octave-dev — GNU Octave developers' list

Re: [OctDev] Re: [Pkg-octave-devel] Bug#344995: octave-forge: FTBFS (ia64, mips, .../unstable): fixed.h: No such file or directory

[David B. <Dav...@mo...>]

Paul Kienzle wrote:

> This sounds familiar and may already be fixed in CVS, so hopefully it  
> is just a matter of releasing a new version.
>
> I think the hordes of octave-forge developers can manage this before  
> the new year.
>
> The steps are as usual:
>
> (1) post all outstanding changes to your packages
> (2) address outstanding bugs on the bug tracker
> (3) address outstanding bugs on the list, or add them to the bug tracker
> (4) purge functions which have made it into 2.1.72
> (5) test build on windows, os x, linux
>
> Please help out where you can.
>
> Also for all the octave-forge functions you are using, add some test  
> cases at the end if there are none there already so that we can 
> release  with more confidence.
>
> - Paul

Paul,

Can we delay a little bit. My reasoning is that I'd like the new 
autoload function in 2.9.x to be used by octave-forge if available to 
allow me to significantly reduce the size of my MinGW build by removing 
all of the symbolic links in octave-forge. The easiest way to do this 
would be to have "make install" diff all of the binary files in the 
directory being installed and see if they are the same and if so don't 
install it but add an autoload command to the PKG_ADD file.  You can't 
just check if it is a symbolic link, as they aren't under MinGW (which 
is the problem).

I'll try and write a script for this rapidly so that it doesn't hold up 
the release too much. As for purging functions, it is a little bit more 
complex than that, as most of the functions merged into octave from 
octave-forge went into 2.9.x release only, and so rather than purging 
these functions, they will need a conditional installation. I can attack 
some of these as well if someone doesn't get there first...

Cheers
David

-- 
David Bateman                                Dav...@mo...
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary

[OctDev] Re: [Pkg-octave-devel] Bug#344995: octave-forge: FTBFS (ia64, mips, .../unstable): fixed.h: No such file or directory

[Paul K. <pki...@us...>]

This sounds familiar and may already be fixed in CVS, so hopefully it  
is just a matter of releasing a new version.

I think the hordes of octave-forge developers can manage this before  
the new year.

The steps are as usual:

(1) post all outstanding changes to your packages
(2) address outstanding bugs on the bug tracker
(3) address outstanding bugs on the list, or add them to the bug tracker
(4) purge functions which have made it into 2.1.72
(5) test build on windows, os x, linux

Please help out where you can.

Also for all the octave-forge functions you are using, add some test  
cases at the end if there are none there already so that we can release  
with more confidence.

- Paul

On Dec 28, 2005, at 5:42 AM, Steve Langasek wrote:

> Package: octave-forge
> Version: 2005.06.13-4
> Severity: serious
>
> octave-forge is failing to build on multiple architectures because it's
> missing an include path while building in the fixed/examples directory:
>
> [...]
> make[4]: Entering directory  
> /build/buildd/octave-forge-2005.06.13/main/fixed/examples'
> Depending ffft.cc
> rm -f ffft.d
> /usr/bin/g++ -M -I/usr/include/octave-2.1.72  
> -I/usr/include/octave-2.1.72/octave -O2 -I../ -DHAVE_OCTAVE_21  
> -DHAVE_RESIZE_AND_FILL ffft.cc | /bin/sed -e  
> 's,^[^:]*/\(.*\.o\):,\1:,' -e 's,ffft\.o,pic/& & ffft.d,g' > ffft.d-t  
> && mv ffft.d-t ffft.d
> Depending fixed_inc.cc
> rm -f fixed_inc.d
> /usr/bin/g++ -M -I/usr/include/octave-2.1.72  
> -I/usr/include/octave-2.1.72/octave -O2 -I../ -DHAVE_OCTAVE_21  
> -DHAVE_RESIZE_AND_FILL fixed_inc.cc | /bin/sed -e  
> 's,^[^:]*/\(.*\.o\):,\1:,' -e 's,fixed_inc\.o,pic/& & fixed_inc.d,g' >  
> fixed_inc.d-t && mv fixed_inc.d-t fixed_inc.d
> mkoctfile -DHAVE_OCTAVE_21 -v -c ffft.cc
> /usr/bin/g++ -c -fPIC -I/usr/include/octave-2.1.72  
> -I/usr/include/octave-2.1.72/octave -O2 -DHAVE_OCTAVE_21 ffft.cc -o  
> ffft.o
> In file included from ffft.cc:30:
> ffft.h:34:19: error: fixed.h: No such file or directory
> [...]
>
> A full build log can be found at
> <http://buildd.debian.org/fetch.php?pkg=octave- 
> forge&arch=mips&ver=2005.06.13-4&stamp=1135763004&file=log>.
>
> The fixed.h file is in the parent main/fixed directory, not in
> main/fixed/examples; for whatever reason, the build is using the -I../
> argument correctly when recording the dependencies, but then proceeds  
> to
> drop it for the actual build, resulting in the shown errors.
>
> There are also other errors shown in the build log; I don't understand  
> why
> the build continues on after these other errors, but I assume at least  
> that
> the final error, if not the *only* reason for the build failure, is at  
> least
> *a* reason for the build failure.
>
> Because ginac is tied to the KDE ABI transition, and octave-forge  
> depends on
> libginac1.3c2a, it would be appreciated if you could address this bug  
> ASAP.
>
> Thanks,
> -- 
> Steve Langasek                   Give me a lever long enough and a  
> Free OS
> Debian Developer                   to set it on, and I can move the  
> world.
> vo...@de...                                    
> http://www.debian.org/
> _______________________________________________
> Pkg-octave-devel mailing list
> Pkg...@li...
> http://lists.alioth.debian.org/mailman/listinfo/pkg-octave-devel

