[ih] FTP Design

[John Day]

Guys, you forget.

Much of the discussions in FTP (and others of those meetings) was 
getting the terminology sorted out.  Things were far from as 
homogeneous as they are now.  None of these machines had been 
intended to talk to others.  They all had their own terminology for 
what they did.

And remember, the people involved had probably only seen their 
machine, which they knew inside out, at least when we started this. 
I remember "intense debates" over mundane things  like "open a file," 
"what does a write do" etc.  Not realizing that it meant very 
different things on different machines.  Sometimes it turned out that 
one operation on one machine, combinations of operations on another, 
impossible on a third, parts of operations, etc. etc.  But it took a 
lot of time to figure out that we were often talking about the same 
thing.  Then of course there were the limitations that each imposed 
and how to finesse them.

Doing FTP in those days entailed figuring out a lot more than you 
might have thought.  ;-)

At 11:47 -0700 2012/07/03, Dave Crocker wrote:
>well...
>
>On 7/3/2012 11:29 AM, SM wrote:
>>Hi Dave,
>>At 10:00 03-07-2012, Dave Crocker wrote:
>>>Quite a bit of current IETF work appears slanted more towards an
>>>initial completeness that attempts to satisfy the /union/ of
>>>participants' desires, rather than the earlier IETF mode of seeking an
>>>initial version that satisfied only the /intersection/, deferring the
>>>remainder for later enhancement efforts.
>>
>>The "process" is fashioned in such a way that it is easier to satisfy
>>the desires of participants than to argue.  That would not be much of a
>>problem if these "enhancements" were pruned by natural selection at the
>>next stage.  The rarely happens.
>
>In earlier days, participants had a sense of urgency, wanting to get 
>something that worked shipped as soon as possible.  This provided 
>pressure for deferring issues that created delay in the design.
>
>As other have noted in this thread, there was also greater 
>appreciation for what we all simplistically call elegance.
>
>It takes more diligence to create tight designs.
>
>>It's a bit more complicated than that.  There is a larger group of
>>people, more external pressure and formalism.  I wonder whether the
>>following would be palatable nowadays:
>>
>>    "The objectives of FTP are 1) to promote sharing of files"
>
>I don't think formalism has anything to do with it; we been formal 
>for a very long time.
>
>As for 'external pressure' I think the loss of urgency actually 
>represents a reduction in pragmatic pressures.(*)
>
>As for the number of people involved, I actually believe the size of 
>many ietf group's active core has gone down, not up.  It's pretty 
>typical to see only 3-5 people being active, now.
>
>d/
>
>(*) A few years ago, I heard an argument that the extended duty 
>cycle for IETF work -- multiple years to produce a standard, in a 
>world with 6-12month product cycles -- can get someone fired because 
>their work is of little relevance to products.  At the least, this 
>means that the folk who attend the IETF often are not the senior 
>engineering talent that we used to attract, but rather professional 
>standards folk.
>
>--
>  Dave Crocker
>  Brandenburg InternetWorking
>  bbiw.net