I don't think these are
formal specifications? I'll give you an example, as you mention SFV, which is used to verify a file has not changed:
The HASH command. It is meant as a way to check the hash values of remote files just like SFV (though directly and not through files). But its authors take the correct way: There exists a formal specification (currently in form of a draft, if it is accepted it will become an RFC and therefore an official standard). Current version can be found here:
http://tools.ietf.org/html/draft-ietf-ftpext2-hash-02. An RFC then leaves no room for interpretations, every developer who wants to implement it, will know exactly how and all implementations will work together (unless there are bugs, ofc).
So unless SFV would be a formal standard (is it, I couldn't find anything?), everone could implement it as (s)he sees fit. Leads to confusion and non-interoperability brtween implementations.
The same for FILE_ID.DIZ. First, let me say that file is well known to me (along with descript.ion, files.bbs etc.), and I certainly recognize its usefulness. But let me quote one sentence from the Wiki article you posted:
Traditionally, FILE_ID.DIZ's should be "up to 10 lines of text, each line being no more than 45 characters long." according to the specification v1.9 — this restriction however is rarely observed nowadays.
What is a specification worth if it isn't formal and binding? Imagine one developer cares for the spec and another implements the loose format - the result is that the former software would fail to properly read the files created by the latter, and nobody can be blamed!