[ih] History of Network Virtual Terminal (NVT)

[Matthias Bärwolff]

Thanks for your suggestions so far; after searching around in early BBN
reports and RFCs I have found three references on the history of NVT
which might be of interest to the discussion:

BBN Report 1928 (the Quarterly Technical Report No.4, 1 October 1969 to
31 December 1969} states:

"While we have come to believe that the IMP should not do character set
conversions, there is still an immediate need for a network-wide
teletype character set into and out of which each Host translates his
messages. The choice is arbitrary, and the need for a decision has
become urgent (already we see Hosts converting to the language of the
destination). We recommend the adoption of 8-bit ASCII with the 8th bit
(checksum bit) set to 1, which is the IMP's internal character set. This
choice has the small additional advantage that Hosts may send messages
to local or remote IMP teletypes without an additional conversion.  As
network use develops, other standards (such as a display language) will
be needed."
(p.\,9)


Then, Steve Crocker's RFC 66 (of mid-1970):

"We next agreed [in a meeting with BBN and MIT representatives] on an
initial network standard console: 7-bit ASCII in 8 bit fields with the
eight bit on, transmitted in contiguous streams.  The specific codes are
listed in appendix H of the IMP Operations manual, BBN report #1877.
This seems to work only some hardship on PDP-10's and be fine for all
others."
(p.\,3)


And, RFC 137 (on Telnet) refers to an "Official Network Virtual Terminal
Code" already. So, by early 1971 at the latest seems there to have been
a firm notion of NVT.

I would think, though, that the TIPs pushed the whole thing quite a bit,
being in a prime position to enforce a standard on this. From the
Ornstein et al (1971) paper:

"Because of the large number of different terminal types used in the
network, the concept of the Network Virtual Terminal was developed. This
is an imaginary but well-defined type of terminal. The TIP translates
typed data to virtual terminal code before shipping it into the network,
and conversely translates the remote system's response back into the
local terminal's code. Thus, each Host system must deal only with this
single terminal type."


Still the precise origin is somewhat foggy. But then again, the concept
seems in a way obvious, so maybe whoever came up with the details (like,
how large or how small to make the standard terminal set, trading off
breadth of applicability and feature-richness, and whom to favor) did
not think it necessary to make a big fuss about it.

Matthias



John Day wrote:
> I may be wrong, but I thought the NVT idea first showed up in the
> "new" Telnet in mid-72.  I had always thought it originated with Alex
> MacKenzie, but he swears it wasn't him.
>
> Take care,
> John
>
> At 10:38 -0500 2010/01/22, Vint Cerf wrote:
>> this was an idea that emerged from the network working group.
>>
>> Steve Crocker led that group and might have better answers.
>> this was not something that BBN designed - it came from the NWG,
>> of course BBN had people involved in the WG. As I recall, Dave
>> Walden was a strong proponent of the Do/Don't/Will/Won't idea.
>>
>> vint
>>
>>
>> On Jan 22, 2010, at 7:33 AM, Matthias Bärwolff wrote:
>>
>>> Hi all,
>>>
>>> I've been trying to find out about the actual origins of NVT, but
>>> cannot
>>> find statements other than (1) that a common terminal representation
>>> would be good, or (2) that it exists in the shape of NVT. So where does
>>> it come from? Did the BBN guys specify it when building the TIPs (which
>>> did just that: the character conversion between the local terminal and
>>> NVT)? Or did it emerge from the work on Telnet in the NWG?
>>>
>>> Thanks for all hints.
>>>
>>> Matthias
>>>
>>> -- 
>>> Matthias Bärwolff
>>> www.bärwolff.de
>

-- 
Matthias Bärwolff
www.bärwolff.de