Thread: [Tuxmath-devel] =?iso-8859-1?q?Bug=3A_0=F7=3F=3D0?=

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Hi all,

There's a bug in tuxmath code: AFAIR the answer of 0÷?=0 should be every
non nul number, not only one particular number.

-- 
yves

Hi Yves et al,

On Friday 27 October 2006 18:05, yv...@yc... wrote:
> Hi all,
>
> There's a bug in tuxmath code: AFAIR the answer of 0=F7?=3D0 should be ev=
ery
> non nul number, not only one particular number.

Good pickup.  I ran across something similar for multiplication when I adde=
d=20
the other question formats - questions like  "0 * ? =3D 0" and "? * 0 =3D 0=
" are=20
also indeterminate.  I worked around that by adding a couple of lines to=20
generate_list() to prevent those questions from getting into the list - it=
=20
would be very easy to do the same for the division example above.  The=20
alternative would be to leave these questions in and accept all answers as=
=20
correct, which would be doable but not quite as trivial.

On other fronts - I have been looking at the tuxtype code with the idea of=
=20
using its system of title and menu screens for tuxmath, as the options scre=
en=20
needs an overhaul.  Any objections?

I have also worked on cross-building tuxtype for Windows using the general=
=20
approach you set up for tuxmath.  I have gotten it to build a nsis-based=20
tuxtype installer, but the tuxtype program doesn't run because it fails to=
=20
load the data files - (probably a simple issue with file paths - need to lo=
ok=20
into it some more).=20

Cheers,
=2D-=20
David Bruce

Hi,

On Monday 30 October 2006 13:20, David Bruce wrote:
> Good pickup.  I ran across something similar for multiplication when I
> added the other question formats - questions like  "0 * ? =3D 0" and "? *=
 0 =3D
> 0" are also indeterminate.  I worked around that by adding a couple of
> lines to generate_list() to prevent those questions from getting into the
> list - it would be very easy to do the same for the division example abov=
e.
>  The alternative would be to leave these questions in and accept all
> answers as correct, which would be doable but not quite as trivial.

I better like the first approach, even though the second slightly has more=
=20
learning value :)

Are you planning on releasing the fix anytime soon? I guess it could still=
=20
make it into etch...

regards,
	Holger

Hi David,

On Monday 30 October 2006 06:20, David Bruce wrote:
> On other fronts - I have been looking at the tuxtype code with the idea of
> using its system of title and menu screens for tuxmath, as the options
> screen needs an overhaul.  Any objections?

I don't know exactly what you have in mind, but I very much agree that the 
options screen needs an overhaul. I've been meaning to look into this myself 
but have been busy, but I would be happy to help as much as I can & where I 
can be useful. Are you still thinking that there might be a separate GUI for 
setting the more "advanced" options, as well as for determining which options 
will be kid-settable?

>
> I have also worked on cross-building tuxtype for Windows using the general
> approach you set up for tuxmath.  I have gotten it to build a nsis-based
> tuxtype installer, but the tuxtype program doesn't run because it fails to
> load the data files - (probably a simple issue with file paths - need to
> look into it some more).

Our elementary school seems interested in tuxmath---my oldest daughter's 
teacher asked her about her rapid progress on her math facts, and she 
attributed it all to tuxmath. (Couldn't have said it better myself. Thanks 
again to all developers!) I'll be demo-ing the program to her teacher soon, 
and can use the opportunity to collect feedback on anything that any of you 
would like to ask.

Our elementary school is Mac OSX-based; does anyone know the status of the Mac 
build? I don't have a Mac to test it on myself...

Best,
--Tim

>
> Cheers,

Hi Holger,

On Monday 30 October 2006 07:55, Holger Levsen wrote:

> Are you planning on releasing the fix anytime soon? I guess it could still
> make it into etch...

I just put it in (a single line change) and it appears to work properly.

