Thank you. We will investigate.
Dean
Dr. Dean C. Adams
Professor
Department of Ecology, Evolution, and Organismal Biology
Department of Statistics
Iowa State University
www.public.iastate.edu/~dcadams/<http://www.public.iastate.edu/~dcadams/>
phone: 515-294-3834
From: Mauro Cavalcanti [mailto:[email protected]]
Sent: Wednesday, December 16, 2015 8:18 PM
To: Adams, Dean [EEOBS] <[email protected]>
Cc: MORPHMET <[email protected]>
Subject: Re: [MORPHMET] Yet another sequel to complaints about 'geomorph' R
package
Dean,
Thanks for you reply.
As I wrote in the first message that started this sequence, the issue with the
missing Biostrings is a minor one, more or less easily solved. I surely am
aware that authors of R packages have no control (fortunately, indeed!) over
the changes done by the authors of other such packages.
Concerning by main complaint, namely the change in scaling of the landmark
coordinates genearated by the digitize2d function, the only mention I found in
the NEWS file (from version 2.16 onwards) wa that: "coordinates returned by
digitize2d() are now unscaled, and SCALE= returns the scale". But this hardly
constitutes proper documentation. WHY has such a change been done? And WHY an
option to allow the user to retain the 'old' scaling system has not been added
to the function, instead of simply implementing a change in scaling overnight?
With all due respect, but this change should not have been done without warning
and without an alternative, in a function dealing with such a critical and
time-consuming activity as data collection -- because it could lead to data
loss (which was just what happened in my case). Furthermore, this change in
scaling made the files digitized with geomorph unreadable by the tps series
programs (they generate a floating-point error), thus making not possible the
highly desirable interchangeable use of morphometrics software (for example,
for comparing algorithms and cross-checking analyses).
I hope the geomorph Development Team will appear with a good and democratic
solution for this issue in due time.
With best regards,
2015-12-16 15:16 GMT-02:00 Adams, Dean [EEOBS]
<[email protected]<mailto:[email protected]>>:
Mauro,
Please re-read the news file. There was a documented change in version 2.1.6
with respect to how the scale was handled (which remains documented in
subsequent updated versions). Please see the News file where this was
mentioned, and the help file for digitize2d which explained the change. We of
course are happy to receive any suggestions for improvements, and if any
unforeseen issues arise as a result of new versions, we would like to know
these as well.
As for inter-linked packages, geomorph depends on ape and rgl, and imports from
geiger, phytools, jpeg, and several other packages. The issue you noted was an
error in Phanghorn, which is called by phytools. Geopmorph does not use
Biostrings, but since geomorph calls phytools, a cascading install error
occurred. While unfortunate, I simply have no control over this.
I have been in contact with the maintainers of both packages, and have alerted
them to the problem. However, it will likely be some time until the corrected
version of Phanghorn is on the main CRAN repository, since they have just
uploaded new versions to CRAN. In the meantime, the workaround you mentioned of
manually installing Biostrings is needed. Clearly this is not my preference,
but I have no control of what others require in their packages.
Dean
Dr. Dean C. Adams
Professor
Department of Ecology, Evolution, and Organismal Biology
Department of Statistics
Iowa State University
www.public.iastate.edu/~dcadams/<http://www.public.iastate.edu/~dcadams/>
phone: 515-294-3834
From: Mauro Cavalcanti [mailto:[email protected]<mailto:[email protected]>]
Sent: Wednesday, December 16, 2015 3:27 PM
To: MORPHMET <[email protected]<mailto:[email protected]>>
Subject: [MORPHMET] Yet another sequel to complaints about 'geomorph' R package
Dear members of the geomorph R package Development Team,
I at least have been able to track down the "missing library" error when
attempting to install even previous versions of geopmorph.
It turns out that geomorph depends upon a package called 'phangorn' -- which
has just been updated to make use of "Biostrings". At least one other R package
(phytools) is also dependent upon phangorn Iand therefore upon Biostrings too).
The fact that the Biostrings package has not been installed as a depedency of
the latest phangorn is due to the fact that the Bioconductor Project (and
associated packages) is not on CRAN. A big mess, indeed...
Hope this helps!
It seems that these complaints about geomorph are turning to become a series
with many sequels as "Star Wars"...
May the Force Be With You! ;-)
--
Dr. Mauro J. Cavalcanti
E-mail: [email protected]<mailto:[email protected]>
Web: http://sites.google.com/site/maurobio
--
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
---
You received this message because you are subscribed to the Google Groups
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to
[email protected]<mailto:[email protected]>.
--
Dr. Mauro J. Cavalcanti
E-mail: [email protected]<mailto:[email protected]>
Web: http://sites.google.com/site/maurobio
--
MORPHMET may be accessed via its webpage at http://www.morphometrics.org
---
You received this message because you are subscribed to the Google Groups
"MORPHMET" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
