Multics > Library > Articles
2000

And They Argued All Night...

Mike Padlipsky

Mike Padlipsky died in 2011.

Mike Padlipsky

{Having very belatedly discovered that I'd uploaded a damaged version of the as-written version of this one, what follows is a combination of the as-written and as-published (in "Matrix News") versions. --MAP}

...over whose claim was right: first at which, and for what, and with whom.

Copyright © 2000 by the author.

[Over the 15 years from 1971 through 1985, Mike Padlipsky authored over 20 RFCs. He ran the Networking Group for the Multics machine at MIT during the early years of the ARPANET, among other things. His The Elements of Networking Style remains the best book on OSI. - Ed.]

Despite my stated views on the futility of "doing" technohistory, I can scarcely find it in what's left of my heart to decline Peter Salus's invitation to take a turn in his Lest They Forget/Be Forgotten series, especially since I've been itching for at least a couple of years to have an excuse to dispute various selfserving claims of Fathers of this and Inventors of that now that the 'Net is the darling of the mediaocrities. And the best place to start is almost certainly with a restatement of why I think technohistory can't be done (lifted from The Elements of Networking Style and Other Essays and Animadversions on the Art of Intercomputer Networking, new ISBN 0-595-08879-1, presumably available via backinprint.com by the time you read this -- which I can now do without any fear of the original publisher, since the copyright reverted to me when They declared The Book out of print and destroyed the remaining inventory without keeping Their promise to inform me in advance if They were ever going to do so ... which didn't surprise me, since I'd long had proof They'd only been paying me royalties on every other copy They sold all the long):

Around [20, by now] years ago I ran into an Old Network Boy friend (Gary Grossman, actually) I hadn't seen for a while. We got to reminiscing about the good old days five or so years previously on the "New" Telnet [...] working group and I said, "You know, I always liked your IAC trick." ( [...] a fairly important second-order mechanism of the ARPANET'S second-pass Virtual terminal protocol.) "Wait a minute," he replied, "IAC was one of yours ."   [Page 24, if you're curious about the fuller context.]

He was right, by the way. The neat ideas zipped around the design meetings at a great rate of speed (and volume, usually; in both senses of the term). Even before we all hit at best EarlyMiddleage (and I fear I crossed over into MiddleMiddleage last May; most of the others still have a few years to go, though, damn them) we didn't really remember who'd come up with which neat notion. Indeed, over at least a 22-year period before the untimely and intensely lamented death of Jon Postel, he and I had frequent conversations trying to reconstruct the origins of a number of neat notions and we almost never succeeded, even on many of the ones I was fairly sure had been his (which, in fact, was most of them). And before the 'Net became big business, it was fairly easy for us to shrug it off; things had worked out, and what did it matter whether he, or I, or Gary, or even one of the BBN guys -- who always seemed to get to write the histories and hence always seemed to have claimed to have invented everything, anyway, perhaps because BBN was the only "for-profit" to furnish key members of the original Network Working Group -- had actually been the first to enunciate an idea that was almost always implicit in the discussion to begin with?

Now, however, there seems to be at least some celebrity value attached to that sort of thing, if not indeed some financial value. There might even just possibly be a minute amount of value to be attached to "historical accuracy", or "intellectual integrity", or some other hopelessly pre-GenX abstraction, but it would doubtless be a tactical error to espouse that sort of thing. Besides, as indicated, some of the current claims stick in my craw, and Peter did ask me to "write something" and didn't get at all specific as to what, so...

Let's start with the one I'm in fact quite certain I was the inventor of, especially because I can't recall which of the BBN guys is claiming it and so I can put off the delicate question of whether I want to name names for a while longer: "anonymous login". I remember pretty clearly, despite being somewhat unsure as to who the other person in the conversation was (Dirk, maybe?), being at SRI for some sort of meeting sometime around 1973 and being told that "the NIC" (or at least Jake [=Elizabeth Feinler - PHS]) was worried about this idea to put the RFCs on-line, because they'd have to establish all sorts of accounts so people could FTP them. "That's easy," I said, "just use my NETML trick." By which I meant, and went on to explain, that just as I'd had to propound a conventional universal "dummy" id and password so that netmail (as we'd called it when we were inventing it, but I'll get to that soon enough) could work via FTP without causing grave harm to the security (and accounting) mechanisms of at least some of the Hosts (mainly Multics, of course, since I was the Multics Network Technical Liaison at the time), all the NIC needed to do was establish a single, known account everybody could use to slurp the RFC's from. "'guest' would be a perfectly fine id," I went on, "and the password should be 'anonymous', since we'd gain some measure of security in that people'd have to know how to spell it and of course not everybody does." Or words very close to that, and to exactly that effect, even if I actually gave the id and password values in the reverse order. [The NETML trick was enunciated in RFC 491, in case you care -- and in case it ever gets scanned in so you can care. -- MAP]

