[ih] History of Anonymous FTP

[Jack Haverty]

I doubt you can find specific times/places for use of such 
names/passwords.  IIRC, there was no effort to standardize such things.  
Some site would start to use a specific name/password, and other sites 
would adopt it if they chose to do so.

My recollection is that the introduction of "anonymous" FTP happened 
when electronic mail started being handled by mail servers. 
Specifically, I recall that I was implementing a mail server on MIT-DM 
(which didn't believe in any passwords at the time), but other systems 
were less cavalier about security.

Prior to mail, FTP usage was pretty similar to Telnet usage.   If you 
wanted to interact using FTP with another machine on the ARPANET, you 
had to have an account there so you could log in to that machine.   Then 
you could move files with whatever permissions your account had on that 
machine's file system.   Some machines had "public" accounts that were 
available (IIRC that was how SRI-NIC provided the Network Information 
Center where all sorts of useful files were available.)

That wasn't workable for mail, since it was desirable to be able to mail 
someone even when you didn't have an account on their system.    Also, 
the "server" had to start running before it knew which user it was 
working for, and somehow deal with mail sent to multiple users.

For systems like ITS on MIT-DM, that wasn't a problem.  For systems like 
MIT-MULTICS (and no doubt others), it wasn't clear how to build such a 
server which wouldn't be operating for any specific user, especially as 
it received incoming mail.

Some systems had trouble accommodating such issues created by 
networking.  E.g., I remember discussing this with Ken Pogran, who was 
implementing mail on MIT-MULTICS, which tried very hard to enforce a 
"secure" system where all users were protected from each other.   Which 
account should it be running on?   How does it switch to another user 
when it later becomes clear who is to receive the incoming mail.   (Ken 
probably remembers more).

So, I *think* the "Anonymous" convention was created as a way to have a 
"user" that could be associated with the activities of a mail server, 
where it wasn't possible to tell which real user of the system should 
actually be associated.   That affected things like allocation of costs, 
disk space quotas, permissions to read/write parts of the file system, 
etc.   Computer OSes would have a "user" who ran the mail server, which 
of necessity had to be able to "serve" mail for all of the real users 
who had accounts there.

This was certainly in the early 70s but I can't remember exactly when.   
IMHO, it was probably the first, or at least early, instance of the 
shift from the "users on their computers running their jobs" model into 
the richer "distributed system" structure with lots of interacting 
servers performing tasks for all users throughout the network, even when 
a user is not "logged in" at the time.

Abhay Bhushan's office was down the hall from mine, and I remember 
discussing various issues with him about FTP's support for "mail". In 
particular, I lobbied for MLFL as an addition to the MAIL command in FTP 
in order to solve some problems which came up while implementing a mail 
system on MIT-DM.   Abhay released several RFCs in the early 70s 
refining the FTP spec, and it's possible that "Anonymous" was included 
in one of those, which might supply a timeframe for when it happened.

So, IMHO some systems were unwilling to easily make that shift as the 
ARPANET took hold.   There had to be "The User" for any compute job to 
run, as had been required since the days of punch cards. "Anonymous" was 
that User.

/Jack Haverty
(JFH at MIT-DM 1970-1977)

On 9/29/21 10:12 AM, touch--- via Internet-history wrote:
> Hi, all,
>
> Does anyone remember when each of the following happened:
>
> 	anonymous/no password
> 	guest/no password
> 	anonymous/anonymous
> 	anonymous/guest
> 	guest/guest
> 	ftp/ftp
>
> I recall seeing these variants (there may have been others).
>
> Joe
> —
> Joe Touch, temporal epistemologist
> www.strayalpha.com
>
>> On Sep 29, 2021, at 5:39 AM, Dave Crocker via Internet-history <internet-history at elists.isoc.org> wrote:
>>
>> On 9/28/2021 8:56 PM, Darius Kazemi wrote:
>>> There is of course Padlipsky's anecdote (that probably won't settle anything to anyone's satisfaction) on the invention of anonymous login for FTP:
>> While one should never take an "I'm the inventor and here are the details" statement on it's own(*), the existence of it can, of course, be helpful.  I'm not an historian, but I'd assume that it would count as part of a collection of bits of complementary information, to form a 'likelihood' assessment.
>>
>> So, for example, I've been saying that I know Anonymous FTP was not in the initial use but appeared quickly, and I guess around 1974 or 1975. Padlisky's claim of 1973 is close enough, IMO, since that leaves some buffer for propagation (and impinging on my slow awareness.)
>>
>> So, no, it doesn't 'settle' the question, but I'd count it as helpful. Thanks!
>>
>> d/
>>
>> (*) As least he didn't claim to invent email.
>>
>> -- 
>> Dave Crocker
>> Brandenburg InternetWorking
>> bbiw.net
>> -- 
>> Internet-history mailing list
>> Internet-history at elists.isoc.org
>> https://elists.isoc.org/mailman/listinfo/internet-history