Re: Single vs Multiple documents

[Dave Crocker <dcrocker@xxxxxxxxxxxxxxxxxxx>]

John (et al),

There seems to be no simple basis for choosing one, larger document,
versus a series of smaller specifications.  It is an issue which
very much has two, reasonable sides to the coin.  So I will simply
cite the reasons we chose to offer three, very small documents
in the alternative SMTP extension specs:

The problem of keeping track of the various "parts" of the SMTP
spec already exists.  I agree with your concern about the difficulty,
but you already need, at least, the original SMTP spec and the Host
Requirements spec.  So, at a minimum, the extension spec is
yet a third. 

One might argue that we should "minimize" the added complexity of the
task, but a different view is to address the question of juggling
multiple references directly.

To some extent, that is exactly one of the Host Requirement doc's
jobs.  To a difference extent, that is the job of the STD document
series.  (STDs are a subset of RFCs; it refers to those specifications
which have attained the status of Internet Status and multiple RFCs
can be contained within one STD citation.)  I won't claim that these
two efforts (HR & STD) make the tracking task trivial, but they are
legitimate efforts along the right lines.

To the extent that one might argue that the extension document, itself,
should contain all the proper citations, to define a full and complete
STMP implementation, I'll note that that wasn't within the scope of
the SMTPext working group -- though I'll also note that it seems a
worthy task.

And lastly, there is the aesthetic issue of large-vs-small.  Smaller documents
seems more tractable.  Tractable to develop and tractable to read.  

Efforts that pursue large and complex goals have a difficult time.  When
possible, dividing the effort into smaller efforts makes things
dramatically easier.

Dave