[OctDev] SIFT image descriptor patch

[Etienne G. <et...@cs...>]

  Hi All,

this patch could be of interest to octave users who do image
processing.  You can apply it to the matlab code in siftDemoV4 [1] to
allow octave to extract SIFT image features [2] with David Lowe's
package.

  The SIFT image detector identifies interest points, somewhat like a
corner detector does, and gives them a local descriptor. The
descriptor can then be used for matching purposes (another way to
perform matching could be correlation of a small window around the
feature). Wrt to corners + local window correlation, SIFT features are
more robust to changes in illumination and viewpoint [3] and they are
increasingly used in computer vision.

  The siftDemo has functions to find sift features and to perform
matching. I've tested the patch under octave+linux, matlab+linux (it
works) and octave+win (it fails, probs w/ imagemagick + double).

  Hth,

  Etienne



[1] http://www.cs.ubc.ca/spider/lowe/keypoints/siftDemoV4.zip

  Synopsis:=20

    unzip siftDemoV4.zip;
    cd siftDemoV4;=20
    patch -p1 < mySift.patch
 =20
    In octave, make sure sift.m is in your path, then

    >> [image, descriptors, locs] =3D sift("/home/etienne/prog/various/si=
ftDemoV4/book.pgm");
   >> showkeys(image, locs)

   See also 'help match'

[2] http://www.cs.ubc.ca/~lowe/keypoints/

  David G. Lowe, =D6bject recognition from local scale-invariant
  features," International Conference on Computer Vision, Corfu,
  Greece (September 1999), pp. 1150-1157.

[3] A performance evaluation of local descriptors Krystian
    Mikolajczyk, Cordelia Schmid International Conference on Computer
    Vision & Pattern Recognition - June 2003

    http://lear.inrialpes.fr/pubs/2003/MS03/mikolajczyk_cvpr2003.pdf


--=20
Etienne Grossmann ------ http://www.cs.uky.edu/~etienne

Re: [OctDev] isunix() returns 1 under cygwin

[William P. Y. H. <wil...@gm...>]

On 12/27/05, Andy Adler <ad...@nc...> wrote:
> The semantics of 'isunix' depends on what you mean by UNIX.
> Strictly speaking, only certain well defined OSes are UNIX. Linux,
> for example, is not.
>
> On the other hand, maybe UNIX means OSes that behave like
> UNIX in most ways. cygwin has UNIX process semantics
> (ie. fork) and file semantics (symlinks, select on files, etc.)
>
> So, is cygwin UNIX? Clearly, a mingwin octave is not unix.
>
> Maybe isunix should make a specific test.
>

In Octave 2.9.4, isunix() is implemented as

function retval =3D isunix ()

  if (nargin =3D=3D 0)
    retval =3D octave_config_info ("unix");
  else
    usage ("isunix ()");
  endif

endfunction

AFAIK, it returns 1 on Cygwin (I know this by reading the sources, so
I'm not very sure)

Re: [OctDev] isunix() returns 1 under cygwin

[Bill D. <de...@se...>]

One thing that I could see as a good option for this would be for isunix=20
to have variable output arguements.  Perhaps something like [true/false,=20
osname, osversion, distro, distroversion].

