2:41 p.m.
Hi,
When I was downloading videolan, I saw there a link to old systems,
like win98. From the comments on the page it seems a good solution.
http://sourceforge.net/projects/kernelex/
I will try to test it (if the computer still works ;)
Bye,
Hélio
4:17 a.m.
On Tue, 15 Nov 2011, Hélio Guilherme wrote:
When I was downloading videolan, I saw there a link to old systems,
like win98. From the comments on the page it seems a good solution.
http://sourceforge.net/projects/kernelex/
I will try to test it (if the computer still works ;)
Let us know. I can put a link to kernelex on the gretl for Windows
page if it works.
Allin Cottrell
6:51 p.m.
Hi,
I did the test but did not manage to get Gretl working. Tested with
snapshot, on my Windows 98 system (4.10.1998).
Installer worked fine.
At first it complained about missing DNSAPI.DLL. After getting this
from Internet and putting it in gretl's folder got a new error:
The DNSAPI.DLL file is linked to missing export NTDLL:NtQueryValueKey.
Another window appears with:
C:\Program Files\gretl\gretlw32.exe
A device attached to the system is not functioning.
Downloaded NTDLL.DLL and put it in gretl's folder, but the same happens.
As I cannot certify the quality of these DLLs, I cannot say that the
problem is with gretl.
There are differences on the builds of these two DLLs. The NTDLL.DLL
says 5.1.2600.2180 (xpsp_sp2_rtm.040803-2158), and the DNSAPI.DLL says
5.2.3790.0 (srv03_rtm.030324-2048).
My feeling is that gretl is trying to get some values from the registry key.
Anyway, I have tested the gnuplot packed with gretl and the demo ran
fine. Also the version verifier was OK.
----
I was hopefull that gretl_cli would work, but got a crash in libgretl.
----
I will test again with the official 1.9.6 package.
(maybe you could mention the link for the program, since it may work
for some setups)
On 11/16/11, Allin Cottrell <cottrell(a)wfu.edu> wrote:
On Tue, 15 Nov 2011, Hélio Guilherme wrote:
> When I was downloading videolan, I saw there a link to old systems,
> like win98. From the comments on the page it seems a good solution.
> http://sourceforge.net/projects/kernelex/
>
> I will try to test it (if the computer still works ;)
Let us know. I can put a link to kernelex on the gretl for Windows
page if it works.
Allin Cottrell
4:56 p.m.
New subject: periodicity 336
I tried to set up a script with periodicity 336 and gretl responded:
? setobs 336 1:1
frequency (336) does not make seem to make sense
Why not? One observation each half an hour in a week: 48x7=336
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.ea3.ehu.es
5:30 p.m.
New subject: periodicity 336
On Thu, 24 Nov 2011, Ignacio Diaz-Emparanza wrote:
I tried to set up a script with periodicity 336 and gretl responded:
? setobs 336 1:1
frequency (336) does not make seem to make sense
Why not? One observation each half an hour in a week: 48x7=336
For a non-standard time-series frequency you need to add the
option --time-series:
<hansl>
nulldata 800
setobs 336 1:1 --time-series
smpl ; 10
print index --byobs
</hansl>
Allin Cottrell
9:28 a.m.
New subject: periodicity 336
El 24/11/11 18:30, Allin Cottrell escribió:
On Thu, 24 Nov 2011, Ignacio Diaz-Emparanza wrote:
> I tried to set up a script with periodicity 336 and gretl responded:
>
> ? setobs 336 1:1
> frequency (336) does not make seem to make sense
>
> Why not? One observation each half an hour in a week: 48x7=336
For a non-standard time-series frequency you need to add the
option --time-series:
<hansl>
nulldata 800
setobs 336 1:1 --time-series
smpl ; 10
print index --byobs
</hansl>
Sorry, I forgot that.
But this was because I previously tried with periodicity 24x7=168
(hourly data in a week) without the --time-series option and it works.
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.ea3.ehu.es
8:40 p.m.
New subject: periodicity 336
On Fri, 25 Nov 2011, Ignacio Diaz-Emparanza wrote:
El 24/11/11 18:30, Allin Cottrell escribió:
> On Thu, 24 Nov 2011, Ignacio Diaz-Emparanza wrote:
>
>> I tried to set up a script with periodicity 336 and gretl responded:
>>
>> ? setobs 336 1:1
>> frequency (336) does not make seem to make sense
>>
>> Why not? One observation each half an hour in a week: 48x7=336
>
> For a non-standard time-series frequency you need to add the
> option --time-series:
>
> <hansl>
> nulldata 800
> setobs 336 1:1 --time-series
> smpl ; 10
> print index --byobs
> </hansl>
Sorry, I forgot that.
But this was because I previously tried with periodicity 24x7=168
(hourly data in a week) without the --time-series option and it works.
gretl's heuristic for rejecting a periodicity specification, pd, is
as follows: pd is less than 1 OR (pd is greater than the number
of observations in the dataset AND the --time-series flag was not
given).
Allin
3:35 p.m.
New subject: scatters --output=display
Hello,
this script works correctly, presenting a window with the graph
<script>
open bjg.gdt
scatters g lg --output=display
</script>
but this does not produce the graph window:
<script>
function void test_graf(series y, series x)
scatters y x --output=display
end function
open bjg.gdt
test_graf(g, lg)
</script>
I think that 'scatters' should work the same as the 'gnuplot' command,
being correct to call it from inside a function, or not?
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.ea3.ehu.es
7:01 p.m.
New subject: scatters --output=display
On Tue, 29 Nov 2011, Ignacio Diaz-Emparanza wrote:
this script works correctly, presenting a window with the graph
<script>
open bjg.gdt
scatters g lg --output=display
</script>
but this does not produce the graph window:
<script>
function void test_graf(series y, series x)
scatters y x --output=display
end function
open bjg.gdt
test_graf(g, lg)
</script>
I think that 'scatters' should work the same as the 'gnuplot' command,
being correct to call it from inside a function, or not?
Yes, I suppose so. Fixed in CVS.
Allin
10 a.m.
New subject: scatters --output=display
El 29/11/11 20:01, Allin Cottrell escribió:
On Tue, 29 Nov 2011, Ignacio Diaz-Emparanza wrote:
> this script works correctly, presenting a window with the graph
>
> <script>
> open bjg.gdt
> scatters g lg --output=display
> </script>
>
> but this does not produce the graph window:
>
> <script>
> function void test_graf(series y, series x)
> scatters y x --output=display
> end function
>
> open bjg.gdt
> test_graf(g, lg)
> </script>
>
> I think that 'scatters' should work the same as the 'gnuplot'
command,
> being correct to call it from inside a function, or not?
Yes, I suppose so. Fixed in CVS.
Allin
I confirm it works now. Thanks!
--
Ignacio Diaz-Emparanza
DEPARTAMENTO DE ECONOMÍA APLICADA III (ECONOMETRÍA Y ESTADÍSTICA)
UPV/EHU Avda. Lehendakari Aguirre, 83 | 48015 BILBAO
T.: +34 946013732 | F.: +34 946013754
www.ea3.ehu.es
4772
days inactive
4787
days old
9 comments
3 participants
participants (3)
-
Allin Cottrell -
Hélio Guilherme -
Ignacio Diaz-Emparanza