Carl Malamud: Internet talk radio, the medi­um is the message. 

This is Geek of the Week and we’re here with Dan Lynch, who is pres­i­dent and founder of Interop com­pa­ny, and he’s also a long ‑time mem­ber of the Internet Architecture Board. Welcome to Geek of the Week, Dan.

Dan Lynch: Well, thanks very much Carl. Glad to be here.

Malamud: You’ve been a mem­ber of the IAB for a long time, and you’ve been offi­cial­ly the indus­try rep. How does your role dif­fer from the oth­er IAB members?

Lynch: Well, the indus­try rep’s kin­da fun­ny since I actu­al­ly you know, have nev­er real­ly been a man­u­fac­tur­er of goods and ser­vices. But I…understand them more. I mean my back­ground was I was basi­cal­ly a com­put­er cen­ter man­ag­er for a lot of these com­put­er sci­ence researchers that devel­oped all of this tech­nol­o­gy. And so I got a chance to help make it work the first time. And the sec­ond time, and the third time. And under­stood the prob­lems that the man­u­fac­tur­ers, the reg­u­lar ven­dors out there in the world were hav­ing in try­ing to you know, cap­ture all of the knowl­edge and all of the meth­ods of this inter­net­work­ing technology. 

So, I kind of went to the IAB peo­ple a num­ber of years ago who were then sim­ply all sci­en­tists, and said, ​“You know, you need to have some­body help­ing you under­stand what the real world’s try­ing to do with this stuff.” 

And they said, ​“Oh, you’re the real world.”

I mean I was the real world to them, right. I mean, I was some­body who was­n’t a PhD in com­put­er sci­ence but almost under­stood it. And I could explain it to oth­ers. So they…you know, just sort of…I was a sur­ro­gate for all those peo­ple out there in the real world. And also since I was­n’t real­ly one of those bad ven­dors try­ing to make mon­ey, I could be hon­est, in their minds. I’m speak­ing a lit­tle bit out of school there, but I think you get the point.

Malamud: Has the IAB had to much of an aca­d­e­m­ic focus? Obviously when they start­ed it was a bunch of sci­en­tists, but has the orga­ni­za­tion changed over time so that it is tak­ing account of the real world?

Lynch: Oh yeah. Very very much so. In the past what, four years, three or four years, the new peo­ple that have come into the IAB have been some of them from—you know, hap­pen to be work­ing for ven­dor com­pa­nies like DEC [pro­nounced ​“deck”] and Data General and BBN. But they’re not there because they’re par­tic­u­lar­ly asso­ci­at­ed with any par­tic­u­lar man­u­fac­tur­er. It’s just they hap­pen to be real­ly good com­put­er design­ers and net­work design­ers, who oh by the way now that it’s impor­tant in the world are work­ing for real com­pa­nies. I mean, this was all just a grand exper­i­ment fif­teen or twen­ty years ago. And it got to be impor­tant, and when it’s impor­tant, some of the brain pow­er obvi­ous­ly is out there deliv­er­ing it to real live users. 

Malamud: Is there still a focus on the Internet in gen­er­al against com­mer­cial users? Do you still see that bias? Is it a research net­work still?

Lynch: It is absolute­ly not a research net­work. The fear— I mean…we’ve had the fear put into us. People are using it to make a liv­ing. And that’s won­der­ful. That’s called suc­cess, right. I mean you know, for most sci­en­tists, they want to design some­thing, build some­thing, that gets used. And the Internet gets used a whole lot. And so we’re all real hap­py that it’s being used, and very con­cerned that the archi­tec­ture of it, the design of it, the underlying…you know, thoughts that guide it are ade­quate to fill the needs of how peo­ple want to use it.

Right now, you know, you’re doing this show over the Internet. We did­n’t design it that way you know, twen­ty years ago. We designed a lot of flex­i­bil­i­ty in at the very begin­ning. But now it’s tak­en off very very fast. I mean, peo­ple are— MIME, you know, the mul­ti­me­dia mail stuff. Conferencing. This is tons of new activ­i­ties going on on the Internet that are…you know, real­ly out­side of every­one’s con­trol. I mean one of the beau­ty— I mean the Internet was designed to have no sin­gle point of con­trol. We have suc­ceed­ed. [laughs] There is no con­trol of that beast.