Bill

On Mon, 26 Dec 2005, Etienne Grossmann wrote:

>
>  Hi Andy,
>
> thx for the answer. I should have added that I want to use isunix() to
> determine whether I want to execute a (linux) binary called (say)
> 'foo' or a windows binary called 'fooWin.exe'. So I really want isunix
> to return true iff it's Linux - my code will fail on other nixes but
> that's ok for now.
>
>  Cheers,
>
>  Etienne
>
> ps: More in detail on what I'm doing, I'm patching David Lowe's SIFT
>    [1] image feature detector code for Matlab so that it'll run both
>    on Matlab and Octave. His code runs both on Win and Linux.
>
> [1] David G. Lowe, "Distinctive image features from scale-invariant
>    keypoints,"International Journal of Computer Vision, 60, 2 (2004),
>    pp. 91-110.
>
>   http://www.cs.ubc.ca/~lowe/keypoints/
>
>
>
> On Mon, Dec 26, 2005 at 05:16:15PM -0500, Andy Adler wrote:
> # On 12/26/05, Etienne Grossmann <et...@cs...> wrote:
> # > octave:8> help isunix
> # > isunix is the user-defined function from the file
> # > /home=E6tienne/prog=F8ctave=F8ctave-forge=F8ctave-forge/main/general/=
isunix.m
> # >
> # > Always returns true.  If you are on a windows machine, be sure to
> # > put an isunix.m which always returns false in your path.
> #
> # The semantics of 'isunix' depends on what you mean by UNIX.
> # Strictly speaking, only certain well defined OSes are UNIX. Linux,
> # for example, is not.
> #
> # On the other hand, maybe UNIX means OSes that behave like
> # UNIX in most ways. cygwin has UNIX process semantics
> # (ie. fork) and file semantics (symlinks, select on files, etc.)
> #
> # So, is cygwin UNIX? Clearly, a mingwin octave is not unix.
> #
> # Maybe isunix should make a specific test.
> #
> # --
> # Andy
> #
>
> --=20
> Etienne Grossmann ------ http://www.cs.uky.edu/~etienne
>
>
>
> -------------------------------------------------------------
> Octave is freely available under the terms of the GNU GPL.
>
> Octave's home on the web:  http://www.octave.org
> How to fund new projects:  http://www.octave.org/funding.html
> Subscription information:  http://www.octave.org/archive.html
> -------------------------------------------------------------
>

--=20
If you make a mistake you right it
immediately to the best of your ability.
   -- /usr/bin/fortune

Re: [OctDev] isunix() returns 1 under cygwin

[Etienne G. <et...@cs...>]

  Hi Andy,

thx for the answer. I should have added that I want to use isunix() to
determine whether I want to execute a (linux) binary called (say)
'foo' or a windows binary called 'fooWin.exe'. So I really want isunix
to return true iff it's Linux - my code will fail on other nixes but
that's ok for now.

  Cheers,

  Etienne

ps: More in detail on what I'm doing, I'm patching David Lowe's SIFT
    [1] image feature detector code for Matlab so that it'll run both
    on Matlab and Octave. His code runs both on Win and Linux.

[1] David G. Lowe, "Distinctive image features from scale-invariant
    keypoints,"International Journal of Computer Vision, 60, 2 (2004),
    pp. 91-110.

   http://www.cs.ubc.ca/~lowe/keypoints/



On Mon, Dec 26, 2005 at 05:16:15PM -0500, Andy Adler wrote:
# On 12/26/05, Etienne Grossmann <et...@cs...> wrote:
# > octave:8> help isunix
# > isunix is the user-defined function from the file
# > /home=E6tienne/prog=F8ctave=F8ctave-forge=F8ctave-forge/main/general/=
isunix.m
# >
# > Always returns true.  If you are on a windows machine, be sure to
# > put an isunix.m which always returns false in your path.
#=20
# The semantics of 'isunix' depends on what you mean by UNIX.
# Strictly speaking, only certain well defined OSes are UNIX. Linux,
# for example, is not.
#=20
# On the other hand, maybe UNIX means OSes that behave like
# UNIX in most ways. cygwin has UNIX process semantics
# (ie. fork) and file semantics (symlinks, select on files, etc.)
#=20
# So, is cygwin UNIX? Clearly, a mingwin octave is not unix.
#=20
# Maybe isunix should make a specific test.
#=20
# --
# Andy
#=20

--=20
Etienne Grossmann ------ http://www.cs.uky.edu/~etienne

Re: [OctDev] isunix() returns 1 under cygwin

[Andy A. <ad...@nc...>]

On 12/26/05, Etienne Grossmann <et...@cs...> wrote:
> octave:8> help isunix
> isunix is the user-defined function from the file
> /home=E6tienne/prog=F8ctave=F8ctave-forge=F8ctave-forge/main/general/isun=
ix.m
>
> Always returns true.  If you are on a windows machine, be sure to
> put an isunix.m which always returns false in your path.

The semantics of 'isunix' depends on what you mean by UNIX.
Strictly speaking, only certain well defined OSes are UNIX. Linux,
for example, is not.