Now, EarlyMiddleAgedMemory (EMAM) being what it was, and MiddleMiddleAgedMemory (MMAM) being what it is, naturally I don't recall whether I read that one of the BBN guys was laying claim to "anonymous login" the other year or saw it on one of those overly-coy little "courses" PBS (the P is for Pious, I like to observe) has taken to showing, before I decided they annoyed me so much I won't watch any more. (Even if I were on one? Irrelevant question. I didn't get rich, nor did I get my claims to've invented things into "the literature" early enough, because my company didn't get commissioned to write the "First 10 Years" report and wind up being visited first by the Internet history book writers.) But I submit that anybody who knows me knows that the crack about security has to be one of mine, and MMAM insists that whoever was laying the claim gave it as his own, which makes me suspect that the charitable explanation of parallel evolution doesn't apply and it was either his EMAM in play or just plain theft of intellectual property. Fortunately, anonymous login isn't really anything to be proud of -- especially since it was exploited for a famous security breach when it was misimplemented on a certain highly-popular Host type -- so I can rise above.

Digression on Internet History Books

I'm not ready to name names yet, but I do want to offer a thought or two about this business of who gets talked to first by the historians. The authors of a couple of Internet history books did, one way or another, happen to get in touch with me during their "research" phases. Both had been in touch with BBN first. One of the authors did do me the courtesy of not buying the BBN version of the anonymous login story when the topic came up during our e-conversations. The other author apparently "bought into" the BBN version of just about everything (I've been warned by several friends not to read the book, in deference to my G.I. Woe), and on happening to meet me a couple of years after cancelling out on a scheduled interview, said "You don't look at all like what I'd expected" and then wandered away, without even letting me explain why my reply had been, "You mean because I'm not 7 feet tall?". Maybe I'm wrong, but the only explanation I could come up with was that the BBN guys had described me with about the same degree of objectivity as Shakespeare had applied to Richard III, and the cognitive dissonance -- is embarrassment over having been duped too much to hope for? probably -- was more than could be dealt with.

There might be some other reason for the apparent personal slight, and I'm taking care not to identify the author in question, but the fact still remains that enormous bias can be introduced into all history books, much less technohistory ones, based on whom the authors talk, or refer, to first ... and whom they find agreeable, of course. (Hey, at least the first author kinda liked me.)

(By the way, the "7 feet tall" crack came from the time I once got a telephone call at work from a programmer who worked for Univac who'd been astute enough to have figured out that I was the right person to ask about the design intent of the Telnet Protocol, and when it came out in the course of the conversation that I'm not above average height she said, quite unsarcastically I believe, "But you write as if you're seven feet tall!")

Back to the Main Thread, Such as it Is

OK, here goes. It'd be not only disingenuous but downright cowardly not to name a couple of names. So first off, I don't believe Ray Tomlinson invented "e-mail". And not because of the quibble that we called it netmail originally, though that does offer an excuse to observe that I personally find the term "e-mail" awfully cutesy, and references to "sending an e-mail" syntactic slime. Nor because of the semi-quibble that "mail" had been around intra-Host on several of the Host operating systems since well before anybody realized they were Hosts, though that one has a great deal of abstract "historical" appeal. No, it's because I have a completely clear memory that Ray wasn't even at the FTP meeting where we decided to add mail to the protocol. Granted, one of the BBN guys mentioned that he'd done a TENEX to TENEX mail hack already, and that he'd used the "@" between id and host name for the addresses, but that was after somebody had said, "Hey, why don't we send mail via FTP?" And for all I know, that somebody might have been me. (Or "been I", I imagine ... if you're a fanatic Syntactic Puritan.) I'm not saying it was, but I don't know it wasn't. What I am sure of is that I was one of the strongest proponents of doing it, whoever broached the idea, and that I could have vetoed the famous "@", but didn't.