Malamud: There is a group of peo­ple, though, a large group of peo­ple that attempts to put tech­nol­o­gy out—standards and pro­to­cols, and you know, MIME is one them. And that body was orig­i­nal­ly just the IAB, and then the IAB begat a group called the IETF—

Lynch: Right, when the work got hard… [both laugh] …we said you know, maybe we need more help. And now the IETF is now five or six hun­dred peo­ple, right. And the IAB is a dozen peo­ple or so. And then what hap­pened what, two or three years ago we real­ized the IETF was so big that it need­ed its own oper­at­ing infra­struc­ture, which became what we call the IESG now the Internet Engineering Steering Group. And those peo­ple, the peo­ple who’re run­ning that and doing that are real­ly the ones clos­est to the prob­lems any­more, right. And so the IAB’s role has changed from being one of being inti­mate­ly involved in archi­tec­tur­al design and detail, to being more of…are we going in the right direc­tion? Are the right ques­tions being asked? You know, and overview kind of activ­i­ty as opposed to an oper­at­ing activity.

Malamud: That evo­lu­tion has­n’t always been very smooth. There was an inci­dent after the IP ver­sion 7 deci­sion or non-decision. And there was a lot­ta soul search­ing and there was almost a minor palace revolt. Do you think the cur­rent struc­ture with an IESG mak­ing stan­dards deci­sions and an IAB out there is gonna last? Are we gonna be able to move for­ward and con­tin­ue deploy­ing tech­nol­o­gy through these results on a peri­od­ic basis?

Lynch: Well… I mean. The easy answer’s I don’t know. If I could pre­dict the future that well I could you know…be find­ing oil instead of sit­ting here. But it’s… I think we’re going to go through a few more steps. And the palace revolt that occurred last year…is no prob­lem. It’s just, you know, more peo­ple anx­ious to get in there and row this boat. And a lit­tle dis­agree­ment about exact­ly how to do it, but there’s no dis­agree­ment about what we all want, which is to have this thing dri­ve, right, and serve mankind. And how it gets done is you know, always a mat­ter of judgment. 

It’s kind of like… I hear these argu­ments and read these argu­ments on the mail­ing list that say oh, we’ve got to make sure we choose the right tech­nol­o­gy, the best tech­nol­o­gy, this should not be a polit­i­cal deci­sion, blah blah blah and all those sorts of things. Well… I… That sounds won­der­ful. But it’s not being real.

What’s real is that the Internet has grown. It’s an obvi­ous suc­cess. It’s a suc­cess pri­mar­i­ly because of its orig­i­nal tech­nol­o­gy of let­ting peo­ple just stretch it any way they saw fit. So, essen­tial­ly what was designed was a very flex­i­ble com­mu­ni­ca­tions tech­nol­o­gy. And now it’s a suc­cess. And peo­ple are makin’ mon­ey. I mean, peo­ple are using it. So if the end users are suc­cess­ful in their lives by using the tech­nol­o­gy… And there’s bil­lions of dol­lars a year worth of gear being sold to real com­pa­nies, okay. And employ­ing real peo­ple you know, design­ing, build­ing, ser­vic­ing, all of this tech­nol­o­gy; the inter­net­work­ing tech­nol­o­gy both in the infra­struc­ture stuff like routers and bridges and phone lines, and things like that. And also you know, all the appli­ca­tions in the end systems. 

