When Info was first created, it was customary to create many small Info files on one subject. Each Info file was formatted from its own Texinfo source file. This custom meant that Emacs did not need to make a large buffer to hold the whole of a large Info file when someone wanted information; instead, Emacs allocated just enough memory for the small Info file that contained the particular information sought. This way, Emacs could avoid wasting memory.
References from one file to another were made by referring to the file
name as well as the node name. (See section Referring to Other Info Files. Also, see section @xref with Four and Five Arguments.)
Include files were designed primarily as a way to create a single,
large printed manual out of several smaller Info files. In a printed
manual, all the references were within the same document, so TeX
could automatically determine the references' page numbers. The Info
formatting commands used include files only for creating joint
indices; each of the individual Texinfo files had to be formatted for
Info individually. (Each, therefore, required its own
@setfilename line.)
However, because large Info files are now split automatically, it is no longer necessary to keep them small.
Nowadays, multiple Texinfo files are used mostly for large documents, such as The GNU Emacs Lisp Reference Manual, and for projects in which several different people write different sections of a document simultaneously.
In addition, the Info formatting commands have been extended to work
with the @include command so as to create a single large Info
file that is split into smaller files if necessary. This means that
you can write menus and cross references without naming the different
Texinfo files.