RE : RE : RE : [ROOT] Gradient Palette when drawing 2D histograms with "col" option

From: ANCIANT E. (
Date: Fri Mar 26 2004 - 16:32:45 MET

Hi Olivier,

> Well, that example is a bit misleading because the background color is
> equal to one of the color in the middle of the palette (the 2nd from
> bottom in fact). One can easily see that the choice of background
color is
> not really accurate.

I know, but naively I would have thought that choosing a background
colour and a palette should be uncorrelated. With the current scheme, a
palette becomes accurate / inaccurate depending on the choice of the
background and vice-versa.

>  I do not think that the current behavior is so weird because at some
> ago (one year or so) I had change it to what you are requesting (IE:
> drawing the empty bins) thinking there was no reasons to treat such
> in a different way. We had immediately MANY complaints and after a
> agreement with several unhappy users it was decided that empty bins
> not be drawn. At that time leaving them painted with the background
> appeared to be the best solution.
> Well, again it was agreed that way ... changing the behavior will be
> really catastrophic. We went through that once I do not really want to
> make it again ... (see below for further details).

OK, I understand there are two different (and incompatible ) approaches
here : "statistical" versus "signal" : If minimum value is 0 and maximum
value is 100000, "statistic" says bins with content 1 should clearly
have a colour different from 0, "signal" says bins with content 1 should
have the same colour than 0 ... (in other words : bin=0, statistic =
"empty", signal= "bin FILLED with 0" )

> Yes I agree it is that way for lego and surface, but that are 3D plots
> they clearly show the Z axis so we get immediately the idea that the
> histogram is cut

again "statistic" versus "signal" : Imagine I have a flat background
signal of value = 10 over my histogram and display without using
SetMinimum/Maximum Do I get the idea that the histogram is NOT cut ?

> whereas when drawn with COL, TEXT or BOX options the
> histogram is seen from top and if we apply your rule:
> "Values below minimum are drawn with the colour of minimum"
> some bins (in particular the empty ones) will appear populated whereas
> they are not ! (actually that is the reason of the test on zmin)

... but populated bins appear as unpopulated ..
+ problems with negative values ...

> May be before implementing it as a new option it would be useful to
> have it as a simple macro (at the end it is just a matter of looping
> the histogram bins and draw colored boxes), play a bit with it to
> understand what is needed.

I perfectly understand that such a change or new implementation is
delicate given the numerous users that already use the option ...

I guess this conversation was useful for clarifying things (at least if
was for me) between the "statistical" and "signal" points of view, ...
hope somebody will come up with an idea that would conciliate the two !

For the moment I will override the Draw methods for my classes, and
adapt the TFrame Fill colour everytime the "col" options are invoked ...



Ce message et eventuellement les pieces jointes sont exclusivement transmis a l'usage de leur destinataire. Leur contenu est strictement confidentiel. Si vous n etes pas le bon destinataire nous vous demandons de ne pas conserver, copier, utiliser ou divulguer cette communication. Merci de le detruire et de me notifier cette erreur.

INTERNET ne permettant pas d assurer l integrite de ce message, SODERN decline toute responsabilite au cas ou il aurait ete intercepte ou modifie.

This message and possibly attachments are transmit for the exclusive use of their address. Their content is strictly confidential. If you are not the intended recipient please do not keep, copy, use or disclose this communication to others. Also please delete it and notify the sender at once.

Because of the nature of INTERNET the sender is not in a position to ensure the integrity of this message, therefore SODERN disclaims any liability whatoever in the event of this message having been intercepted and/or altered.

This archive was generated by hypermail 2b29 : Sun Jan 02 2005 - 05:50:07 MET