Mauro, I do agree with both things of course. Getting rid of radiation errors should be the goal. But it is a very difficult and complicated matter. And there is no good or bad choice for amateurs here. Very few are so lucky to have a farmer nearby providing grassland to setup your system out in the open free of charge like I did. In fact I have several farmers in my neighbourhood. So to me it is essential since I have the same surroundings to come as close as possible to the KNMI as this is my reference. I cannot compare to cities surely not when radiation plays a big role. When you are however measuring things in the cities it becomes a lottery and I think money permitting your would do well to eradicate as much of radiative and other "errors" if you want to measure as correct as possible.
Interesting story on the screens with snow covered ground. These indeed gave me the most extreme differences due to radiation in the past. Some years ago I measured -17,2 C in february with snowcoverd ground (just 7 cm btw) and drove around that night. It was amazing to see how every tree in our flat and open landscape made the mercury jump by 5-7 C! so -17 to -18 C where I drove and one tree and boom: -11 C. Which was exactly the temperature in villlages and cities. -11 C. Every little bit of wind made the mercury jump too by 3-4-5 C. I recorded every minute. So at 7.01 after a small gust of wind, may b 0,9 m/s made the mercury simply plumeted!
7.01: -13,3
7,02: -14,0
7.03: -15,2
7.04: -16,3
7.05: -16.5
7.06: -17,2
So the gust just mixed up some air around the screen, the response (Vaisala) was incredible. It also went up with wind just like that. Now this also can be suspected of a screen that is to open. In which case the instrument itself radiates heat away hence you get an instrumeent that is in fact colder than the air surrounding it. But that would also mean that in summer it would overheat because radiation could all to easily enter the screen. Which is not the case. Also the temperature was similar to other KNMI stations situated more inland. These were a few tenths of a degree colder.
When we look into the study done in Algeria in 2008 there was a small, but significant note on the Davis 7714 that is now used as a reference over here as it is supposed to be so good. The note was that the outcome was in contrast with what individual members found to be the case with that screen. I am guessing gthe collaborators did some tests in the field near their home and found less satisfying results. Also there was a notion of jamming ventilators inside artifically aspirated screens. Sadly: in both cases as far as I can remember none of those remarks were elaborated but I got the impression that the setting in Algeria was done with extreme radiation. From what I gathered there erre clear differences between lower and higher radiation situations per screen. I live at 52 degrees north, quite different than 30 degrees in a desert.
What I am trying to say is that they hinted the Davis did not do so well under less extreme circumstances, which are probably the ones most of us in Europe and northern USA/Canada live in. You can counter by saying that the most troublesome soiotuation still are those when we do get a lot of radiation.
Annyways back to the Meteoshield Pro: regardless of our preferences and our "needs" based on where we measure, it needs to be determined what causes what in my view. Helical shape vs black inside...What is it. I have not found anything meaningful here. I won't promise I will do some work on it next year in order to give an indication. I might. But I am quite busy and have many other hobbies so. But who knows.
The Barani setup is interesting because they do everything according to WMO standards. It is why I tried the lacklustre Peetbros sustem because for some parametres it did the same.