The svn commit also contains a more substantial change - Last week, I put in a 
set of small functions to change the game_options struct rather than having 
other parts of the program change the struct directly (making it more like a 
C++ class).  This fixes a number of ways in which it was possible to crash 
the program by feeding it impossible settings in the options file.

Cheers,

-- 
David Bruce

Hi David,

On Monday 30 October 2006 17:59, David Bruce wrote:
> I just put it in (a single line change) and it appears to work properly.
>
> The svn commit also contains a more substantial change - Last week, I put
> in a set of small functions to change the game_options struct rather than
> having other parts of the program change the struct directly (making it
> more like a C++ class).  This fixes a number of ways in which it was
> possible to crash the program by feeding it impossible settings in the
> options file.

Great. Please tell me when a new version is ready which should be uploaded =
to=20
debian. I think, now is a good moment (before the rewrite of the menu=20
system), but you should do a release first :)

regards,
	Holger

Hi Tim,

On Monday 30 October 2006 09:21, Tim Holy wrote:

> I don't know exactly what you have in mind, but I very much agree that the
> options screen needs an overhaul.

TuxType 2 has a nice-looking menu system for selecting options, and the code 
supports the screens for a hierarchical menu without having to write each 
individual screen from scratch.  It would be nice for TuxMath and TuxType to 
have a more unified appearance. 

> Are you still thinking that there might be a
> separate GUI for setting the more "advanced" options, as well as for
> determining which options will be kid-settable?

Yves Combe has written a quick Qt4-based config program that can be built for 
either Windows or Linux.  The code is at his web site 
(http://www.ycombe.net/tuxmath).  It is sort of a "proof of concept" program 
at this point.  I perhaps could put the code into the tuxmath svn repo under 
something like "trunk-configurator".

> Our elementary school seems interested in tuxmath---my oldest daughter's
> teacher asked her about her rapid progress on her math facts, and she
> attributed it all to tuxmath. (Couldn't have said it better myself. Thanks
> again to all developers!) I'll be demo-ing the program to her teacher soon,
> and can use the opportunity to collect feedback on anything that any of you
> would like to ask.

Great!
>
> Our elementary school is Mac OSX-based; does anyone know the status of the
> Mac build? I don't have a Mac to test it on myself...

I have been asking around for an old Mac to try to get a OSX build together, 
so far with no luck.  Thanks to Yves, I can cross-build Tuxmath and (almost) 
Tuxtype for Windows on my Debian system.  If anyone knows how to cross-build 
for MacOSX under Linux, I would be interested in learning more.

Someday, I would like to have everything set up and scripted  to build 
TuxMath, TuxType, and TuxPaint from source and spit out an ISO image 
with .debs, .rpms, Windows and Mac binaries, and (of course) all of the 
source, so I could just hand out an all-in-one Tux4Kids CD to whomever is 
interested.

-- 
David Bruce

On Mon, Oct 30, 2006 at 12:18:30PM -0500, David Bruce wrote:
> Someday, I would like to have everything set up and scripted  to build 
> TuxMath, TuxType, and TuxPaint from source and spit out an ISO image 
> with .debs, .rpms, Windows and Mac binaries, and (of course) all of the 
> source, so I could just hand out an all-in-one Tux4Kids CD to whomever is 
> interested.

That'd be sweet.  If you need anything on the makefile end of things in
Tux Paint, feel free to send me patches, or ask me for CVS access on
SF.net so you can poke at things.

Some useful folks are on the tuxpaint-devel list, too:

  John Popplewell - Win32 installer and ZIP builds
  TOYAMA Shin-Ichi - Fedora CORE and Red Hat RPM (x86) builds
  Ben Armstrong - Debian maintainer (as in, he's part of Debian)
  Martin Fuhrer - Mac OS X 10.3+ Universal (PPC/x86) builds

They may have feedback on how to script all this, if it's possible.

-- 
-bill!
bi...@ne...
http://www.newbreedsoftware.com/

Thread: [Tuxmath-devel] =?iso-8859-1?q?Bug=3A_0=F7=3F=3D0?=

tuxmath-devel