Vetoed it? Huh? Well, it takes a little, um, er, history to explain that one. When the "TIP" (or Terminal IMP, the ancestral terminal support machine, and at least weakly arguably the grandfather of the "ISP") was introduced -- and MMAM assures me EMAM had long believed it was at the 1971 Network Working Group meeting in Atlantic City, though both could be wrong, as usual -- the BBN guys announced, with grins, that the "intercept character" was going to be the "at-sign" and all looked at me. The reason was that on Multics, the at-sign was the reserved line-delete character. I thought about it for a second or two (being much younger and quicker of wit then) and said something very much along the lines of, "I guess. After all, everybody hits it a few times to be sure it takes, and you tell me doubling it will let it get through, so it shouldn't be that hard to remember that if we're TIPping to Multics we need to hit it an odd number of times when we want it intercepted but an even number of times when we want Multics to get it." And with that in mind, since in those days we tended, or pretended, to subscribe to the "all researchers together" myth, I had one coming, and could have insisted that it was too inconvenient for Multics users to have to type \@ when sending netmail, so let's pick another character. Or even use the more Multicsy syntax of "-at". But since it was the at sign -- even though I could swear it was called the "commercial at sign" in my high school typing class, just to pick a nit now that I might well have picked then, too -- and since I am a Semantic Puritan, I couldn't bring myself to argue that it wasn't appropriate, and let it pass.

So when it comes to who invented e-, or net-, mail, I submit: that as usual it was a group thing; that Ray wasn't at the meeting where it became part of the 'Net, although, granted, he was, as far as I know, the first person to send mail from one computer to another over a network, even though they were like computers, which isn't particularly interesting or challenging even if the world's most over-"capped" corporation can't seem to do consistently better than that these days, and for all I know somebody at the National Physical Laboratory in England had already done at least the same, and maybe even used unlike operating systems (the Brits turn out to have better claims to've invented a surprising number of things, I've noticed over the years, but apparently the historians' travel budgets are never adequate...) -- and he did have the insight to realize that "@" is commonly pronounced "at", of course; and that if I really wanted to I could always pretend to be sure I'd been the one to propose doing mail via FTP at the meeting and I don't think anybody who was there and who is intellectually honest could gainsay me, but I don't really want to because I'm simply not sure who proposed it. I am, however, completely sure that it was implicit in what we were doing, so in a fairly real sense we all invented it.

One other thing about mail: Realize that FTP-based mail required the other Host to be up at the time (though people did come up with queuers for delayed sending if it wasn't), and it's arguable that what made "e-mail" really take off was when the model was changed so that that was no longer the case. The change came with SMTP. Jon was, of course, the author of that protocol, and knowing him he probably was the one who proposed it at the meeting, but I wasn't at the meeting so I don't know, but I'm perfectly content to assert that his claim to inventing "e-mail" is as good as anybody's.

"It's a Wise Child that Knows Its Own Father"

Or is it "It's a wise father that knows his own child"? (Or, in PC -- for Philologically Corrupt -- contempobabble: It's a wise child that knows their own father, maybe. Let's hope it wouldn't have to be their own parent in PCorruption of the first version, anyway, which MMAM does think is the original: the use of "their" with a singular antecedent to avoid choosing between "him", "her", or "him or her" is nauseating enough; if They insist on destroying the sense by ignoring that motherhood is a matter of fact, fatherhood is the one that's a matter of conjecture, They're not even worth trying to talk to.) Anyway, let's get to this "Father of the Internet" crap.

Normally, I'd say "In my humble but dogmatic opinion", but I don't want to be misunderstood. In my considered opinion, Len Kleinrock's claim to be the father of the Internet is ... no, don't get "personal", make it, I know: unfounded. And/or lacking in merit. And/or utterly unfounded. And/or utterly lacking in merit. And/or ... get the picture?

Now, between the word count Peter asked me to live with and the fact that my carpal tunnels are on the brink of cave-in after some 37 years of "keyboarding", I'm not going to go into much detail as to why I think a number of people's claims for that fundamentally meaningless but emotively compelling title are better than Len's, though I'll touch on some of them briefly. And perhaps surprise you by not including myself among them. But for Clio's sake, if she was the Muse of History, or for x's sake assuming Peter will fill in the x for me [Clio is correct. - PHS], all Len was was the faculty guy who got the grant from ARPA to do the Network Measurement Center, where the first IMP happened to be installed. That doesn't even make him the father of the ARPANET, which ISN'T the Internet, if you're not a mediaocrity and care something about technical accuracy as opposed to "selling papers". Granted, three people who worked for him have good claims to the title -- far, far better claims, imhbdo -- but that would at best let him claim to be the grandfather, or more accurately great-grandfather, not the father. And depending upon how you define "the Internet", it's eminently arguable that their claims aren't as good as those of three other guys, whose efforts led to Len's getting the grant. (There's a fourth guy who worked at the NMC, who's still a good friend, whose claims to netpaternity are no better than my own non-existent ones, but he was also an important "player" in the early days and he's so incensed by Len's claim that in his honor I'm spending more time, and subjecting my wrists and fingers to more wear and tear, on this point than I might otherwise have done. However, the "all Len was" point just made is, at some level, sufficient refutation.)

