That makes no difference:

     OSM    <-----> Wikipedia

      ^                 ^
      |                 |
      |                 |
      v                 v

     Atlas  <-----> Encyclopedia

It's just which parallels of the commutativity diagram do you focus on. The map-versus-text operator (left to right), or the digital-versus-dead-tree operator (top down).

It's logically equivalent, but it seems much easier to parse that OSM [is digital version of] an atlas vs OSM [is a map instead of text like] Wikipedia

I don't agree; to me, it looks like the the atlas <-> encyclopedia operator is a more complex, nuanced relationship than dead-tree <-> digital.

So that is to say, it's more interesting to focus on that relationship and look for analogies of it elsewhere.

More concretely, this equation seems to me to be mildly weird and interesting to try to solve for X:

     Bicycle  <----->   X 

      ^                 ^
      |                 |
      |                 |
      v                 v

     Atlas  <-----> Encyclopedia

This one doesn't seem to invite erudite thought:

      X  <----->    Wikipedia

      ^                 ^
      |                 |
      |                 |
      v                 v

     Fish <-----> Encyclopedia

This is because the difference between the Wikipedia and Encyclopedia is narrowly defined: one is digital, online and massively edited all the time. That tends to narrowly restrict the thinking to be along the lines of what is like a fish, but digital, online and massively edited? And that quickly fizzles out.

Another angle on it is that atlas + encyclopedia form a parcel; they are artifacts of the pre-digital world. Historically, people had an instance of each in their library, often on the same bookshelf. The original sentence preserves this encapsulation:

  digital: {X, wikipedia}  <--->  pre-digital: {atlas, encyclopedia}

What is X? Oh, possibly OSM. OSM is to Wikipedia as atlas is to encyclopedia.

The big divide is between pre-digital and digital. Users that use Wikipedia don't reach for a paper atlas. Users of paper atlases probably still use encyclopedias too.

The only way in which it might be confusing is that the encyclopedia and wikipedia seem to be a red herring. There is no need to bring them in at all simply to ask "what is the modern, digital equivalent of an atlas?" But, ah, this is not so because that loses a sublety. The digital equivalent of an atlas can be something quite different from OSM, such as the proprietary firmware and map data installed in a navigator. Like the Wikipedia, The OSM is (1) online, and freely accessible everywhere by anyone, and (2) editable by its users. The analogy to the relationship between an encyclopedia and Wikipedia is entirely relevant.

Thanks for the feedback, I updated it.

Do you think the other version is clearer, though? I think he’s right. I will probably update it.

I had to re-read the original sentence a time or two, but never thought it was outright wrong. I do think OSM : Atlas :: Wikipedia : Encyclopedia is slightly more clear, though.

OSM:wikipedia :: atlas : ?

- vs -

OSM:atlas :: wikipedia: ?

Both seem about the same to me(both clear).