having old branches around can lead to confusion (see
https://github.com/matplotlib/matplotlib/pull/3748#issuecomment-61372162).
>
> Tom

What about having a "matplotlib/matplotlib-archive" fork where old branches
and tags can be kept?  That way someone can always go back and install an
old version if they really need to.

2003	Jan	Feb	Mar	Apr	May	Jun	Jul	Aug	Sep	Oct (1)	Nov (33)	Dec (20)
2004	Jan (7)	Feb (44)	Mar (51)	Apr (43)	May (43)	Jun (36)	Jul (61)	Aug (44)	Sep (25)	Oct (82)	Nov (97)	Dec (47)
2005	Jan (77)	Feb (143)	Mar (42)	Apr (31)	May (93)	Jun (93)	Jul (35)	Aug (78)	Sep (56)	Oct (44)	Nov (72)	Dec (75)
2006	Jan (116)	Feb (99)	Mar (181)	Apr (171)	May (112)	Jun (86)	Jul (91)	Aug (111)	Sep (77)	Oct (72)	Nov (57)	Dec (51)
2007	Jan (64)	Feb (116)	Mar (70)	Apr (74)	May (53)	Jun (40)	Jul (519)	Aug (151)	Sep (132)	Oct (74)	Nov (282)	Dec (190)
2008	Jan (141)	Feb (67)	Mar (69)	Apr (96)	May (227)	Jun (404)	Jul (399)	Aug (96)	Sep (120)	Oct (205)	Nov (126)	Dec (261)
2009	Jan (136)	Feb (136)	Mar (119)	Apr (124)	May (155)	Jun (98)	Jul (136)	Aug (292)	Sep (174)	Oct (126)	Nov (126)	Dec (79)
2010	Jan (109)	Feb (83)	Mar (139)	Apr (91)	May (79)	Jun (164)	Jul (184)	Aug (146)	Sep (163)	Oct (128)	Nov (70)	Dec (73)
2011	Jan (235)	Feb (165)	Mar (147)	Apr (86)	May (74)	Jun (118)	Jul (65)	Aug (75)	Sep (162)	Oct (94)	Nov (48)	Dec (44)
2012	Jan (49)	Feb (40)	Mar (88)	Apr (35)	May (52)	Jun (69)	Jul (90)	Aug (123)	Sep (112)	Oct (120)	Nov (105)	Dec (116)
2013	Jan (76)	Feb (26)	Mar (78)	Apr (43)	May (61)	Jun (53)	Jul (147)	Aug (85)	Sep (83)	Oct (122)	Nov (18)	Dec (27)
2014	Jan (58)	Feb (25)	Mar (49)	Apr (17)	May (29)	Jun (39)	Jul (53)	Aug (52)	Sep (35)	Oct (47)	Nov (110)	Dec (27)
2015	Jan (50)	Feb (93)	Mar (96)	Apr (30)	May (55)	Jun (83)	Jul (44)	Aug (8)	Sep (5)	Oct	Nov (1)	Dec (1)
2016	Jan	Feb	Mar (1)	Apr	May	Jun (2)	Jul	Aug (3)	Sep (1)	Oct (3)	Nov	Dec
2017	Jan	Feb (5)	Mar	Apr	May	Jun	Jul (3)	Aug	Sep (7)	Oct	Nov	Dec
2018	Jan	Feb	Mar	Apr	May	Jun	Jul (2)	Aug	Sep	Oct	Nov	Dec

matplotlib-devel Mailing List for matplotlib (Page 25)

matplotlib-devel — matplotlib developers