does not run, exception in gtk+ about biodiverse HOT 10 CLOSED

Thanks for the feedback.

The first Gtk warning is possibly a clash with another Gtk app using a more 
recent
version of the file history (the Gimp is one example). This does not affect the
Biodiverse program itself.  

The second issue is due to the input file containing a non-numeric or null 
value in a
group column (the "[BASEDATA] Non-numeric cell size field 7, check your data or
cellsize arguments" warning).  

The system has been modified to optionally ignore these in the next release 
(version
0.11), or alternately it will print a warning in a popup window so it is easier 
to
see the error.  See Issue #26

Until version 0.11 is released the solution is to delete records with 
non-numeric or
null group columns.  

Closing the progress dialogue will stop the import in other cases.  

Neither of those were the problem. I simply changed all spaces to commas and let
Biodiverse guess the separator and quote character. That DID load the file. 
There
were no NULL data and no non-numerics in the file.

I got some result, but I wanted to analyze more of the data (mostly on -1,1, 
except
for an angle). Cell size 0.6, 8 groups, one label. Crashed in clustering.

Explanation of why the space-delimited version failed (probably). In two places 
I had
(intentionally for a different use) put in a pair of spaces. 

Pairs of spaces would cause a problem if they were the delimiter.  It would 
result in
a null value between the two spaces for one of the columns.

Please let me know which columns you are using for the groups and labels so I 
can
reproduce the cluster error.

The first two columns were used as labels. All 8 of the floating point values 
were
used as groups (ang,dist,x,y,z,basex,basey,basez or whatever I called them). The
strange thing was that Vista did not even notify that there was a crash. I just
noticed that the window wasn't there any more.

Just tested the data set using version 0.11 and it clusters with no problems
(sp_select_all() was the spatial condition).  It does take a while, though, as 
the
matrix is very large (~1.4*10^6 values).  It also runs very close to the memory
limits on WinXP (2GB), and exceeds them when two cluster analyses are run and 
then
the project is saved.  This causes a failure and a crash we cannot trap.

Many of the silent error issues have been rectified through issue #29.

update label and owner

version 0.11 released so issue marked as fixed.