Friday, January 15, 2021

memories from uw and cornell

This is something of a random blog posting. 

As the new year rolls around, I became thoughtful of the page of milestones. These include my time at the University of Washington here in Seattle https://www.cs.washington.edu/ getting my CS Masters during the 1997-1999 time frame. 

I spoke a bit about the UWCSE in http://vzimmer.blogspot.com/2018/06/ along with the now-closed https://www.livingcomputers.org/. Given the recent interest in retrocomputing https://www.nytimes.com/2021/01/08/style/retrocomputing.html the museum would be overrun with aficionados if it were open.

I frame many of my UW memories via the professors. These included John Zahorjan https://www.cs.washington.edu/people/faculty/zahorjan on computer performance. I recall one project with a classmate Amanda Barrett (then an employee at Teledescic, Macaw's attempt at a satellite communications network in the late 90's) on modeling different web server scheduling policies, such as Round Robin DNS (RR-DNS) using C++Sim. The most interesting part of the effort was the ability to drive the simulation with anonymized, real-life web traffic from Metacrawler by way of UW alum Brian Pinkerton alumni https://en.wikipedia.org/wiki/WebCrawler https://www.w3.org/Conferences/WWW4/Papers/169/.

The next professor I recall is Anna Karlin for algorithms https://www.cs.washington.edu/people/faculty/karlin. She taught my first class at UW. The take-away I have from that course was the value and extent of mathematical rigor behind algorithms. From my undergraduate and 5 years prior industry experience I saw algorithms more as rote recipes than evolve mathemtical objects.

Next up was the artificial intelligence course with Dan Weld https://www.cs.washington.edu/people/faculty/weld. Like the performance class above, my strongest impression was the project course. The specific project included writing a movie recommendation system for movies. We would create Java wrappers for websites, such as for the recently launched https://www.imdb.com/, to support queries written in Datalog https://link.springer.com/referenceworkentry/10.1007%2F978-0-387-39940-9_968. It would allow the end user to write queries, such as 'Show me all of the movies in Seattle starring Tom Hanks.' The downside of the system is that this work predated the semantic web and the website wrappers had to continually get updated based upon the changes in sites like IMDB. 

The other part I recall from the AI adventure is that my partner was a local Intel DuPont employee in another team. His manager was much more liberal about taking classes so he had the opportunity to work on the course during the work day. My management, who had initially replied to me when I requested funding for the masters project with 'why do you want to take classes, you are already smart enough?' didn't permit such liberties. So like my patent writing of the last 20 years, my masters work was always a late-night after-hours and predominately weekend activity.

From AI I recall taking a second algorithms class with Richard Ladner  https://www.cs.washington.edu/people/faculty/ladner. I still recall a quote from Ladner early in the quarter, namely "I cannot teach you everything about algorithms since the field is so broad and continually changing, but what I can do is teach you have to do research and learn on your own." The deep project work done in that class involving assessing recent publications has stayed with me. And the wisdom still holds true today, every field is continually changing. Sort of the academic analogy to the pre-Socratic Heraclitis quote "You cannot step into the same place in a river twice."

Another part of the UW journey was sorrow, too, including the passing of my advisor https://s3-us-west-2.amazonaws.com/www-cse-public/publications/msb/msb11.1.pdf

Just like my undergraduate journey, I didn't have the luxury to take so many courses, so I calculated the exact number of credits I needed to get my degree. For undergraduate urgency the timing was economic based, whereas for my masters it was lifestyle based (i.e., high pressure job with hardware power-ons, new-borne daughter, etc). As such, one way to complete the requirements was through research credits, and one effort involved working on a project with Susan Eggers  https://homes.cs.washington.edu/~eggers/. She was a huge influence on me in computer architecture, and after meeting her, some of the stories I late reach https://egc.yale.edu/how-job-yale-1960s-set-susan-eggers-groundbreaking-path-computer-science did not surprise me at all.

Since distance learning was a bit nascent in the late 90's, I still recall hurrying from DuPont, WA Intel site to the U District in Seattle. I tried to time my arrival such that when the street parking became free at 6pm I could find a slot.

And the old CS building Sieg Hall, ah......


Definitely quite a change from the new EE/CS building, especially the Amazon auditorium. I luckily managed to catch a couple of interesting talks there in the last couple of years, including Patterson preaching about RISC-V https://www.youtube.com/watch?v=mD-njD2QKN0 (and getting one of the single-page ISA descriptions printed on green paper) and David Bacon on the evolution of quantum computing https://phys.washington.edu/events/2019-12-10/cse-colloquium-quantum-computer-versus-supercomputer.

