Hi,
I tend to agree with the idea that the reply to the midasreg question is not that obvious unless you are in the very inner circle of gretl developpers, which I am obviously not. Actually, I consider myself as a long time fan of Gretl but with just a mild technical knowledge of it. Recently I lost my credentials in Gretl and I was unable to upload an updated version of Kapetanios. I tried to get a new account; when I saw the midasreg question, I tried to find the answer by googling it, but I only obtained, apparently, wrong solutions. At the end, I simply tried multiples times to open the new account by changing the reply to the midasreg question. Perhaps not so surprisingly, I exhausted all the wrong answers before getting the right one. This procedure is really discouraging; I honestly prefer the other 3 questions. Perhaps it could be replaced with a "recaptcha"? It's just an opinion.
Friendly regards,
D.
  


Missatge de Sven Schreiber <svetosch@gmx.net> del dia dg., 6 de març 2022 a les 6:10:
Am 04.03.2022 um 19:05 schrieb Allin Cottrell:
> On Fri, 4 Mar 2022, Sven Schreiber wrote:
>
>> I've just noticed the current state of the procedure how to obtain a
>> login to upload a contributed package (as an author). This quiz strikes
>> me as prohibitive: "The midasreg command was introduced in"...
>> Seriously? I would fail bitterly, thank God I already have an access
>> account.
>
> Seems to me anyone serious about gretl knows there's a change log
> online (linked on the main gretl page) and can easily search through
> it via their browser to find the earliest reference to midasreg. I
> suppose we could insert a link to the change log in that question but
> I don't think that should really be necessary.

Hm, I really think we shouldn't overestimate the extent to which people
outside our devel bubble know the various bits and pieces of the gretl
universe. I mean the changelog isn't even shipped in the gretl
installation. (or is it?) And most people wouldn't care about it. Just
think how little even the gretl inner circle people are aware of when a
certain feature was introduced and what the resulting gretl version
requirement is.

I know we've had spamming problems in that area, but maybe we have
overshot the target. I think with contributed packages it's a bit like
voting in the US: it should be made easier, not harder.

>
>> And question number 1 is ill-posed I think, sigma isn't defined.
>
> Conventionally, in the context of a regression, sigma^2 is the
> variance of the error term unless otherwise specified, no?

Well, yes, but since there are multiple choice options like
"hyperelasticity" one wonders if perhaps the sigma equation part isn't
also some kind of trick question :-)

cheers

sven


>
> Allin
> _______________________________________________
> Gretl-devel mailing list -- gretl-devel@gretlml.univpm.it
> To unsubscribe send an email to gretl-devel-leave@gretlml.univpm.it
> Website:
> https://gretlml.univpm.it/postorius/lists/gretl-devel.gretlml.univpm.it/
_______________________________________________
Gretl-devel mailing list -- gretl-devel@gretlml.univpm.it
To unsubscribe send an email to gretl-devel-leave@gretlml.univpm.it
Website: https://gretlml.univpm.it/postorius/lists/gretl-devel.gretlml.univpm.it/