`no-attribute-formation` lint may be helpful

[yegor256]

How about we prohibit formations with no void attributes:

[] > foo
  42 > bar

This code should lead to a warning, since foo doesn't have any attributes. In Java, this would be a "utility class", which is a bad idea. Of course, we should not complain about tests.

[yegor256]

@maxonfjvipon WDYT?

[maxonfjvipon]

@yegor256 take a look at this org.eolang.go object:

[] > go
  malloc.of > identifier!
    8
    m.put m.id > [m]

  [body] > to
    try > @
      body token
      [e] >>
        if. > @
          identifier.eq e.id
          e.value
          error e
      true

    [] > token
      error > backward
        jump
          to body

      [value] > jump
        identifier > id

      [res] > forward
        error > @
          jump
            res

Here you can see token object which is passed to body scope.

go.to
  [g] # here g is token
    g.backward

Also check this org.eolang.structs.map.with object:

# Returns the new `map` with added object
# Replaces if there was one before.
[key value] > with
  hash-code-of key > hash!
  this. > @
    map.initialized
      with.
        origin.
          filtered.
            list entries
            (hash.eq entry.hash).not > [entry] >>
        [] >>
          ^.key > key
          ^.value > value
          ^.hash > hash

Here you can see auto named abstract object without void attributes which is used as kind of DTO with pre-defined key, value and hash attributes.

Such cases are met quite often and I'm not sure we should complain about them

[yegor256]

@maxonfjvipon why can't we replace them with:

[body] > go-to

[x] > map-initialized

I can't see any limitations for this. Do you?

[maxonfjvipon]

@yegor256 maybe we can but it's not in the scope of the ticket and it won't solve

[yegor256]

@maxonfjvipon I mean, we can make all objects attribute-full, can't we?

[maxonfjvipon]

@yegor256 I don't think so

[h1alexbel]

@maxonfjvipon maybe we can try this lint in experimental mode, and then try to re-write EO sources? WDYT?

[yegor256]

@h1alexbel good idea about "experimental" mode. Let's try to create such a flag for all our defects: Defect.experimental(). Then, eo-maven-plugin will not fail the build if such defects are found. WDYT? Let's make a ticket?

[h1alexbel]

@yegor256 extracted to #476