9:33 a.m.
Hi,
the German translation is more or less finished, although IIRC there is
still a .pot-change to come?
Some remaining small issues I discovered in the latest windoze snapshot:
- opening the pdf manual failes with the English message: "failed to
download file"; at the very least, this message should be localized
- similar for "there are no new files on the server right now"
- the version number in About... is still at 1.5.1
- the new menu entry "seed for random numbers" is missing in the po-file
- in preferences, can somebody think of a better tab title than "HCCME"?
I used "H(A)C" for now, but it's not much better I think. Maybe
"Robust
Cov"?
- why is opening of PcGive and RATS files under database instead of import?
- in the ADF dialog, shouldn't constant/trend/quad trend be mutually
exclusive radio buttons instead of checkboxes?
- in spectrum estimation: I can't see any difference between selecting
the sub-menu entries "periodogram" or "Bartlett lag window", the same
things appear to happen (namely table and graph).
Thanks, I think gretl 1.6.0 is going to be pretty cool!
-Sven
3:20 p.m.
I won't try to answer all your questions right now, but here are
the quick ones:
On Mon, 4 Sep 2006, Sven Schreiber wrote:
- why is opening of PcGive and RATS files under database
instead of import?
Because they are opened in "database mode": you get a window
showing all the series, from which you can import selected
series to gretl's workspace. Both Rats and PcGive databases can
contain series of mixed frequencies.
- in the ADF dialog, shouldn't constant/trend/quad trend be
mutually exclusive radio buttons instead of checkboxes?
No, that's correct, you can run several tests in one go, if you
like.
- in spectrum estimation: I can't see any difference between
selecting
the sub-menu entries "periodogram" or "Bartlett lag window", the
same
things appear to happen (namely table and graph).
Well yes, but take a look at the details. One gives the "raw"
sample periodogram and the other gives a smoothed version (along
with fractional integration tests).
Allin.
8:25 p.m.
Thanks for your answers...
Allin Cottrell schrieb:
> - in spectrum estimation: I can't see any difference between
selecting
> the sub-menu entries "periodogram" or "Bartlett lag window", the
same
> things appear to happen (namely table and graph).
Well yes, but take a look at the details. One gives the "raw" sample
periodogram and the other gives a smoothed version (along with
fractional integration tests).
ok, but doesn't the term "periodogram" only refer to the non-smoothed
raw estimate? I'm asking because even the smoothed values in the table
have that heading. And vice-versa, isn't the term "spectrum" a little
misleading as a graph title if it's the unsmoothed periodogram?
(I'm not 100% sure though what the most widespread conventions are in
the frequency domain.)
cheers,
sven
8:57 p.m.
On Mon, September 4, 2006 22:25, Sven Schreiber wrote:
Thanks for your answers...
Allin Cottrell schrieb:
>> - in spectrum estimation: I can't see any difference between selecting
>> the sub-menu entries "periodogram" or "Bartlett lag window",
the same
>> things appear to happen (namely table and graph).
>
> Well yes, but take a look at the details. One gives the "raw" sample
> periodogram and the other gives a smoothed version (along with
> fractional integration tests).
>
>
ok, but doesn't the term "periodogram" only refer to the non-smoothed
raw estimate? I'm asking because even the smoothed values in the table
have that heading. And vice-versa, isn't the term "spectrum" a little
misleading as a graph title if it's the unsmoothed periodogram?
(I'm not 100% sure though what the most widespread conventions are in
the frequency domain.)
cheers,
sven
As far as I know, the spectrum is a theoretical construct, whereas the
periodogram is its sample counterpart. As a consequence, the periodogram can
be viewed as an estimator of the spectrum. The trouble is that, unlike
autocorrelations, the sample statistic is not a consistent estimator, so you
normally want to use a smoothed version (which _is_ consistent under mild
assumptions); it's no surprise people sometimes use the word "spectrum"
when
they really mean "estimate of the spectrum", be it consistent or not, the same
way we often call "autocorrelations" what really are "sample
autocorrelations".
Riccardo "Jack" Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona
7:24 p.m.
Round 2 of my replies.
On Mon, 4 Sep 2006, Sven Schreiber wrote:
the German translation is more or less finished, although IIRC there
is
still a .pot-change to come?
I've just committed a new .pot to CVS.
Some remaining small issues I discovered in the latest windoze
snapshot:
- opening the pdf manual failes with the English message: "failed to
download file"; at the very least, this message should be localized
I hope this should be fixed now. I'll try testing the next time
I have Windows access, but I'm putting a new snapshot in place
too.
- the version number in About... is still at 1.5.1
Well, 1.6.0 is not really out yet.
- the new menu entry "seed for random numbers" is missing
in the po-file
Should be fixed.
Allin.
5:24 p.m.
new snapshot, next round...
Sven Schreiber schrieb:
Some remaining small issues I discovered in the latest windoze snapshot:
- opening the pdf manual failes with the English message: "failed to
download file"; at the very least, this message should be localized
- similar for "there are no new files on the server right now"
loading the pdfs now works fine, but the last message remains
non-localized (and it doesn't seem to be in the new .pot)
- in preferences, can somebody think of a better tab title than
"HCCME"?
