According to the Arrow paradigm, everything is an arrow. It means that information is expressed in the form of arrows linked to each others. There is nothing like “nodes”.
However, arrows can be used to form signs. See:
The idea is to define a set of interlaced arrows so to make it impossible to find smaller closed constructs inside. One calls such a construct an “entrelacs”.
An “entrelacs” is fully defined and can’t be divided. It doesn’t contain any smaller entrelacs. One may say an entrelacs is equivalent to an atom.
And even more important: it’s actually possible to distinguish entrelacs from each other.
For example, the arrow whose both ends links to itself is usually named “Ouroboros” (or Eve in Entrelacs documentation). Two drawings of “Ouroboros” are two physical representations of the same entrelacs.
Consequently, even in a pure space of relations, entrelacs can be used as atoms so to define further Domain Specific Languages, like lambda-calculus.
The set of entrelacs is a countable set. It implies that by defining an adequate bijection, one can easily prove that any binary sequence is equivalent to a corresponding “entrelacs”. In conclusion, a binary sequence may be seen as the condensed form of a corresponding entrelacs.
A practical conclusion is that handling binary sequences in an arrows based system doesn’t deny the overall paradigm. All in all, everything is still an arrow. But it also implies that raw data should be proceed like regular arrows, especially regarding uniqueness and immutability.