xyplot ticks in 0.86 - WHAT HAS HAPPENED?

Michal Szymanski (msz@ogle.ctio.noao.edu)
Mon, 24 Mar 1997 22:25:50 -0300 (CDT)

To subscribers of the xforms list from msz@ogle.ctio.noao.edu (Michal Szymanski) :

Hi XFormers,

Whoops, I changed to 0.86 and now I am thinking about going back to 0.81.

One of my most used apps using XForms is an FTS image display. As one of
its functions it displays a profile of a star, using 'xyplot' object.

Now, in the 'Changes' file I found: "better tics generation"
Well, I find this 'better' very peculiar. Here is an example:

with bounds and ticks set with:

fl_set_xyplot_xbounds(plot, -0.3, 8.3);
fl_set_xyplot_ybounds(plot, -1554.2, 14960.2); /* computed values !!! */
fl_set_xyplot_xtics(plot, 9, 1);
fl_set_xyplot_ytics(plot, 8, 1);

I get following labelling of the ticks:

XForms 0.81:
X: 0.0 1.0 ... 8.0
Y: 0 2e+03 4e+03 6e+03 8e+03 1e+04 12000 14000

XForms 0.86:
X: -0.3 0.7 1.7 ... 7.7
Y: -2e+03 4e+02 2e+03 4e+03 6e+03 8e+03 10445.8 12445.8 14445.8

I have never understood why it is writing 2e+03 instead of 2000
(especially if it writes 12000), or even more tricky 1e+04 instead of
10000, but I got used to it. But what I get now is really UNACCEPTABLE.

TC, please add 'fl_set_xyplot_ticks_WORSE' function for those poor guys
who do not want to have them 'better' ;-)))

Speaking more seriously, I think either the ticks generation algorithm
should be fixed (the previous one was not so bad, after all) or
(MUCH BETTER) there should be a possibility to set not only the number
of major ticks but also their start point and the step. This should be
easy task as one has just to skip the automatic scaling routine and
supply user-defined values, possibly checking if they are not out of
bounds.

Another small thing: Interpretation of the 'align' parameter in
fl_add_xyplot_tex() has changed which is not mentioned in 'Changes'.
In 0.81 I used FL_ALIGN_LEFT to get labels starting just to the right
side of (x,y) point - which seems intuitive, as the labels written at
(x, y), (x, y+dy), (x, y+2*dy) were "left-aligned". Now with the same
parameter the labels end to the left of (x,y) point, getting actually
"right-aligned". FL_ALIGN_RIGHT does in 0.86 what FL_ALIGN_LEFT did in
0.81.

regards, Michal.

****************************************************************************
Michal Szymanski Internet: msz@sirius.astrouw.edu.pl
Warsaw University Observatory Bitnet: msz@plwauw61.bitnet
Al. Ujazdowskie 4 Voice: 48-22-6294011 ext 23
00-478 Warszawa, POLAND FAX: 48-22-6294967
****************************************************************************
_________________________________________________
To unsubscribe, send the message "unsubscribe" to
xforms-request@bob.usuf2.usuhs.mil or see
http://bob.usuf2.usuhs.mil/mailserv/xforms.html
Xforms Home Page: http://bragg.phys.uwm.edu/xforms
List Archive: http://bob.usuf2.usuhs.mil/mailserv/list-archives/xforms-archive/