So, it’s too impor­tant to be left to sci­en­tists. It’s too impor­tant to be left to any­body to make all the deci­sions. I sort of believe what we have right now is there’s momen­tum. I mean, it’s mov­ing. There’s a mar­ket­place our there. If the IETF, IESG, IAB or what­ev­er is the right place…if it does a good job of advanc­ing the tech­nol­o­gy, of advanc­ing designs so that more peo­ple can get more things done, then great. Then if the cur­rent place where all that work is being done, and there’s a lit­tle bit of heat and light about you know, who’s run­ning it, how it’s being orga­nized. But if that turns out to be too sludgy a process, or too much…you know, if it’s not mak­ing for­ward progress—that’s the sim­ple measure—then some­thing else is gonna hap­pen, some­one else—something else is going to hap­pen in the real world. And you know, a com­mit­tee will form itself of five or ten or twen­ty peo­ple who say you know, ​“A pox on all of you, we’re gonna go off and you know, here’s IP ver­sion 7.” Whatever it is. Whether it’s TUBA or SIP or IKE[?] or some com­bi­na­tion of them or CLNP. I mean, they’re just gonna say that the mar­ket demands it. Customers demand it. Okay we’re run­ning out of address space, we’re run­ning out of man­age­abil­i­ty of the rout­ing tables, and we’ve got to solve the prob­lem. And you know, nobody’s come up with a straw man. And you know, the beau­ty is they don’t need any­body’s per­mis­sion. There’s real­ly no per­mis­sion…the per­mis­sion is the per­mis­sion of the buy­ers. If the buy­ers are say­ing, ​“I like this, this does what I want. Here’s some mon­ey. I’ll take some and I’ll run it.” You know, the IAB does­n’t have any…holiness about it. Nor does ISO. I mean, you know, ISO is a UN treaty activ­i­ty, okay. But it’s real­ly just human beings say­ing, ​“Okay, you guys are going to run this for us, and we’ll take what you put out.” But if you don’t put out rea­son­able stuff no one uses. And so I’m not too wor­ried about the pol­i­tics of where the good ideas or the next ideas come from. I see the momen­tum of the mar­ket­place using all this stuff is so big that there will be a solution.

Malamud: You’re lis­ten­ing to Geek of the Week. Support for this pro­gram is pro­vid­ed by Sun Microsystems. Sun Microsystems, open sys­tems for open minds. Additional sup­port for Geek of the Week comes from O’Reilly and Associates, pub­lish­es of books that help peo­ple get more out of computers. 

Don’t touch that mouse, Internet talk radio will be right back.

[Incidental Tourist seg­ment omitted]

Internet Talk Radio. Asynchronous times demand asyn­chro­nous radio. 

You don’t think that momen­tum is gonna stop. The Internet is a glob­al infra­struc­ture and it’s going to be here for a long time?

Lynch: Oh yeah. 

Malamud: It’s not just a step­ping stone towards the real intel­li­gent network?

Lynch: Uh…no you know, I— One of the beau­ties of the Internet is its sim­plic­i­ty. There’s a few sim­ple rules that it has. I’m a member…or asso­ci­at­ed with, I’m a mem­ber of the board of the Santa Fe Institute, which is a non­prof­it think tank kind of activ­i­ty of mul­ti­dis­ci­pli­nary sci­en­tists. I mean, the peo­ple are sci­en­tists of either med­ical or math­e­mat­i­cal or com­put­er, eco­nom­ics and archae­ol­o­gists. And they get togeth­er and basi­cal­ly teach each oth­er their own meth­ods of how they do research and how they explore in their own fields. And of course what they find out is that a method that works in one field is quite applic­a­ble to anoth­er field. Immunologists, real live immu­nol­o­gists describ­ing how the autoim­mune sys­tem works in a human being have been lis­tened to by com­put­er sci­en­tists who lis­ten that and say, ​“Hey, that sounds like the right way to design and write virus pro­tec­tion soft­ware.” It’s fun­ny the word ​“virus” we stole from them, okay, actu­al­ly to describe bad pro­grams or naughty pro­gram­mers, is actu­al­ly the source of the cure, right. I mean, go back to the same field and learn from them. 

