Re: [libreoffice-users] Re: precision of Draw dimensions - 3 place

On 19/11/15 15:39, V Stuart Foote wrote:
> No, the all measurements in the StarOffice -> OpenOffice Draw/Impress legacy
> are derived from raster screen bitmap graphics which has maintained the text
> "print" centric practice and uses a smallest "Field_Unit" of twips for
> calculations of placement onto the canvas.
>
> A twip  (1/1440") --from Postscript 1/72" point-- is a screen independent
> value converted (and rounded)  in calculating vertex placements as pixel
> dimensions.  There are routines to convert measurement from twip to mm or
> inch--but they are equally imprecise for use in CAD.
>
> Meaning there is no history for vector based measurements (vertex, angles,
> radius, and distances) needed for CAD. What is provided is all geared to
> rendering bitmap onto a raster canvas--for  resample to screen pixels.
>
> So, the "arbitrary" lack of precision is a function of the resampling to
> "fit" bitmap elements to pixels of a raster screen. wo Floating point
> precision of measurements remains much less significant than performance for
> rendering the twip derived document canvas. Yes there was a tradeoff, but an
> easy one to make.
>
> If you need the precision for CAD you want a program designed to handle
> vectors under the hood, not bitmaps--period!


Whatever the internals within LO, the present behaviour is irrational. 
Yes, obviously everything /eventually/ has to be adjusted some arbitrary 
grid size. However, there's no reason at all for arbitrary roundings 
before printing.

In particular, I've just experimented to see what's stored. I draw a 
single line and position it (or try to!) at 1.234cm from the margin 
(itself set at 1cm). I do this twice, with units set to cm (which forces 
a rounding to 2.34cm) and them mm (retains 2 dp's in/mm/ this time).

The generated contents.xml has a draw:line element, whose position is in 
the first case svg:x1="2.23cm" and in the second svg:x1="2.234cm". So the

They really, really ought to be the same. If the user chooses to work in 
cm, mm, or cubits, the available physical precision should nevertheless 
be ultimately defined by what the printing system can do.

OTOH, if the stored measurements are rounded to some arbitrary grid, 
then my two files should have the same rounded content.

Final experiment. Set the units to feet - and see how the precision 
drops to about 3mm because the entries are rounded by the GUI to 2 dp in 
feet. Rational behaviour? Not from a user's POV.



--
Mike Scott (unet2 <at> [deletethis] scottsonline.org.uk)
Harlow Essex England

--
To unsubscribe e-mail to: users+unsubscribe@global.libreoffice.org
Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/
Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette
List archive: http://listarchives.libreoffice.org/global/users/
All messages sent to this list will be publicly archived and cannot be deleted