On the other hand, maybe UNIX means OSes that behave like
UNIX in most ways. cygwin has UNIX process semantics
(ie. fork) and file semantics (symlinks, select on files, etc.)

So, is cygwin UNIX? Clearly, a mingwin octave is not unix.

Maybe isunix should make a specific test.

--
Andy

[OctDev] isunix() returns 1 under cygwin

[Etienne G. <et...@cs...>]

  Hi All,

since

octave:8> help isunix
isunix is the user-defined function from the file
/home=E6tienne/prog=F8ctave=F8ctave-forge=F8ctave-forge/main/general/isun=
ix.m

Always returns true.  If you are on a windows machine, be sure to
put an isunix.m which always returns false in your path.

and isunix returns 1 under the octave-2.1.50 cygwin package, am I
right to conclude that every user of that package who wants isunix to
answer correctly should modify that function?

  TIA,
=20
  Etienne

--=20
Etienne Grossmann ------ http://www.cs.uky.edu/~etienne

Re: [Pkg-octave-devel] Re: [OctDev] compiling o-f against 2.1.72

[folajimi <fol...@sp...>]

On Wed, 21 Dec 2005, Quentin Spencer wrote:

> Rafael Laboissiere wrote:
>
> >* Michael Creel <mic...@ua...> [2005-11-23 15:46]:
> >
> >
> >
> >>I have had some trouble getting o-f CVS to compile against Octave
> >>2.1.72. Has anyone else had problems? Michael
> >>
> >>
> >
> >This is a belated reply but, indeed, compilation of octave-forge fails
> >miserably in my Debian unstable system with the latest version of octave.
> >
> >Question to the octave-forge crew: are there any plans to release a new
> >version of octave-forge compatible with Octave 2.1.72?
> >
> >
>
> I had just been wondering the same thing. As far as I can tell, it looks
> like we've reached a point where it's time to branch octave-forge. The
> current aging release of octave-forge compiles with octave 2.1.72, but
> not 2.9.4. The CVS doesn't compile with 2.9.4, and apparently (I haven't
> tried it) not with 2.1.72 either. It doesn't compile with 2.9.4 because
> the sparse stuff conflicts with the sparse stuff that's now in octave
> 2.9.x, but it compiles if the sparse directory is disabled. Furthermore,
> there are some functions like print.m that have been included in octave
> that are no longer necessary in octave-forge. I think octave-forge needs
> a little house cleaning and a new release, perhaps two: a 2.9.x
> compatible branch and a 2.1.x compatible branch. Has anyone kept track
> of what has been added to octave that can be removed from octave-forge?
> I'm not sufficiently familiar with CVS to know how to create branches.
> How does one go about doing that?
>
> I ought to raise one further question: what's the future of
> octave-forge? It seems like the new package system will result in it
> eventually going away, with things either being absorbed into octave or
> into smaller, more manageable packages. Until then, who's in charge of
> octave-forge? It seems like Paul doesn't have the time he used to for
> this--does it make sense to designate some other project leader(s) (I am
> not volunteering)?

This sounds like what Colin & myself has been tasked with doing. The cynosures
around here will probably weigh in on this sooner or later...

JA

[OctDev] updated index

[David B. <ad...@gm...>]

I just updated the index 
(http://octave.sourceforge.net/index/index.html) against the octave and 
octave-forge CVS versions, the previous version was from May against a 
2.9.3 release. I tried to update the INDEX files for the major changes 
since then, but if I missed something tell me and I'll add it...

Cheers
David

-- 
David Bateman                                Dav...@mo...
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary

Re: [OctDev] compiling o-f against 2.1.72

[David B. <Dav...@mo...>]

Quentin Spencer wrote:

> Rafael Laboissiere wrote:
>
>> * Michael Creel <mic...@ua...> [2005-11-23 15:46]:
>>
>>  
>>
>>> I have had some trouble getting o-f CVS to compile against Octave 
>>> 2.1.72. Has anyone else had problems? Michael
>>>   
>>
>>
>> This is a belated reply but, indeed, compilation of octave-forge fails
>> miserably in my Debian unstable system with the latest version of 
>> octave.
>>
>> Question to the octave-forge crew: are there any plans to release a new
>> version of octave-forge compatible with Octave 2.1.72?
>>  
>>
>
> I had just been wondering the same thing. As far as I can tell, it 
> looks like we've reached a point where it's time to branch 
> octave-forge. The current aging release of octave-forge compiles with 
> octave 2.1.72, but not 2.9.4. The CVS doesn't compile with 2.9.4, and 
> apparently (I haven't tried it) not with 2.1.72 either. It doesn't 
> compile with 2.9.4 because the sparse stuff conflicts with the sparse 
> stuff that's now in octave 2.9.x, but it compiles if the sparse 
> directory is disabled.