An interesting aspect of the university, versus industry, is that the rank of a professor is much more explicit, such as the laddering of associate versus assistant versus full versus emeritus professor, respectively. Compare this with the wild variability of job titles in the technology industry https://www.levels.fyi/, for example. A principal at one company versus a partner at another versus....although https://news.ycombinator.com/item?id=25758663 recently had an interesting discussion thread on the topic.

Speaking of partner title, I still recall one MS partner architect telling me that partner is the 'throw the other guy under the bus' job band. This spoke to the highly competitive nature of the job category since like many domains the higher you advance leads to few openings and broader expectations with the competition to advance and challenge to sustain the level being commensurately intense. 

Even in my early late teens I saw a glimpse into this. At Cornell I recall a fellow undergraduate who had interned at Goldman. He related tales of other interns sniping at each other openly during presentations each was required to make. The same calculus applies - exclusive job category with high compensation, thus some 'throw them under the bus' behaviors.

Speaking of Cornell working on my BS EE https://www.ece.cornell.edu/ece during the 1988-1991 window, a few memories come to mind. The first includes the luminaries who visited campus, such as Normal Mailer and Roger Penrose to give talks. Penrose, Kip Thorne, and other physicists were drawn to a department that still had Hans Bethe https://en.wikipedia.org/wiki/Hans_Bethe. This was the same physics department that hosted the Feynman lectures. And for literature Cornell was the abode of the likes of Vladimir Nabakov, too.

Just like UW the professors made some of the largest impacts. These include Dynkin for real analysis https://news.cornell.edu/stories/2014/11/mathematician-eugene-dynkin-dies-90, Terrence Fine on probability https://www.engineering.cornell.edu/faculty-directory/terrence-fine, and the father of information retrieval Gerard Salton https://en.wikipedia.org/wiki/Gerard_Salton for discrete math. Cornell didn't just send TA's to instruct undergraduates but instead offered access to world class researchers in their fields.

Moving from maths to engineering, Prof Torng https://news.cornell.edu/stories/1997/12/cornell-professor-honored-intel-corp-his-computer-chip-improvement taught the course on digital design. K-maps, Mealy-Moore machines, etc. Good stuff. 

And closest to home there was Professor Gries https://www.cs.cornell.edu/home/gries/gries.html class on computability theory. This started with computational complexity, big-O notation, and other rudiments, leading into Hoare triples, such as pre-conditions, post-conditions, and invariants.  I still remember walking down the hall of the computer science building and seeing a room filled wall-to-wall with silver-covered Springer-Verlag "Texts and monographs in computer science."






Beyond visiting speakers and professors I recall lots of lake-effect snows and winds during the winter, and beautiful changes of seasons and the rugged environment of Ithaca, New York. There was a common theme on T-shirts and car stickers of 'Ithaca is gorges' as a play on the terrain and 'gorgeous' scenery up upstate New York.


Since I was so far from home and living on a budget, I would only return back to Houston for events like Christmas or summer. I recall one thanksgiving holidays eating at gas station since I had forgotten to do grocery shopping ahead of time. Other times I would travel to visit my aunt in New Jersey by way of New York Port Authority. Port Authority was definitely a stark contrast from bucolic Ithaca, NY or suburban Houston. I recall one trek through Port Authority when there was a body in a wheel chair parked outside of the bathroom. A sheet covered the bulk of the person but the hands were protruding with what looked like the stiffness of rigor. The crowd in the bus terminal where walking around the body as if it were just another obstruction on the path of their daily routine. Another trick I learned going through Port Authority with my suitcase was to put my money in my shoe. I'd leave $5 in my wallet so that when someone invariably wanted to 'help me' find my connecting bus terminal and then requested a gratuity because "you know I could be doing much worse to make money", I could satisfy the request and not get rolled for all of my traveling money.

When I think of growing up in Houston, or the US Southwest, doing undergrad in Ithaca in the US Northeast, and now leaving in the US Pacific Northwest, I'm mostly covered the continental US. Heat in the SW, cold in the NE, and rain in the NW. I'm only missing living in the US Southeast, such as Florida. But after reading Thomas McGuane's 'Ninety-two in the Shade' https://www.goodreads.com/book/show/51598.Ninety_two_in_the_Shade in my youth, I wasn't so anxious to move to that area.

No comments: