anybody running gretl on tablets?
by cociuba mihai
JM2C
I'm using R on my tablet by connecting to an online server with RStudio
server http://www.rstudio.com/ide/server/ and I'm wondering if a similar
solution would be feasible for gretl?
On Thu, 9 Jan 2014, Sven Schreiber wrote:
>* Am 09.01.2014 11:28, schrieb Riccardo (Jack) Lucchetti:
*>>* On Thu, 9 Jan 2014, Sven Schreiber wrote:
*>>>>>* Hi,
*>>>>>>* has anybody out there tried to run gretl on a tablet device?
The general
*>>>* idea is of course not new, but now there are Windows tablets appearing
*>>>* with Intel Atom processors and apparently regular Windows versions (not
*>>>* the earlier embedded variant), so it would seem like being as easy as on
*>>>* a desktop/laptop.
*>>>>>>* Obviously it would be even more interesting to do it on some kind of
*>>>* Linux (Android, FirefoxOS, Ubuntu, etc.), but as mentioned before, I
*>>>* think the necessary software ingredients are still missing on those
*>>>* platforms. (Of course, if anybody has counterexamples we would all like
*>>>* to hear them...)
*>>>>* The lion's share of the tablet OS market belongs to iOS and Android. I
*>>* can't speak about iOS, but porting gretl to Android would be quite a
*>>* feat. First, you'd have to port the libgretl library; I understand there
*>>* is some way to use gcc to cross-compile for the ARM platform, but then
*>>* one should find (or build from source) all the libraries libgretl needs:
*>>* lapack, blas, fftw3... On the other hand, apparently somenoe did manage
*>>* to port Octave to Android, so it would seem that there's some hope on
*>>* that front. Still, it wouldn't be easy.
*>>>>* Exactly. That's why it sounds attractive to me have to have a regular
*>* x86-style CPU (Atom) in a tablet, as in the Dell Venue (just an
*>* example), instead of ARM. I guess that it would even be possible to load
*>* a standard Linux distro on them. Maybe even dual-boot??
*
I've thought (to a limited extent) about this sort of thing, and I
came to the conclusion that there's nothing useful to be done right
now.
A determined geek could get gretlcli running on just about any
modern device (from raspberry pi on up, and including ARM). But the
GUI is another matter, for two reasons: (1) gtk is not available for
(and/or would not be permitted on) the devices one might want to
target, and (2) even if gtk were available, gretl's desktop-type
interface is just not right for use on tablet / smartphone / toaster
/ whatever.
My motivation to write a touch GUI for gretl in Java or iOS is zero
but if anyone else wants to try I'll be happy to make encouraging
noises and help from time to time, to the extent I can.
Allin
10 years, 11 months
gretl build on ubuntu 13.10
by cociuba mihai
>
> I am building gretl on ubuntu 13.10 with the openmp option and there is
> the following error:
make[1]: *** No rule to make target `../lib/src/bfgs_mp.c', needed by
`gretl_bfgs.lo'. Stop.
make[1]: Leaving directory `/home/asus/gretl_cvs/gretl/lib'
make: *** [lib] Error 2
what should I do?
10 years, 11 months
adding description to a variable
by Leandro Zipitria
Dear Gretl users,
I have a dataset that have 200 variables, all by the name prod_# (that is,
a list of products from 1 to 200). I also have a csv with the description
of that product (oil, ham, etc.).
Is there any (simple) way of adding the information of each product to the
description of the variable by reading the csv file? I know that it could
be done "by hand" by editing each variable and copy and paste on the
description of the variable.
I have read the manual, and do not find such way to add descriptions to a
variable.
Regards,
Leandro
10 years, 11 months
Time Series Decomposition
by Mehdi Elharti
First of all, I want to thank you for the effort you have put in to provide
us this wonderful software for free.
My question is: I looked for a decomposition function (additive and
multiplicative) as in Eviews, but I couldn't find it on gretl!!! Are you
planning on adding such function to gretl?
Thank you again and have a nice day.
10 years, 11 months
on gretl's capacity to handle a large matrix
by artur.bala.tn
Dear Allin,
Is there something like Stata's "set matsize" command in gretl?
gretl crashed when estimating a panel including 600 unit dummies (LSVD estiamtion); Stata did stop over and asked to increase the matsize capacity to handle a larger matrix (the upper limit is 11 000)
Merry Christmas to you and to the gretl community!
Bets,
Artur
10 years, 11 months
anybody running gretl on tablets?
by Sven Schreiber
Hi,
has anybody out there tried to run gretl on a tablet device? The general
idea is of course not new, but now there are Windows tablets appearing
with Intel Atom processors and apparently regular Windows versions (not
the earlier embedded variant), so it would seem like being as easy as on
a desktop/laptop.
Obviously it would be even more interesting to do it on some kind of
Linux (Android, FirefoxOS, Ubuntu, etc.), but as mentioned before, I
think the necessary software ingredients are still missing on those
platforms. (Of course, if anybody has counterexamples we would all like
to hear them...)
thanks for any info,
sven
10 years, 11 months
Durbin's h statistics
by Lovrenc Pfajfar
Dear Gretl-users,
Can somebody tell me, how exactly (formula) is calculated Durbin's h
statistic in Gretl.
Thanks a lot
L. Pfajfar
10 years, 11 months
nonlinear restriction
by Artur T.
First of all: Happy new year to the members of the gretl community!
Some years ago the possibility of testing non-linear restrictions within
the single-equation framework (based on the Delta method) was added. The
help file provides a simple example on this:
<hansl>
function matrix restr (const matrix b)
matrix v = -b[3]/b[2]
return v
end function
open denmark --quiet
ols diff(LRM) 0 LRM(-1) LRY(-1) IBO(-1)
restrict
rfunc = restr
end restrict
</hansl>
The issue I face is given by the fact that I would like to test not only
the restriction -b[3]/b[2] = 0 but also e.g. -b[4]/b[2] = 0 separately.
Atm I have to write down another function
function matrix restr2 (const matrix b)
matrix v = -b[4]/b[2]
return v
end function
if I want to test -b[4]/b[2] = 0.
Is this right or is there an alternative way to obtain a vector of test
results? It would be helpful if one could extend the function by
something like this:
function matrix restr (const matrix b, int pos1, int pos2)
matrix v = -b[pos2]/b[pos1]
return v
end function
Best,
Artur
10 years, 11 months