NOINSTALL in main/sparse fixes this

> Furthermore, there are some functions like print.m that have been 
> included in octave that are no longer necessary in octave-forge. I 
> think octave-forge needs a little house cleaning and a new release, 
> perhaps two: a 2.9.x compatible branch and a 2.1.x compatible branch. 
> Has anyone kept track of what has been added to octave that can be 
> removed from octave-forge? I'm not sufficiently familiar with CVS to 
> know how to create branches. How does one go about doing that?
>
There are some examples in main/miscellaneous/Makefile of some 
conditionally compiled. There is also regexp.cc that needs to be updated 
to the octave CVS syntax that is compatiable with matlab in octave forge 
and disabled from 2.9.5 on.. I'd suggest going this path rather than 
seperate 2.1 and 2.9 releases, as I believe a 2.9.x release marked as 
testing hopefully should be that far away. At that point octave-forge 
should be cleaned and then converted to the new package system.

The admin/make_index script gives information about shadowed functions 
in octave and octave_forge and can be used to find the migrated files.

> I ought to raise one further question: what's the future of 
> octave-forge? It seems like the new package system will result in it 
> eventually going away, with things either being absorbed into octave 
> or into smaller, more manageable packages. Until then, who's in charge 
> of octave-forge? It seems like Paul doesn't have the time he used to 
> for this--does it make sense to designate some other project leader(s) 
> (I am not volunteering)?

I'd like to see at least one more release of octave-forge, and more if 
2.9.x doesn't pass into testing some time soon. I'm definitely not 
volonteering either... My main reason for a desire for an octave-forge 
release is so that I can build a MingW release against a fixed 
octave-forge release rather than the CVS. I'd like at least one major 
change to octave-forge prior to that, as I'd like octave-forge to use 
the autoload function if it is available to reduce the size of the mingw 
release I'm making.

Cheers
David

-- 
David Bateman                                Dav...@mo...
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary

Re: [OctDev] compiling o-f against 2.1.72

[Quentin S. <qsp...@ie...>]

Rafael Laboissiere wrote:

>* Michael Creel <mic...@ua...> [2005-11-23 15:46]:
>
>  
>
>>I have had some trouble getting o-f CVS to compile against Octave 
>>2.1.72. Has anyone else had problems? Michael
>>    
>>
>
>This is a belated reply but, indeed, compilation of octave-forge fails
>miserably in my Debian unstable system with the latest version of octave.
>
>Question to the octave-forge crew: are there any plans to release a new
>version of octave-forge compatible with Octave 2.1.72?
>  
>

I had just been wondering the same thing. As far as I can tell, it looks 
like we've reached a point where it's time to branch octave-forge. The 
current aging release of octave-forge compiles with octave 2.1.72, but 
not 2.9.4. The CVS doesn't compile with 2.9.4, and apparently (I haven't 
tried it) not with 2.1.72 either. It doesn't compile with 2.9.4 because 
the sparse stuff conflicts with the sparse stuff that's now in octave 
2.9.x, but it compiles if the sparse directory is disabled. Furthermore, 
there are some functions like print.m that have been included in octave 
that are no longer necessary in octave-forge. I think octave-forge needs 
a little house cleaning and a new release, perhaps two: a 2.9.x 
compatible branch and a 2.1.x compatible branch. Has anyone kept track 
of what has been added to octave that can be removed from octave-forge? 
I'm not sufficiently familiar with CVS to know how to create branches. 
How does one go about doing that?

I ought to raise one further question: what's the future of 
octave-forge? It seems like the new package system will result in it 
eventually going away, with things either being absorbed into octave or 
into smaller, more manageable packages. Until then, who's in charge of 
octave-forge? It seems like Paul doesn't have the time he used to for 
this--does it make sense to designate some other project leader(s) (I am 
not volunteering)?

-Quentin

[OctDev] docstring change to spearman.m suggested

[D G. <de...@gn...>]

My spearman.m in octave2.1's forge calculates spearman as     

rho = cor (ranks (x), ranks (y));
this was as seen in
/usr/share/octave/2.1.69/m/statistics/base/spearman.m.



Since then, I read here that the implementation has been changed.
http://octave.sourceforge.net/index/f/spearman.html says that:

     For some (unknown) reason, in previous versions Spearman's rank
     correlation
     r = corrcoef(ranks(x)).
     But according to [1], Spearman's correlation is defined as
     r = 1-6*sum((ranks(x)-ranks(y)).^2)/(N*(N*N-1))
     The results are different. Here, the later version is implemented.

The 2 implementations should differ only when there are tied ranks.

