Douglas Comer: Well there are lots of peo­ple I would like to thank. I thank the advi­so­ry board for induct­ing me. I thank, way back, Peter Denning and Larry Landweber for includ­ing me on the CSNET project that got it all start­ed for me. 

They made me in charge of pro­to­cols at a time when I knew very lit­tle about pro­to­cols. I talked to Vint Cerf about what to do and we need­ed to build a net­work to get across all com­put­er sci­ence depart­ments and we did­n’t have enough mon­ey to afford a new back­bone. So, we decid­ed to use X.25, an exist­ing pub­lic ser­vice that we could use to trans­port data from one site to the other. 

And I explained to Vint, I did­n’t know how to do pro­to­col trans­la­tion, which was the way every­body did it those days. And so I decid­ed to use some­thing that became known as tun­nel­ing. I thought I was pret­ty well set. They approved it. Then we sent in the CSNET pro­pos­al, and the National Science Foundation had anony­mous review­ers review the pro­pos­al. And they came back with very neg­a­tive reviews. 

Now some of them did­n’t both­er me at all. They said for exam­ple this is a multi-institutional grant with mul­ti­ple insti­tu­tions run­ning bud­gets, we’re gonna have to do acqui­si­tions, and nobody on this project—it’s just a bunch of professors—nobody has had any real pro­gram man­age­ment expe­ri­ence. There ough­ta to be a real pro­gram manager.

That did­n’t both­er me because I real­ly did­n’t know about bud­gets and how to acquire equip­ment and lines and…that was okay. But, they also sin­gled out my part. They said, point blank, that idea about putting IP pack­ets over X.25 will not work. You can’t do it that way. 

Now, I was an assis­tant pro­fes­sor at the time. I was gonna come up for tenure. As you can imag­ine, I did­n’t want to have a black mark on my record. I had noth­ing to back me up except a hunch that it would work. So, I dug in. I said to myself I bet­ter find out about these pro­to­cols. What do these experts know that I don’t? 

I worked and worked and worked. I stud­ied all the protocols—not just indi­vid­u­als but I study how they inter­act­ed togeth­er. I thought about rout­ing, I thought about address­ing, I thought about cor­ner cas­es. I thought about maybe we should do an imple­men­ta­tion, so I imple­ment­ed the pro­to­cols, put em across my tun­nel… Well a few years lat­er, it all turned out to worked just fine. 

But, that study had a pro­found impact on my career and on my books. When it came time to write the books, I already knew first-hand all about the pro­to­cols. I had stud­ied them so intent­ly that I under­stood cor­ner cas­es, I under­stood how to imple­ment them, I under­stood what not to do. So I boiled down all the prin­ci­ples, all the details, and all my find­ings about pro­to­cols and put em into books. And I have always thought that that’s one of the key rea­sons that my books were so wild­ly suc­cess­ful. Because they weren’t just some pro­fes­sor writ­ing a text­book, they were from first-hand experience. 

So, my final thanks for intim­i­dat­ing me into study­ing pro­to­cols goes to those obnox­ious and absolute­ly incom­pe­tent NSF review­ers who forced me to study.

Further Reference

Internet Hall of Fame pro­file