So any­way, that’s all a big segue that says that the sim­plic­i­ty of the Internet is its strength. And all it takes is a few sim­ple rules. And the sim­ple rules of the Internet are… Let there be no one in con­trol. Let asso­ci­a­tions form, let peo­ple just con­nect to each oth­er who wish to. Put the respon­si­bil­i­ty for the com­mu­ni­ca­tion in the end sys­tems. That’s real­ly a corol­lary of let no one be in con­trol. I mean because every­one if you wan­na have some assur­ance that some­thing hap­pened, the insur­ance has to come from some­where so it real­ly has to be you know, in the end sys­tems if it’s not in the network—you were talk­ing about the intel­li­gent net­work before. I mean that’s…you know, the phone com­pa­ny PPC right, okay, want­i­ng to sells tons and tons of ser­vices when in fact they and no one can imag­ine and bring to bear all the ser­vices that peo­ple real­ly can dream up by them­selves. You’re real­ly giv­ing peo­ple a very fun­da­men­tal new com­mu­ni­ca­tion capa­bil­i­ty with the Internet tech­nol­o­gy. And it’s just fool­hardy to think that some­one who can design all that in advance and put it on a tar­iff sheet and for­ev­er more you’ll be eter­nal­ly hap­py with those ser­vices and those ser­vices only.

So the Internet real­ly is a groundswell kind of thing. I mean, you were INET with me Kobe, Japan last sum­mer, okay, and saw and felt the emo­tion of those peo­ple from the emerg­ing coun­tries, right, who are just excit­ed as can be about get­ting con­nect­ed in a very intel­lec­tu­al and emo­tion­al way to the rest of the world, and such a rapid way of doing it. I mean, just by buy­ing a few lit­tle modems and a PC or two or some­thing like that and some lit­tle piece of soft­ware, they’re right into the main­stream late 20^th cen­tu­ry tech­nol­o­gy. And they have the brains to do it. And it’s not very cost­ly to do it. And it’s just you know, that infec­tion when you see new­com­ers come into it and you see how they react, boom, you know this is the right thing. It’s got a long play.

Malamud: You’re right, it is an amaz­ing effect. In some ways the Internet is most use­ful the far­ther away from the core you get. If you’re at a nation­al lab­o­ra­to­ry it’s nice. If you’re iso­lat­ed some­place in the mid­dle of the desert, it may be essential.

Lynch: Yeah.

Malamud: I’ve noticed that a lot of the TCP/IP-based com­put­ers that’re sold end up in some lit­tle LAN island. They’re not on the Internet. And I guess that rais­es a cou­ple ques­tions. One is, are these islands gonna con­nect up to the Internet. And also, are we meet­ing the needs of these peo­ple that are not con­nect­ing to the net. Is the TCP/IP stan­dards process meet­ing the needs of a small cor­po­rate consumer?

Lynch: Well, the answer’s no, okay. And they’re iso­lat­ed for two rea­sons. You know, num­ber one it’s eas­i­er to stay iso­lat­ed because they don’t have to wor­ry— They hear well, I’m not a gov­ern­ment thing or what­ev­er, I don’t ha— The accept­able use pol­i­cy. I don’t know if I should be on the net or some­thing like that. And so they just stay off. Why bother.

But, if they want­ed to get on, and they almost all can—I mean I sort of believe that this whole AUP thing is sil­ly. I mean…you know. But if they want to get on, how do they find any­thing? And that’s the hard part, right? You know, you have this…the White ages, I mean. We’ve had these var­i­ous X.500 exper­i­ments and DNS things and… You know, I for­get what the exact tech­nol­o­gy is, okay. The idea of hav­ing this giant White Pages in the sky you know, where every­thing is con­nect­ed to every­thing and you can learn any­thing about any­body you want to and all that sort of jazz…it…it just…it does­n’t work. It does­n’t make sense from an eco­nom­ic standpoint…you know, from a social stand­point, from a legal stand­point. So, we have this prob­lem of when you get on the Internet you know, well what do ya get on to? Because there’s nobody in charge. There’s just peo­ple out there com­mu­ni­cat­ing who want to com­mu­ni­cate and who form sub­groups, or col­lec­tions of peo­ple, and they pass each oth­ers’ address­es around in names of files, and you know… It’s seduc­tive to think oh well that could all just be in one giant file, you know what I mean? One giant data­base in the sky. It’s all just data, right. And it’s all pretty—you know, there’s prob­a­bly only a thou­sand dif­fer­ent kinds of entries. Maybe it’s 200, maybe it’s 10,000—you know. But I mean there’s some finite num­ber of attrib­ut­es that you might look up things by. We’ll all need— You know, Steve Chen will nev­er go out of busi­ness, okay. We’ll all need Super Crays just to do a lookup of a name. 