I used "H(A)C" for now, but it's not much better I think. Maybe
"Robust
Cov"?
well, what do you think?
And then, some things I hadn't noticed before:
- I couldn't see how to set a variable to "discrete" in the denmark data
set. Not that it would make any sense there, but in case you disabled
discreteness for all time series data, isn't that a little too radical?
It's not really my field, but I don't see why we can't have time series
of discrete variables.
- Also with a discrete variable, there seems to be a bug when fitting a
normal distribution, it's way off (see attached saved graph
"histogram-discrete.png"). However, this didn't happen with another
discrete variable I had, so this bug may be subtle.
that's it for now,
Sven
7 p.m.
On Tue, 5 Sep 2006, Sven Schreiber wrote:
> - similar for "there are no new files on the server right
now"
the last message remains non-localized (and it doesn't seem to
be in the new .pot)
This is not trivial to fix. The phrase occurs in a a native
Windows auxiliary program. Since the updater should be able to
update gretl DLLs, it can't be part of gretl itself (at least I
don't think it can).
And then, some things I hadn't noticed before:
- I couldn't see how to set a variable to "discrete" in the denmark data
set. Not that it would make any sense there, but in case you disabled
discreteness for all time series data, isn't that a little too radical?
It's disabled for variables any of whose values have a
fractional part.
- Also with a discrete variable, there seems to be a bug when fitting
a
normal distribution, it's way off (see attached saved graph
"histogram-discrete.png").
Thanks, I'll take a look at this.
Allin.
7:05 p.m.
Allin Cottrell schrieb:
>
> - I couldn't see how to set a variable to "discrete" in the denmark
data
> set. Not that it would make any sense there, but in case you disabled
> discreteness for all time series data, isn't that a little too radical?
It's disabled for variables any of whose values have a fractional part.
Is that really necessary? For example, I was just told that in a popular
German panel dataset (GSOEP) the years-of-schooling variable can also
take on half-year (.5) values. But nothing in between, so it's still
discrete.
-sven
7:12 p.m.
On Wed, September 6, 2006 21:05, Sven Schreiber wrote:
Allin Cottrell schrieb:
>>
>> - I couldn't see how to set a variable to "discrete" in the denmark
data
>> set. Not that it would make any sense there, but in case you disabled
>> discreteness for all time series data, isn't that a little too radical?
>
> It's disabled for variables any of whose values have a fractional part.
>
Is that really necessary? For example, I was just told that in a popular
German panel dataset (GSOEP) the years-of-schooling variable can also
take on half-year (.5) values. But nothing in between, so it's still
discrete.
You can always multiply it by 2. Not very clean, but...
Riccardo "Jack" Lucchetti
Dipartimento di Economia
Facoltà di Economia "G. Fuà"
Ancona
7:18 p.m.
Riccardo Jack Lucchetti schrieb:
On Wed, September 6, 2006 21:05, Sven Schreiber wrote:
> Allin Cottrell schrieb:
>
>>> - I couldn't see how to set a variable to "discrete" in the
denmark data
>>> set. Not that it would make any sense there, but in case you disabled
>>> discreteness for all time series data, isn't that a little too radical?
>> It's disabled for variables any of whose values have a fractional part.
>>
> Is that really necessary? For example, I was just told that in a popular
> German panel dataset (GSOEP) the years-of-schooling variable can also
> take on half-year (.5) values. But nothing in between, so it's still
> discrete.
>
You can always multiply it by 2. Not very clean, but...
Yeah right ;-)
Seriously though, it's an explicit user action, so why restrict it at all?
Also, IIRC you guys agreed to mark discrete variable in the main list
(Icon, or 'd', or whatever). Is that still on?
-sven
9:17 p.m.
On Wed, 6 Sep 2006, Sven Schreiber wrote:
Riccardo Jack Lucchetti schrieb:
> On Wed, September 6, 2006 21:05, Sven Schreiber wrote:
>> Allin Cottrell schrieb:
>>
>> Is that really necessary? For example, I was just told that in a popular
>> German panel dataset (GSOEP) the years-of-schooling variable can also
>> take on half-year (.5) values. But nothing in between, so it's still
>> discrete.
>
> You can always multiply it by 2. Not very clean, but...
Yeah right ;-)
Seriously though, it's an explicit user action, so why restrict it at all?
OK, I can see the case for making it unrestricted. But I'm not
goint to do that for 1.6.0 because it might break some things.
BTW, I've fixed the frequency plot business for discrete
variables in CVS.
Also, IIRC you guys agreed to mark discrete variable in the main
list
(Icon, or 'd', or whatever). Is that still on?
Probably after 1.6.0.
Allin.
6445
days inactive
6447
days old
10 comments
3 participants
participants (3)
-
Allin Cottrell -
Riccardo Jack Lucchetti -
Sven Schreiber