However, it seems to me that the older implementation was more
logical.  The logical definition of spearman is
corr(ranks(x),ranks(y)), and the formula just happens to be what that
reduces to when there are no tied ranks.  Schaum's Probability and
Statistics derives the spearman formula from
corr(ranks(x),ranks(y)).. (but curiously, still uses the formula in an
exercise even though there are tied ranks.)

The following websites seem to address the issue of tied ranks, and
suggest that the formula is not valid in case of ties.

----
http://faculty.vassar.edu/lowry/ch3b.html -- says that the formula is
valid only for rankings without ties.

http://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient
-- the wikipedia article also says the same: "The formula becomes more
complicated in the presence of tied ranks, but unless the tie bands
are large, the effect of ignoring them is small"


http://www.netnam.vn/unescocourse/statistics/13_6.htm : 

When there are no ties, the formula for rs, reduces to .... where d is
the difference between the values of x and y corresponding to a pair
of observations. This simple formula will provide a good approximation
to rs when the number of ties in the ranks is small.


http://web.uccs.edu/lbecker/spss80/ctabs2.htm : 
says: CAUTION: Please note that the formula given above is
inappropriate when there are tied ranks. Our example data has many
ties. The rank order correlation computed by that formula is .513,
whereas the correct value (given by SPSS) is .370. So how should the
Spearman rank-order correlation be computed if there are ties? The
Spearman correlation is a special case of the Pearson product-moment
correlation. If you compute a Pearson product-moment correlation on
the ranked data the result will be the correct value of the Spearman
rank order correlation.


http://www.nyu.edu/its/socsci/Docs/correlate.html even provides a correction.

----

It seems then, that the previous version was perhaps more accurate
logically?

It seems that some (most) sources say that corr(ranks(x), ranks(y)) is
the correct definition, whereas others (like the last one) imply that
the formula is the correct definition, but then ask us not to rely on
the spearman formula in case of ties, but rather to use corr(ranks).  


In either case, perhaps it doesn't make sense for us to try to decide
what is the true definition.  Maybe we should keep the spearman
implementition as is (i.e. use the formula), but to add a doc like
this to the end of the docstring of spearman.m


"Note that this function simply implements the Spearman formula.  A
more logical indicator, often regarded as the true
definition[1,2,3,4,5] of Spearman Rank Correlation is
corr(ranks(x),ranks(y)), so you might prefer to use that instead. The
latter coincides with the Spearman formula when there are no ties
among ranks, but differs when there are ties among ranks.  For a small
number of ties, the difference can be ignored.

[1] http://faculty.vassar.edu/lowry/ch3b.html 
[2] http://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient
[3] http://www.netnam.vn/unescocourse/statistics/13_6.htm 
[4] http://web.uccs.edu/lbecker/spss80/ctabs2.htm 
[5] http://www.nyu.edu/its/socsci/Docs/correlate.html 
"



I guess the point is that we are distinguishing Spearman Correlation
from the formula that Spearman came up with. 


PS: Please cc me, i am not subscribed yet.

Sincerely,

DG                                 http://gnufans.net/
--

Re: [OctDev] compiling o-f against 2.1.72

[Rafael L. <ra...@de...>]

* Michael Creel <mic...@ua...> [2005-11-23 15:46]:

> I have had some trouble getting o-f CVS to compile against Octave 
> 2.1.72. Has anyone else had problems? Michael

This is a belated reply but, indeed, compilation of octave-forge fails
miserably in my Debian unstable system with the latest version of octave.

Question to the octave-forge crew: are there any plans to release a new
version of octave-forge compatible with Octave 2.1.72?

-- 
Rafael

Re: [OctDev] date/time functions

[William P. Y. H. <wil...@gm...>]

On 12/20/05, William Poetra Yoga Hadisoeseno <wil...@gm...> wrot=
e:
> OK, I thought it would be more useful, but later I decided unnamed
> tokens are good enough. I'm thinking of using strptime, though, to
> reduce the size of __date_str2vec__ and increase the likelihood of it
> being accepted ;) Any ideas? Suggestions?
>

I've submitted a patch for strptime (I sent it to bu...@oc...).
I've also merged __date_str2vec__.m into datevec.m (attached). Please
comment on it.

--
William Poetra Yoga Hadisoeseno

Re: [OctDev] date/time functions

[William P. Y. H. <wil...@gm...>]

Sorry, I forgot to attach the files.

