Stuff Yaron Finds Interesting

Technology, Politics, Food, Finance, etc.

Host Routing Multicast Engine (HRME)

I liked university. It gave me time to think about things. It gave me interesting people to talk to. But then I had an idea. It was called the Host Routing Multicast Engine (HRME). The idea was to get computers above the TCP/IP layer (e.g. not IP based multicast, see here for the problems with IP multicast) to join up in a spanning tree and distribute information down the tree. By each machine volunteering to redistribute content to every other machine one could very efficiently distribute large amounts of data. I'm sure this all sounds very familiar but these thoughts occurred to me back in 1994. 56Kb modems were state of the art, e-mail was still mostly 7 bit text, the Internet bubble was just getting started and I smelled money in them 'dar' hills.

I wrote a paper about my idea for a computer science class [1]. Reading the paper now is interesting if only to see what assumptions I had 10 years ago and how they turned out. I already seemed to have a hang up on standards, I promoted the idea of using shareware to distribute the program and I expected to distribute digitally signed source code. Of course I also expected per packet pricing charges (although I used this to motivate why an efficient distribution algorithm was needed) and was sure that the Internet community would get very cross at 'inefficient' commercial usage (another motivator for the algorithm).

Of course I didn't mention HRME directly in the paper since it was a 'secret'. The reason it was a secret is that I had started a company (That at various times was called Modulus and CoreTech) with my college friend Adam Zell. The company's purpose was to commercialize the algorithm. Padgett Peterson, someone I knew from my anti-virus days, introduced me to Robert Cringely who at the time was working for the International Data Group (IDG). Cringely understood the potential for the Internet and wanted to find a way to let IDG distribute its content over it. Web Pages were not yet seen as an optimal way to do this because the web page technology was very primitive and publishers as well as advertisers are neurotic about controlling exactly how their material appears to readers. What Cringely wanted was a way to distribute high quality multi-media files with lots of pretty graphics with a fidelity far beyond what web pages could handle. But given the size of the data and the speed of the average Internet connection at the time this didn't look like a solvable problem.

When Padgett told me what Cringely was looking for I figured I had hit gold. Adam and I met up with Cringely and we explained the HRME algorithm to him. Cringely really liked what he heard and agreed to pay us to write up a feasibility report that would examine how well HRME could be used to meet IDG's needs. Everything was going well until we delivered a status report giving more detail on the HRME technology (here is an excerpt describing the HRME algorithm). The paper was marked property of CoreTech. It appeared that IDG thought that it had purchased the HRME technology. Of course, they hadn't, they had only paid (a tiny) amount of money to pay us for a feasibility report.

At this point we had to fly out to San Jose (we were based in LA) to meet with IDG's lawyers, Wilson Sonsini Goodrich & Rosati (WSGR). I still remember their offices, they had their own building with a two story foyer and enough marble and wood to make me think I was in a cave in a forest. We were then taken to a suitably impressive conference room with a suitably impressive lawyer with a suitably impressive title. What happened next was the interesting part.

I don't claim to know much about the legal profession but near as I can tell these folks have a fixation on 'pedigree' where one's pedigree is determined by which law school one went to and which law firm one is a member of. Go to the wrong school or firm and you are trash. This certainly was the way the WSGR lawyer treated our lawyer, Jonathan Barsade. Jonathan is a wicked smart lawyer but he made the unforgivable error of going to an Israeli law school, passing the bar in Israel, practicing law in Israel, coming to America and in less than a year and a half passing both the California and New York bars (both on the first try) and then starting his own law firm. But none of this mattered, he had the wrong pedigree so the WSGR lawyer figured he had a hick from the sticks on his hands.

So when the meeting began the WSGR lawyer very calmly explained to us in extremely polite language how we were screwed and should take the pittance of money IDG was willing to throw at us (in the form of jobs, not even cash) in return for our technology. What the WSGR lawyer didn't understand is that Jonathan, not IDG's in house legal staff, had written our contract with IDG to do the research report. Apparently the WSGR lawyer had never actually seen the contract. So we showed it to him.

The guy excused himself for a few minutes to go make a "photocopy" of the contract for his own records. Ten or Fifteen minutes later he came back with a crestfallen look on his face. The meeting was over. We were shown out. Don't mess with Jonathan.

Later on another, more junior lawyer, from WSGR was sent to negotiate with us. We made the same offer to this new lawyer as we had made to the previous lawyer, in return for essentially no money we would give IDG a non-exclusive, perpetual, royalty free license to the technology. We figured it was worth taking a bath on this deal in order to establish the technology and make money on other customers. But for IDG it was total ownership or nothing, so we both ended up with nothing.

After that we also had a conversation with Pointcast (remember them?) but that went nowhere. By this point I had graduated college and I figured it was time to stop mooching off my parents. I saw two options – get a job or go to graduate school. I had always wanted to go to graduate school to study AI but while I was mulling over my options I had sent in an ASCII resume with no cover letter to a resume e-mail drop at Microsoft. The joke being that I had no interest in working for the Evil Empire but my mother insisted I at least submit a resume and I felt childish saying no. To make a long story short during my meetings with Microsoft I met some of the smartest people I'd ever run into and realized I'd be an idiot not to go work there. The rest, as they say, is history.

[1] I think I might be going senile. The paper I link to above claims to be for Professor Karplus's CS 171 class. But CS 171 was a systems and signals class, it was about things like fiber optics. The paper doesn't seem appropriate for that class. I do remember that the professor liked me and after taking CS 171 and CS 171L he asked me to work with him on some of his research but he was working on fiber optics and that wasn't my area of interest. So perhaps he let me turn the paper in for credit as a favor? It's embarrassing but I just don't remember. I would ask the professor but I found out, with great sadness, that he died of cancer in 2001.

Leave a Reply

Your email address will not be published. Required fields are marked *