By the way, I left out one strikingly negative (imhbdo) contribution Ray made in order not to get "personal", and I'm leaving out two (ditto) Len made for the same reason; but I'm not anywhere near noble enough not to mention that I'm doing so. But wait: I've belatedly realized that I also left out an amusingly discreditable sidelight on the second "history book's" author, so I fear nobility has nothing at all to do with it, it's just that I seem to be getting soft, in my MMA. Not soft enough, after due deliberation, to omit mention of one other craw-sticker, though, namely the ... interesting aspect of the BBN "corporate culture" which led the nominal author of the "New" Telnet RFC to effusively acknowledge the assistance of his BBN colleagues in the document and totally ignore the fact that not only had I edited it extensively but around a quarter of the actual words printed were mine. (Not that I choose to offer him the publicity of mentioning his name, of course, but he's also the one who declined to co-sign my Host-Front End Protocol RFC despite the fact that 20some others in the NWG did and he'd privately told me he agreed with it, on the stated grounds that "it wouldn't be good for the company" to admit publicly to being for something that would make more work for the IMP -- which tells us some more about the corporate culture, I'd say. Indeed, MMAM belatedly assures me that for a number of years I was quite fond of saying that we'd done the 'Net despite BBN, not because of it, but I'm already over the word-count as I add this and probably couldn't remember enough of the points on which that was based anyway, so will let it go at that.)

"Historically speaking", let's touch on the three without whom Len wouldn't have gotten his grant in the first place: J.C.R. Licklider, Bob Taylor, and Larry Roberts. Lick is usually credited with having had the germ of the idea to network computers, and I liked him so much personally I can't be objective: if there must be a "the" Father of the 'Net to satisfy the mediaocrities, then as They seem to non-define the 'Net, Lick's the one. Period. Bob Taylor, whom I didn't know personally but as his wife happened to tell me at a party at Bob Metcalfe's apartment in 1978 did pick up the torch from Lick and hired Larry Roberts to run with it, has a very good claim too. (Speaking of Metcalfe, a non-PBS show I made the mistake of watching recently semidemihinted that he, as the acknowledged "father" of Ethernet, might have some claim to the Internet title as well. I don't think he admitted the soft impeachment, as it were, but just for the record the Internet wasn't built to link together only Ethernets, but in fact various and sundry comm subnet "technologies", nor did Bob have anything to do with developing IP, as I recall, so that's a non-starter.) And Larry was introduced at the '71 NWG meeting with the phrase "And now a word from Our Sponsor", so to those of us doing the work of building the ARPANET, Larry's claim feels pretty good, too -- especially to those of you who weren't fortunate enough to've worked in the same building as Lick for what, five or six years, one or two of which he was your boss's boss's boss.

Nor, turning to the three who originally worked at the NMC, should we overlook the person who was chairing the meeting and introducing Larry: Steve Crocker. Steve was hired by the ARPA office to run the Network Working Group, and deserves full credit for getting things going w/r/t the 'Net, even unto writing RFC #1. I don't feel at all filial about his role, but his claim still strikes me as better than Len's. Then there's Vint Cerf's role. He, too, started at the NMC and went on to the ARPA office. The difference is that if you define the Internet at all strictly, Vint's participation in the "invention" of TCP/IP coupled with his role in coordinating the contracts and contractors who implemented the protocol suite makes me feel that if he chose to say "I was the Father of the Internet" on his "web site" I might flinch a bit but I certainly wouldn't sneer. Finally, we shouldn't overlook Jon Postel, the way almost everybody did until we lost him and then people suddenly noticed how important, how central, how vital his participation had been all the long. He was the only one of the Old Network Boys whose views I took with fewer grains of salt than I took my own, of course, but the real point is that almost all the protocol RFC's were under his authorship, and not because he was that slick a writer. It's because he understood how the protocols had to work better than any of us, usually because the ideas they embodied were his, even when he wouldn't acknowledge that, after the fact.