So you know, think about the voice tele­phone sys­tem which is rather per­va­sive world­wide. There’s no sin­gle White Pages, right? I you wan­na get some­one’s name in Lyon, France…I mean, you know, you go to the expense of call­ing up the oper­a­tor in Lyon, okay. A spe­cial kind of direc­to­ry assis­tance thing, you know, fish­ing around, right. And if you don’t speak French well then you got­ta find a trans­la­tor, I mean you know, what­ev­er, in order to get some­one’s phone num­ber. So, that’s just not all done automag­i­cal­ly in the sky. And there’s just lots of good social rea­sons and busi­ness rea­sons for that.

Malamud: So how are we gonna get these peo­ple on the net, then? I’m assum­ing that’s a good thing to do, right, to con­nect these lit­tle islands up so we can talk to ​’em if we want to. Not say­ing we have to talk to ​’em.

Lynch: Yeah.

Malamud: What’s it gonna take to get—

Lynch: Well I think the first thing you got to get ​’em on is just email. Okay that’s the first part, right. That’s the ben­e­fit, Okay. And of course we all know that email is a…you know, you don’t have to be com­plete­ly and ful­ly con­nect­ed in order to have email con­nec­tiv­i­ty. It might take a few hours of polling and dial­ing and things like that as a sort of store-and-forward sys­tem. And it’s not a full live IP-connected way. 

And so the first way is to get them hooked on email. And that prob­lem’s basi­cal­ly solved. What I mean by that is you can buy from…you can have your lit­tle local mail sys­tem. And whether it’s a SMTP email sys­tem, or a Novell mail sys­tem, or Microsoft or what­ev­er that does all your local iso­lat­ed mail. And then buy an account at MCI, or buy an account at Spring, or buy an account at AT&T. Or CompuServe or you know, any of these ser­vices. And then you can plug all your mail through that and use those as for­ward­ing agents. That’s what I see hap­pen­ing, okay, is those com­mer­cial mail ser­vices are real­ly becom­ing the back­bone and mail for­warders. They’re becom­ing mail for­warders and not so much that peo­ple actu­al­ly use them to do the mail on. 

For instance I know a com­pa­ny that start­ed out an MCI Mail, and they had like a cou­ple thou­sand peo­ple using MCI Mail. And then they then they bought Lotus Notes, okay, which gave them a rich­er way of manip­u­lat­ing their infor­ma­tion. And they use Lotus Notes now as their mail sys­tem, but they still use MCI to actu­al­ly deliv­er the bits. But they no longer use the MCI front-end or any­thing like.

Malamud: So MCI in that case is turned into a tran­sit net.

Lynch: It’s turned into a tran­sit net, okay—

Malamud: They’re a part of the Internet backbone.

Lynch: They’re a part of the Internet back­bone as far as I’m con­cerned, absolute­ly right you know. I mean, I get mail from my head­quar­ter’s com­pa­ny through MCI. And they have NetWare— I mean it’s just this giant hodge podge…and it just kin­da works. And so that gets peo­ple infect­ed, gets them going of, you know— And you can sign up for MCI Mail in you know, Ethiopia. It does­n’t mat­ter. They’ll give you a num­ber to call. I’m not sure it’s absolute­ly 100% uni­ver­sal, but I’ll bet it’s very close. 

