[TUHS] Re: Mills' initial implementation of FTP - best citation?

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Noel Chiappa]

    > From: Paul Ruizendaal

    >> the ambiguous phrase "had the first implementation of FTP", which
    >> has been flagged as needing clarification

    > From RFC 354 ... and from RFC 414

Those are NCP FTP, a slightly different protocol, and implementation, from TCP
FTP. (The code from the NCP one was sometimes recycled into the TCP one; see
e.g.:

  https://github.com/PDP-10/its-vault/blob/master/files/sysnet/ftpu.161

which has both in one program.)

These RFC's you listed are obviously pre-TCP; the first TCP RFC is
RFC-675. (The first RFC that even mentions TCP seems to be RFC-661.)  RFC's
are all NCP-related until around #700 or so, when the mix starts to change.


Maybe the "needing clarification" refers to these two different FTP's? Without
an explicit classifier, does that text refer to NCP FTP or TCP FTP?

	Noel

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Paul Ruizendaal]

> On 22 Jan 2024, at 22:29, Noel Chiappa <jnc@mercury.lcs.mit.edu> wrote:
> 
> Maybe the "needing clarification" refers to these two different FTP's? Without
> an explicit classifier, does that text refer to NCP FTP or TCP FTP?
> 
> 	Noel

That is a good point. I looked at a longer section of his talk and indeed the FTP reference is part of his discussion the NCP->TCP transition. Here it becomes easier to find references to support the claim.

The strongest one that I could find is in the TCP-IP Digest mailing list from that era. The first issue has a summary of what is already out there and David Mills work is mentioned:

https://jaist-g.dl.sourceforge.jp/pub/RFC/museum/tcp-ip-digest/tcp-ip-digest.v1n1.1

====
COMSAT

   Date:  30 Apr 1980
   From:  Dave Mills <Mills@ISIE>

   1. The TCP/IP implementation here runs in an LSI-11 with a homegrown 
   operating system compatible in most respects to RT-11. Besides the 
   TCP/IP levels the system includes many of the common high-level 
   protocols used in the ARPANET community, such as TELNET, FTP and 
   XNET.
=====

Although this date precedes the date of the RFC by a few months, I do believe that it is correct and not a typo. For example, IEN-194 (July 1981) discusses the system from a perspective of experience, not of recently completed code.

All that said, I wonder if “doing the first FTP” is a true reflection of his contribution in this era. After all, porting NCP FTP to TCP FTP is not all that big a deal (see for instance https://www.tuhs.org/cgi-bin/utree.pl?file=BBN-Vax-TCP/src/ftp). Maybe the sentence on Wikipedia should say that beyond NTP, he was an important contributor to several other early internet protocols.

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[George Michaelson]

NIFTP was the UK Coloured book File Transfer Protocol. The JANET X.25
based comms stack which included different coloured fasicles for each
applications protocol. If memory serves me right NIFTP was grey book
but I'd have to go and check.

I worked on Yorkbox, the LSI-11 based X.25 system which had a UNIX
applications suite connected to X25 by a DR11-W link. I did
applications level bugfixes, I completely bolloxed it up, probably
wrecked the York University IPR in this project, pissed off my project
partners at least one of whom left because I was such a crap coder.
Happy days. Later on I worked at UCL at the end of the SATNET era when
Bob Braden wrote up the project:
https://dl.acm.org/doi/pdf/10.1145/1035237.1035286

Funny story: UCL was very unhappy with my ex-employer York Uni because
they had coded the LSI-11 work and then lost the contract to maintain
it. They kept operating a gateway from JANET to ARPANET which required
use of fiddly kermit logins to get from a PAD X.25 login into a system
to send real FTP commands, and there was a translator to take NIFTP
command streams and somehow gateway them into FTP like in JCL (there
was another applications stack in coloured. book for real JCL which
was coded at Bristol uni if I recall correctly)

BTW Noel is right that written reports outweigh faulty memory every
time. But my memory of what a bad job I did continues to haunt me.

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Noel Chiappa]

    > From: Bakul Shah

    > He was part of NSFNet, so could have got first FTP on NSFnet or a
    > later version of FTP.

You all are talking about _two separate FTP's_ (as I pointed out
previously). If you all would stop confusing yourselves, you'd be able to sort
out the bogons.

In this particular case, the NSFnet appeared at a _much_ later stage of the
growth of the Internet (yes, it is spelled with a capital 'I'; the morons at
the AP were not aware that 'internet' was a pre-existing word with a
_different meaning_) than when Dave was working with the Fuzzball, and by that
point there were _many_ TCP FTP's (e.g. the ITS one I previously sent the URL
to the source for), so the 'first FTP on NSFnet' is a non-concept.