Re: [OctDev] date/time functions

[William P. Y. H. <wil...@gm...>]

Ignoring the regexp issue for the moment, I've tidied up the functions
+ calendar: ready for inclusion into octave
+ datenum: ready for inclusion into octave
+ datestr: ready for inclusion into octave, but it uses strfind, which
is not in octave yet
+ datevec: ready for inclusion into octave, but it relies on
__date_str2vec__, which is not ready yet
+ eomday: ready for inclusion into octave
+ now: ready for inclusion into octave
+ weekday: ready for inclusion into octave

Please check the functions (included in the attachment) and comment on
them. In particular, I think __date_str2vec__ (was datesplit) needs to
be re-rewritten. Can someone please help me with this?

I think strfind is also easy. Can someone tidy up strfind in
octave-forge and import it into octave?

--
William Poetra Yoga Hadisoeseno

[OctDev] test_octcdf.m

[Paul K. <pki...@us...>]

Alex,

test_octcdf.m with %! tests and nothing else isn't the correct way to 
write an independent test file.

For one thing, it means you have to type "test test_octcdf" to use it, 
which is awkward.

For another thing, test_octcdf shows up on the list of functions but it 
isn't one.

The technique used for main/sparse is slightly better.  There I wrote 
build_tests.sh which creates main/sparse/sptest.m with the following 
header:

##
## sptest
##
##    run preset sparse tests.  All should pass.
function [passes,tests] = sptest
   disp('writing test output to sptest.log');
   test('sptest','normal','sptest.log');
endfunction

%!...


Perhaps an even better solution is to create the file octcdf.test and 
modify admin/mktests.sh to include files with the .test extension.

These should probably be installed as well, and the test suite should 
know to look for files with the extension .test so that we can type 
'test octcdf'.

Comments?

- Paul

On Dec 15, 2005, at 12:39 PM, alex wrote:

> --- NEW FILE: test_octcdf.m ---
>
>
> %!shared fname, nc
> %! fname = [tmpnam '-octcdf.nc'];
> %! nc = netcdf(fname,'c');

<snip>

Re: [OctDev] interp1q.m wrapper?

[William P. Y. H. <wil...@gm...>]

On 12/16/05, Frederick (Rick) A Niles <fn...@mi...> wrote:
> On Fri, 2005-12-16 at 08:22 +0800, William Poetra Yoga Hadisoeseno
> wrote:
> > On 12/16/05, Frederick (Rick) A Niles <fn...@mi...> wrote:
> > > Can someone add a function called interp1q.m to octave-forge that sim=
ply
> > > calls interp1.m?
> > >
> > > It's suppose to be the "quick" version of interp1.m that assumes the
> > > points are already sorted.
> > >
> > > I'm working with some Matlab developers and would like to run the
> > > project in Octave.  I can make this "fix" in my local copy, but I
> > > thought it might make sense for the others.
> > >
> >
> > Are you talking about developers of Matlab or developers of a project
> > that uses Matlab?
>
>
> The second.  They're windows-only users that were given Matlab when they
> started working here, so they seen no reason to even try Octave. :(
>
>

OK, then they should be more acceptive to Octave (than Matlab developers :p=
).

> > If you can write the script, then it would be nice to contribute it to
> > octave-forge. Just remember that it should be free, otherwise all of
> > us might not be able to benefit fully from it.
> >
> > Anyway, from what I can gather from the interp1 documentation,
> > interp1q should be the one called by interp1, right? Or is it the
> > other way round?
>
> Well, Ideally.  interp1 would call interp1q after it sorted the input.
> However, I was thinking that in might be more easy to get accepted by
> the Octave-forge maintainers if I didn't try to modify an existing
> function and just made interp1q an alias for interp1.
>

No, actually whether or not your contribution gets accepted depends on
the quality of your script, not whether you modify an existing
function or not. Just try ;)

--
William Poetra Yoga Hadisoeseno

[OctDev] interp1q.m wrapper?

[Frederick (R. A N. <fn...@mi...>]

Can someone add a function called interp1q.m to octave-forge that simply
calls interp1.m?

It's suppose to be the "quick" version of interp1.m that assumes the
points are already sorted.

I'm working with some Matlab developers and would like to run the
project in Octave.  I can make this "fix" in my local copy, but I
thought it might make sense for the others.

Rick Niles.

[OctDev] admin/mkpkgadd

[<bu...@pi...>]

Hi,

during make install, if the variable CDPATH is set, the command

cd $dir

in the script admin/mkpkgadd echos $dir to stdout. This causes the final 
  PKG_ADD to contain that $dir when it should in fact be empty.

	Gerd

Re: [OctDev] Proposed changes to cell2mat and inclusion in Octave

[John W. E. <jw...@be...>]

On 14-Dec-2005, William Poetra Yoga Hadisoeseno wrote:

| It seems that no one has objections at this point, so I'm braving
| myself to submit this file (imported from octave-forge). The file is
| attached.

I included this function.