At the risk of getting a bit sloppy here, and almost certainly commiting a metaphor that should have been blocked, I simply can't resist observing that it's not particularly interesting who the "father" of the Internet was, because what really matters is who its "mother" was, in the sense of whose labor gave it birth. And if you need me to spell the rest of it out for you, I assume you wouldn't understand it anyway...

Come to think of it, though, if that's all a bit too equivocal for you, maybe we should remember our Time Enough for Love (and if you haven't read everything Heinlein wrote at least twice, it's your loss) and consider how the fleshly version of the computer "Minerva" came to be: a chromosome pair from Justin Foote for math, a chromosome pair from Lazarus Long presumably for longevity, and so on. So maybe the Internet wasn't fathered or mothered but rather created out of Lick's genes for inspiration and, I dunno, charm, and Jon's genes for call it technical intuition or brilliance and maybe composure, or whatever it was that enabled him to forge consensus among all us prima donnas, and Vint's gene for, I suppose, leadership, and ... so on. (Gee, maybe that way even I'd get a piece of the action myself, since I'd like to think that my gene for, what, righteous indignation? Constructive Snottiness? belongs in the mix, too, for being willing to take up the cudgels in public against the ISORMite idiocy -- a/k/a "OSI" for those who haven't read The Book -- probably did help, some.)

"In Conclusion Bear in Memory"

Keep the password in your mind.... Oops, wrong song (but a wonderful one; do searchengineer it if you don't know it -- provided anybody's bothered to webulate it, of course). No, what to keep in mind is that from my point of view it's all rather moot, when you get right down to it:

Jon's been taken from us, after all, so sticking up for his "rights" is merely an exercise in sentimentality -- though not, one hopes, futility. (And I say "taken from us" advisedly, though this isn't the time or place to go into why since I'm being careful not to risk any libel suits by not being personally insulting, even to those who have insulted me -- nor am I afraid of the original publisher, since Truth is the classically best defense against libel charges if They were stupid enough to take me on even though I haven't even mentioned 'Em by name -- but couldn't possibly avoid it if I got started on the topic of whom I view as the villain of Jon's last years.)

And I, after all, went into de facto retirement years ago, when I left the second job in a row over not being allowed to smoke in my own office, so I've got nothing to gain from winning any technopaternity suits. (Not even the mild egoboo of a cameo on one of the tellyjoke shows, since I'd insist on approval of the final cut lest I be accorded the same gratuitously snotty treatment Ted Nelson -- eminently arguably the "Father of the Web", by the way -- was, and clearly that'd never be granted.) Unless ... gee, what if The Book reprint'd be a bestseller if it were advertised as being by the Father of ...? Nah, anybody who read it would still need to be able to deal with real sentences, and there are far too few of us left.

So unless some or all of the over-"capped" companies that owe their existence to us (cisco, Network Solutions, and, save the mark, AOL come immediately to mind) were to decide to give chunks of stock to the Fathers of this and Inventors of that in appreciation, all that's at stake is another one of those hopelessly pre-GenX abstractions: "for the record". (Actually, what I think would be far more seemly would be to award, say, 1,000, but even a couple of hundred would be nice, shares per RFC to the authors of all substantive [i.e., non-repetitive, in the sense of the monthly reports on Host traffic or even the every-hundred indexes] RFC's with numbers under 1000 [inspired by the time Interop gave out t-shirts to the authors of the first thousand RFCs, even if they only allowed us to pick one RFC number to have printed on it, which led some of us to agonize a lot]. That'd be nice. And Jon's lion's share of the shares could go to funding the Old Programmers Home, as we discussed when I first had idea ... the day I heard Netscape's stock was over 150, to which I'd also will the proceeds of my own 20some lots. Of course, with my usual luck if they did do it the market'd make October of '29 look like a mere blip before I could convert to tax-frees anyway, so it's all still moot, isn't it.)

Meanwhile, what price "resource sharing"? The Bandwidth Bandits reign, even turning off "autoload images" doesn't spare one from slithering signs, everybody's committed to being HTML-trendy, "browsers" are bigger than operating systems need to be, and the main difference between "the Web" and the Sunday paper is that the p-ads can be used for wrapping fish...

Let at least one tradition be upheld, though: we conclude with the usual

cheers, map

[ In the August, 1999, issue of Matrix News we published a piece by Len Kleinrock. Last month's MN carried the first installment of a reminiscence by Les Earnest. Future authors of memoirs in this series will include Steve Crocker, Alex McKenzie, and Dave Walden. - Ed.]