Mail Archives: geda-user/2015/09/15/00:51:32
--089e010d82126f23a4051fc1ed33
Content-Type: text/plain; charset=UTF-8
On Mon, Sep 14, 2015 at 4:40 PM, John Doty <jpd AT noqsi DOT com> wrote:
>
> On Sep 14, 2015, at 3:04 PM, Peter Stuge (peter AT stuge DOT se) [via
> geda-user AT delorie DOT com] <geda-user AT delorie DOT com> wrote:
>
> > John Doty wrote:
> >>> I disagree very strongly with floating point, but using a fixed-size
> >>> decimal is an important improvement!
> >>
> >> The trouble is that common computer numerics do not actually obey the
> >> same rules as mathematical numbers.
> >
> > Nod.
> >
> >
> >> Rational numbers fix these problems.
> > ..
> >> For rendering on a grid, use fixed or floating point. The rationals
> >> that fall on your grid are a set of measure zero, anyway.
> >
> > Output (rendering on grid) is one issue, and is easy enough to deal
> > with in isolation.
> >
> > But input (rotate by 60 degrees) is another issue, and less easy to
> > handle, because it's very important for usability that user
> > input->output and vice versa is also closed.
If you define your set of permissible orientations sufficiently coarse
relative to the accuracy of the angle calculations you need to make, and
put those orientations at the center of your bins, there is no chance of a
rotate/rotate sequence putting you in a different bin.
This application doesn't need long rotation sequences, so its easy to
ensure that your calculations are sufficiently accurate with respect to
your bin size.
--089e010d82126f23a4051fc1ed33
Content-Type: text/html; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
<div dir=3D"ltr"><br><div class=3D"gmail_extra"><br><div class=3D"gmail_quo=
te">On Mon, Sep 14, 2015 at 4:40 PM, John Doty <span dir=3D"ltr"><<a hre=
f=3D"mailto:jpd AT noqsi DOT com" target=3D"_blank">jpd AT noqsi DOT com</a>></span> w=
rote:<br><blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;borde=
r-left:1px #ccc solid;padding-left:1ex"><span class=3D""><br>
On Sep 14, 2015, at 3:04 PM, Peter Stuge (<a href=3D"mailto:peter AT stuge DOT se"=
>peter AT stuge DOT se</a>) [via <a href=3D"mailto:geda-user AT delorie DOT com">geda-use=
r AT delorie DOT com</a>] <<a href=3D"mailto:geda-user AT delorie DOT com">geda-user AT d=
elorie.com</a>> wrote:<br>
<br>
> John Doty wrote:<br>
>>> I disagree very strongly with floating point, but using a fixe=
d-size<br>
>>> decimal is an important improvement!<br>
>><br>
>> The trouble is that common computer numerics do not actually obey =
the<br>
>> same rules as mathematical numbers.<br>
><br>
> Nod.<br>
><br>
><br>
>> Rational numbers fix these problems.<br>
> ..<br>
>> For rendering on a grid, use fixed or floating point. The rational=
s<br>
>> that fall on your grid are a set of measure zero, anyway.<br>
><br>
> Output (rendering on grid) is one issue, and is easy enough to deal<br=
>
> with in isolation.<br>
><br>
> But input (rotate by 60 degrees) is another issue, and less easy to<br=
>
> handle, because it's very important for usability that user<br>
> input->output and vice versa is also closed.</span></blockquote><di=
v><br></div><div style=3D"">If you define your set of permissible orientati=
ons sufficiently coarse relative to the accuracy of the angle calculations =
you need to make, and put those orientations at the center of your bins, th=
ere is no chance of a rotate/rotate sequence putting you in a different bin=
.</div><div style=3D"">This application doesn't need long rotation sequ=
ences, so its easy to ensure that your calculations are sufficiently accura=
te with respect to your bin size.</div><div><br></div></div></div></div>
--089e010d82126f23a4051fc1ed33--
- Raw text -