The best bet for accurate data is to look at the TCP meeting minutes from the
IEN series:

    https://www.rfc-editor.org/ien/ien-index.html

Looking quickly, the first one that Dave appears in might be IEN-160,
"Internet Meeting Notes -- 7-8-9 October 1980". (He wasn't in the "Attendees"
lists of any of the earlier ones I looked at.) Look in the "Status Reports"
sections to see if he says anything about where he's at. The one for this one
says:

  "Dave described the configuration of equipment at COMSAT which consists of a
  number of small hosts, mainly LSI-11s. ... COMSAT has also used NIFTP to
  transmit files between their hosts and ISIE.  The NIFTP software was provided
  by UCL. ... COMSAT plans to .. arrange a permanent connection to the ARPANET."

I have no idea what a "NIFTP" might be. Also, there is a reason that serious
historians prefer contemporary written records, not people's memories.

	   Noel

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Bakul Shah]

[-- Attachment #1: Type: text/plain, Size: 1740 bytes --]

On Jan 22, 2024, at 8:15 AM, Paul Ruizendaal <pnr@planet.nl> wrote:
> 
>> I see that the wording on his Wikipedia page has the ambiguous phrase "had
>> the first implementation of FTP", which has been flagged as needing
>> clarification, so I intend to provide it.
>> 
>> In both this interview:
>> 
>> https://conservancy.umn.edu/bitstream/handle/11299/113899/oh403dlm.pdf
>> 
>> ... and this video recording of Mills himself giving a lecture at UDel:
>> 
>> https://youtu.be/08jBmCvxkv4?t=428
>> 
>> ... it's quite clear that it's literally true - he authored, compiled,
>> installed, implemented, and tested the very first (and apparently second)
>> FTP server.
> 
> It may be impossible to provide hard evidence. From RFC 354 it seems to me that the protocol took on a recognisable shape around July 1972 and from RFC 414 it seems to me that there were a number of implementations by November 1972, and unfortunately Dave Mills is not mentioned. His recollection may well be correct, but finding proof he was the first in a 4 months time slot 50+ years ago may be too ambitious.
> 
> https://www.ietf.org/rfc/rfc354.txt
> https://www.ietf.org/rfc/rfc414.txt
> 
> Maybe the internet history list can shed some more light on the matter:
> 
> https://elists.isoc.org/mailman/listinfo/internet-history

I asked Abhay Bhushan (author of the above two RFCs) and he had this to say:

Dave Mills recently honored was not part of FTP development and specifications.  Many groups were implementing early versions of FTP, so I don't know who got it working first. It was not Dave as he joined the Arpanet community later. He was part of NSFNet, so could have got first FTP on NSFnet or a laterversion of FTP.

FYI.

[-- Attachment #2: Type: text/html, Size: 2345 bytes --]

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Paul Ruizendaal]

> I see that the wording on his Wikipedia page has the ambiguous phrase "had
> the first implementation of FTP", which has been flagged as needing
> clarification, so I intend to provide it.
> 
> In both this interview:
> 
> https://conservancy.umn.edu/bitstream/handle/11299/113899/oh403dlm.pdf
> 
> ... and this video recording of Mills himself giving a lecture at UDel:
> 
> https://youtu.be/08jBmCvxkv4?t=428
> 
> ... it's quite clear that it's literally true - he authored, compiled,
> installed, implemented, and tested the very first (and apparently second)
> FTP server.

It may be impossible to provide hard evidence. From RFC 354 it seems to me that the protocol took on a recognisable shape around July 1972 and from RFC 414 it seems to me that there were a number of implementations by November 1972, and unfortunately Dave Mills is not mentioned. His recollection may well be correct, but finding proof he was the first in a 4 months time slot 50+ years ago may be too ambitious.

https://www.ietf.org/rfc/rfc354.txt
https://www.ietf.org/rfc/rfc414.txt

Maybe the internet history list can shed some more light on the matter:

https://elists.isoc.org/mailman/listinfo/internet-history

[TUHS] Re: Mills' initial implementation of FTP - best citation?

[Marc Donner]

[-- Attachment #1: Type: text/plain, Size: 1078 bytes --]

We need a code repository for original code like this.  Cue the backup
restoration.

On Mon, Jan 22, 2024, 02:59 Royce Williams <royce@techsolvency.com> wrote:

> What is the best public, unambiguous, non-YouTube reference I can cite for
> the late David Mills' initial FTP work?
>
> I see that the wording on his Wikipedia page has the ambiguous phrase "had
> the first implementation of FTP", which has been flagged as needing
> clarification, so I intend to provide it.
>
> In both this interview:
>
> https://conservancy.umn.edu/bitstream/handle/11299/113899/oh403dlm.pdf
>
> ... and this video recording of Mills himself giving a lecture at UDel:
>
> https://youtu.be/08jBmCvxkv4?t=428
>
> ... it's quite clear that it's literally true - he authored, compiled,
> installed, implemented, and tested the very first (and apparently second)
> FTP server. But Wikipedia's guidelines discourage YouTube-only citations,
> and the text in the interview seems insufficiently detailed to have
> citation value.
>
> What is the best reference I can cite?
>
> Thanks!
>
> --
> Royce
>

[-- Attachment #2: Type: text/html, Size: 1918 bytes --]