And then you get hooked. And then you say okay, well now I want some real real-time con­nec­tiv­i­ty. So I can send files around of arbi­trary size and all that sor­ta jazz, and lis­ten to Internet Talk Radio. Whatever it is you want. Then you’re moti­vat­ed to go the next step, okay, which is to go buy a router from some­body, or lease a router or lease a ser­vice. I mean it’s kin­da too bad— You know, I come from a tech­ni­cal world and I always think in terms— And I was a com­put­er guy, you know, so I always think in terms of get­ting hard­ware, right, buy­ing the soft­ware, con­fig­ur­ing it, and run­ning it and all that jazz, you know. Most peo­ple real­ly don’t want to both­er with that. They want to buy a ser­vice, right. I mean they just wan­na call a phone num­ber up and say, ​“Here’s what I want, and where do I bring some­thing over and plug it in. And send me the bill. And I’ll argue about the bill, but I just wan­na see a bill and some service.” 

And that’s the next step. That’s the next com­mer­cial step for the Internet, is to have ser­vice providers who just do that and do it for a rea­son­able price, but take the headache of oper­at­ing your net­work away from you. It is a headache. 

Malamud: This is Geek of the Week, fea­tur­ing inter­views with promi­nent mem­bers of the tech­ni­cal com­mu­ni­ty Geek of the Week is brought to you O’Reilly and Associates and by Sun Microsystems. 

This is Internet Talk Radio. You may copy these files and change the encod­ing for­mat but may not alter the con­tent or resell the programs. 

You can send us mail to mail@​radio.​com.

Internet Talk Radio, same-day ser­vice in a nanosec­ond world.

It sounds like you don’t par­tic­u­lar­ly care if it’s NetWare on this island and if it’s TCP/IP on that. It’s just—it’s a hodge podge that works.

Lynch: It’s a hodge podge that works. I mean, I used to have this grand hope that it was all gonna be one thing, you know? And at first I thought it was going to be XNS. The stuff that came out of Xerox—

Malamud: You’re show­ing your age.

Lynch: Yeah. Cuz I saw that stuff well, back at Xerox PARC when it was called PUP, PARC Universal Protocol back in the mid-70s. And it was pret­ty neat. And they were actu­al­ly doing some­thing with net­work­ing besides just send­ing files around. Which I call send­ing com­plet­ed units or work. That’s what a file is. That’s dead data, right. Okay, and some­one else wants to look at the dead data. That’s fine, there’s a lot of val­ue in that. But you know, what about doing things because there’s a net­work in place, that you could­n’t do with­out a net­work like you know, com­bine cal­en­dars as a sim­ple exam­ple, okay, or big sales meet­ings or any kin­da orga­ni­za­tion­al meet­ing in real time in many dis­parate locations. 

But… The Inter— See, where was I. I lost my train of thought. 

Malamud: Question being um…if there’s a hodge podge out there does it real­ly mat­ter if we’re using TCP/IP or…

Lynch: Okay. Yeah. So I thought there was going to be you know, one— First it was gonna be XNS, okay. Then Xerox did­n’t let go of every­thing and so that was­n’t going to be it. And then it looked like well, maybe it’s going to be TCP. And I saw TCP real­ly take off in the mid-80s, and it was solv­ing a lot of prob­lems. It was a bit bare­bones. And then ISO stepped up to the bar and said, ​“Oh no, TCP is just the train­ing wheels, okay. ISO, the real thing.” And I look at all the ISO def­i­n­i­tions and said gee whiz, yeah that is sor­ta TCP on steroids, right. It’s got almost every­thing you want, plus things—

Malamud: Plus more.

Lynch: Plus more, that you may or may not want, okay. But you know—

Malamud: Feature-rich.

Lynch: It’s feature-rich, okay. And so I thought well that’s going to be the great hope, okay, and the great sav­ior for all of us. And I watched it try, and essen­tial­ly stum­ble in a lot of places, and have some suc­cess but you know, not take over the uni­verse. And then along­side in the PC world is NetWare, okay, which I sor­ta laughed and said heh, that’s just a lit­tle print­ing pro­to­col, or a file-sharing pro­to­col and a print­ing pro­to­col. And well, son of a gun, 20 or 30 mil­lion com­put­ers lat­er, it’s sort of big, and impor­tant, and does a lot of great work. 

