[Kst] [Bug 221749] Clearer font size definition

Barth Netterfield netterfield at astro.utoronto.ca
Fri Jan 8 14:28:38 CET 2010


Based on your comments, here is my new proposal:
	-Remove reference font size from the settings.  All font sizes will be 
absolute relative to the reference page size.
	-Change the default reference page size to Letter or A4, and have a pull down 
list of common sizes (maybe Letter, A4, 1 column Journal plot, Custom).
	-Add 'increase all fonts' and 'decrease all fonts' actions to scale 
everything.
	-Add check box options to re-scale all fonts when plots are added to a tab.

On Friday 08 January 2010 05:26:55 Nicolas Brisset wrote:
> 1) I find it is nice if the fonts scale down when the number of plots
>  increases (resp. when the size of a plot decreases, for whatever reason:
>  maximze, window resized, plot resized, etc...) -> a completely fixed font
>  size (S) does not sound good

Kst 1.x did this.  The problem is that it became very difficult to make fonts 
the same size if the plots were different sizes, as font sizes were relative 
to the plot not the window.

The compromise I have tried with the data wizard is that if you add plots to a 
window that already has plots, it resizes the fonts of all plots in the tab.  
This is dangerous if someone has already crafted font sizes, but often does 
what you want.

Suggestion: have a check box wherever you add plots to a tab to optionally 
resize the fonts of all plots in the tab, but keep the general calculations 
relative to the page, not the plot, so consistency remains easy.

> 2) my biggest gripe against the current system is that changing fonts
>  sometimes requires a lot of trial and error and results in quite some
>  confusion, as described in bug #109469. 

This is mainly do to the reference font size issue.

>  In fact, the relative font size
>  concept is hard to grasp, especially when it has no effect (i.e. the user
>  expects that changing the value in the spin box will change the font size
>  by one point, and it's not what happens) and the value in the spinbox
>  can't be correlated to anything (has no "unit" if I dare say so)

It is points.  P = S + D.  But on some systems, the font sizes are/were 
quantized quite coarsely in Qt, so you may not see a change.  Because I 
intimately understand what is going on, it is easy for me to understand what 
happens, but the fact that, as far as I can tell, no one else does informs me 
that the interface is pretty broken :-)

> 3) font sizes can't be decreased infinitely -> we need to keep a minimum
>  font size (preferably set to something that should suit all, but possibly
>  that does not exist and should be settable from the UI as today)

It is in the setting dialog along with the reference font size and reference 
plot size.

> 4) in the typical KDE4 approach, I'd say we should strive to find good
>  defaults and remove some confusing options from the UI (e.g. the reference
>  view size, with apparently defaults to 12 cm x 16 cm: ???). 

A reference plot size is absolutely critical for publication.  You want to say 
that when the plot is published as a 12x16 cm plot in a journal, the font will 
be 12 pt, or whatever.  The default reference size has not had much thought 
though.  IIRC, the size I used what from taking a ruler to a plot in a paper 
that happened to be on my desk.

>  I think there
>  are a couple of things kst should be able to find out by itself: whether
>  we are printing (and then on which media size), or displaying on screen
>  (and then with which resolution). Based on that it should be feasible to
>  determine an appropriate font size. Maybe only offer the user an option
>  between small fonts / medium fonts / large fonts?

> 5) in the end I think it is a matter of proportions. Instead of specifying
>  a font size, we should determine how big a font must be relative to the
>  plot, and compute the font size from that of the plot. 
> One thing which is not clear to me is what we should do for windows where
>  the plots have different sizes: I believe it would be better to have
>  homogeneous font sizes, but that breaks a bit the central concept in point
>  5)...

In 2.x, size is relative to the page not the plot.  This is important I think.


More information about the Kst mailing list