Thanks,

jwe

Re: [OctDev] Proposed changes to cell2mat and inclusion in Octave

[William P. Y. H. <wil...@gm...>]

Part of the discussion on octave-dev follows:

On 12/9/05, David Bateman <Dav...@mo...> wrote:
> William Poetra Yoga Hadisoeseno wrote:
>
> >On 12/9/05, David Bateman <Dav...@mo...> wrote:
> >
> >
> >>Again, why? Here is matlab's output
> >>
> >> >> C =3D {[1], [2, 3, 4; 6 7 8]; [5; 9],[10 11 12]}
> >> >> cell2mat(C)
> >>??? Error using =3D=3D> cat
> >>CAT arguments dimensions are not consistent.
> >>
> >>Error in =3D=3D> cell2mat at 92
> >>        m{n} =3D cat(2,c{n,:});
> >>
> >>The error message seems to indicate that matlab just lets cat generate
> >>the error as well.
> >>
> >>
> >>
> >
> >Well, in that case then we should just let cat do the job ;)
> >
> >
> >
> >>Matlab doesn't respect their own documentation either
> >>
> >> >> C =3D {[1, 2], [3, 4]; [5; 9],[6 7 8; 10 11 12]};
> >> >> cell2mat(C)
> >>
> >>ans =3D
> >>
> >>     1     2     3     4
> >>     5     6     7     8
> >>     9    10    11    12
> >>
> >>This is with R14sp2..
> >>
> >>
> >>
> >
> >OK, this is consistent with the old (current) cell2mat.
> >
> >
> >
> >> >> C =3D {[1;2],ones(2,2)=3D=3D1}
> >>
> >>C =3D
> >>
> >>    [2x1 double]    [2x2 logical]
> >>
> >> >> cell2mat(C)
> >>??? Error using =3D=3D> cell2mat
> >>All contents of the input cell array must be of the same data type.
> >>
> >>However, this seems stupid to me, as logical arrays are supposed to be
> >>promoted to double in normal usage... I'd prefer that the above remaine=
d
> >>legal..
> >>
> >>
> >>
> >
> >Yes, I also think the above should work (logical -> numeric). Also, I
> >think char arrays should be allowed (using ascii). What do you think?
> >
> >
> >
> Promotion of type is implicit in the concatenation operation. If cat
> allows it so should cell2mat in my opinion. Therefore just letting cat
> generate the error messages for an illegal concatenation of types makes
> sense to me, as it will report the two offending types in the
> concatenation operation. I believe this means that cell2mat needs no
> changes...
>
> Regards
> David
>
>
> --
> David Bateman                                Dav...@mo...
> Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph)
> Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax)
> 91193 Gif-Sur-Yvette FRANCE
>
> The information contained in this communication has been classified as:
>
> [x] General Business Information
> [ ] Motorola Internal Use Only
> [ ] Motorola Confidential Proprietary
>
>

It seems that no one has objections at this point, so I'm braving
myself to submit this file (imported from octave-forge). The file is
attached.

--
William Poetra Yoga Hadisoeseno

Re: [OctDev] date/time functions

[Paul K. <pki...@us...>]

You are welcome to submit anything in octave-forge for inclusion in=20
Octave.

Can you do what you need in ver.m using the native octave functions=20
strftime and strptime?

Ultimately datestr and datenum should be part of octave since they=20
extend date handling beyond the 1970-2038 supported by unix functions.

- Paul

On Dec 10, 2005, at 12:14 AM, William Poetra Yoga Hadisoeseno wrote:

> My new version of ver.m needs datestr and datenum, which are in
> octave-forge. Are they ready for inclusion into Octave?
>
> Paul, since most of the functions in main/time have your copyright on
> them, would you like to tidy them up for inclusion into Octave? ;)
>
> --
> William Poetra Yoga Hadisoeseno
>
>
> -------------------------------------------------------
> This SF.net email is sponsored by: Splunk Inc. Do you grep through log=20=

> files
> for problems?  Stop!  Download the new AJAX search engine that makes
> searching your log files as easy as surfing the  web.  DOWNLOAD =
SPLUNK!
> http://ads.osdn.com/?ad_idv37&alloc_id=16865&op=3Dclick
> _______________________________________________
> Octave-dev mailing list
> Oct...@li...
> https://lists.sourceforge.net/lists/listinfo/octave-dev
>

[OctDev] date/time functions

[William P. Y. H. <wil...@gm...>]

My new version of ver.m needs datestr and datenum, which are in
octave-forge. Are they ready for inclusion into Octave?

Paul, since most of the functions in main/time have your copyright on
them, would you like to tidy them up for inclusion into Octave? ;)

--
William Poetra Yoga Hadisoeseno