And so I sor­ta real­ized there’s not gonna be one win­ner. There even can’t be. And the rea­son there can’t be is the tech­nol­o­gy keeps mov­ing ahead so fast that no one cen­ter of excel­lence can encom­pass it all. And you know, we’ve got this whole new thing tak­ing off with mobile com­put­ing, right, and mobile data. And the peo­ple who are sweat­ing out NetWare, or sweat­ing out TCP, or what­ev­er it is just…are also pay­ing strong atten­tion to mobile data needs. And so the peo­ple who are pay­ing strong atten­tion to that are devel­op­ing their own slight­ly dif­fer­ent pro­to­cols, and slight­ly dif­fer­ent meth­ods for com­mu­ni­cat­ing. And when that stuff becomes use­ful, it’ll get incor­po­rat­ed. There’ll be yet anoth­er kludge box—hardware, soft­ware, whatever—that you can buy for some­body to hook togeth­er some mag­ic cap stuff, and some MIME stuff, and some UUNET, and on and on and on. So it’s just going to be this giant hodge podge in the sky. Perhaps forever.

Malamud: And that’s a very dif­fer­ent dream from what for exam­ple the OSI com­mu­ni­ty had orig­i­nal­ly intend­ed [crosstalk] which would be a sin­gle uni­ver­sal ulti­mate protocol.

Lynch: Yeah. Yeah. I don’t see it… I don’t see it hap­pen­ing. I don’t see the tech­ni­cal force for that hap­pen­ing. I don’t see the mar­ket force. I think if you get the thing down to half a dozen rea­son­able com­peti­tors at any one point in time, that’s enough. N squared where n is six is only thirty-six, you know. And that’s not too bad a price to pay for diver­si­ty. See, for hav­ing diver­si­ty. The oth­er side of that is, you get diver­si­ty. And we know from our knowl­edge of the world and the uni­verse that diver­si­ty is a good thing. You get new ideas. 

Malamud: Mutations are good.

Lynch: Mutations are good, you know. And some are bad. And the bad ones don’t get used, or don’t prop­a­gate for what­ev­er rea­son and go away, and the good ones get you know, jumped on. 

Malamud: As a mem­ber of the IAB, I’ve noticed that mem­bers of the Internet Architecture Board seem to have an appre­ci­a­tion for fine wines. Was at one of the selec­tion cri­te­ria for IAB members?

Lynch: Well, no no no no no no. I’ve always been interested—well not always. I mean, not as a cal­low youth. But when I moved to Northern California in the ear­ly 70s, I hap­pened to fall into a group of peo­ple at SRI who were the founders and own­ers of a lit­tle bou­tique win­ery called Ridge Wines, and they made real­ly good wines. And they used to have them at the par­ties on Friday after work and things like that, and at their homes. I went gee, I’d been used to drink­ing you know, Gallo Hearty Burgundy, right, which I thought was okay. Until I got intro­duced to some real­ly good wines. And so no, I’ve always ever since then been a wine col­lec­tor, con­nois­seur, call­ing it what you like. 

Malamud: You’ve got a bit of a collection?

Lynch: Oh sure. Oh sure, Carl, you know. You’ve been there. 

Malamud: [laughs] I’ve appre­ci­at­ed it. Definitely fine wines.

Lynch: It’s won­der­ful. It’s like net­work­ing. It some­thing you can share with friends.

Malamud: Well this has been Geek of the Week, and we’ve been talk­ing to Dan Lynch. Thanks a lot Dan.

Lynch: Good night, Carl.

Malamud: This has been Geek of the Week, brought to you by Sun Microsystems and by O’Reilly and Associates. To pur­chase an audio cas­sette or audio CD of this pro­gram, send elec­tron­ic mail to radio@​ora.​com. Internet Talk Radio, flame of the Internet.