Should highway type be set according to physical characteristics?

In my local community in a developing country with poor and variable infrastructure, there is disagreement on how to classify the highway system, and there are basically two competing philosophies: that they should be classified by official status, and that they should be classified by physical characteristics, as those can vary significantly along roads of same official status.

I’ve found the following comment in the wiki:

Would you agree?

I know that, in the UK, where the OSM terminology for highways (trunk, primary, secondary, tertiary) originated, a highway’s class is defined by official status, and is indicated by signs, so it is very easy to decide which class a road should be. Wikipedia on trunk suggests that physical characteristics play a minor role in determining a road’s class in the UK. How about other countries? Do they follow the same philosophy, or has any country adopted physical characteristics to define some/all highway classes?

I’m writing from Brazil, and I’ve been using two regions as examples of extremes we have here: São Paulo (densely populated, rich) and Amazonas (sparsely populated, mostly just the Amazon rainforest, struggling to get resources for development). In São Paulo, roads are of a high standard, comparable to those in developed nations. In Amazonas, many of the main national routes between its capital and adjacent states’s capitals are unpaved, officially mostly of the surface=compacted type, but with maintenance issues in some parts. Two examples of such are BR-319 (full voyage) linking state capitals Manaus and Porto Velho, and BR-230 linking those capitals to state capital Santarém.