Diary (This Diary has been replace by this)

My time in Cambridge as a Visiting Scientist at MIT/CSAIL.

Sunday, 30-AUG 2009

Arrived with Ronee at Boston Logan.  Unpacked.  Went out and had dinner at the Marriot "Champions" restaurant.  She was disappointed that only the highlights of the Red Sox game were on TV at the restaurant, which is a sports bar.  After dinner took a walk and bought some stuff for breakfast at the 7-Eleven nearby.  Unpacked.  Apartment nice, but telephone doesn't work, a TV doesn't work, and ice machine doesn't  to work.  Oh well, it's a new apartment.

Monday, 31-AUG 2009

Took a walk to get my ZipCard at 25 First Street.  Cool.  Had lunch at the Cambridgeside Galleria, then walked back to Third and picked up the car and drove to Shaw's market.  Spent a bunch of money stocking up on groceries and still missed some things.  Returned the car on time, unpacked the groceries, and went for dinner at Vicky and Dick's house.  Jessica joined us.

Tuesday, 01-SEP 2009

Went to Orientation at MIT.  Met my office-mate, Masatoshi Yoshida ("Masa"), an engineer from Hitachi.  He is here for a year with his wife.  Filled out some paperwork with the CSAIL HR folks (John Merriman, Rachel Avery), got my office key, and then Masa and I went to get our ID cards.  Came back and started fixing up the office.  Spent a lot of time downloading certificates, getting mail set up, web site set up, etc.  This continued over the next few days.  Finished the latest Eric Van Lustbader book: First Daughter.

That evening, had dinner with Ed and Arlyne Westerman at the East Coast Grill near Inman Sq.  Very good.  It was good to see them.

Wednesday, 02-SEP 2009

Did a bunch of email in the morning.  It's taking me about 2 hours each morning to get out of the apartment, most of the time being keeping up with emails.  Went to the office in the afternoon.  Had a problem with the SpashID upgrade.  Went back to the apartment, and had an ABI Board affairs committee call.

That evening, had dinner with Bob Redd, Marie Enochty, and Shayna Redd.  We went to Legal Seafoods and had the fried clams.  Excellent.  Will we get Red Sox tickets as Shayna suggests?

Thursday, 03-SEP 2009

Went to the office.  Pretty much completed the environmental computing setup, but needed a little help from TIG.  Met a few members of the staff, and also bumped into Tom Knight, who I haven't seen in years.

That evening, had dinner with Rich Stroum and his new lady friend, Linda (Wantman) Levine.  Dinner was at L'Osteria, 104 Salem Street, North End.  The Eggplant/Veal Parmigiana was quite good.  Definitely will go back.

Went back and did some more email.

Friday, 04-SEP 2009

Again: email in the morning.  Stopped by Au Bon Pain on the way to the office.  They have French Roast Decaf (my favorite).  Also picked up a sandwich for lunch.  Attended a talk StarCluster - High Performance Computing on Amazon's Elastic Compute Cloud by Speaker: JUSTIN RILEY, MIT.  He described a set of tools to manager virtual clusters on AWS (Amazon Web Services).  Mostly stuff I already knew.  Saw Peter Mager, who I haven't seen in years, there but didn't get a chance to talk with him.  After the talk, I bumped into an old colleague, Randy Rettberg (BBN Butterfly, Apple).  He's working with Tom Knight.

Came back downloaded some new software, a little more email. and attended the CSAIL annual ice cream social at 3:30.  Saw and talked to Victor Zue, Lab Director, Jack Dennis, Dorothy Curtis, and Randy Davis, all of whom I know.  (Dorothy is married to Craig Schaffert, Barbara Liskov's first Ph.D. student; Craig worked for me at DEC.)

Dinner tonight at the SummerShack right next to the Alewife T station.  Excellent stuffed lobster.  Ronee had a boiled lobster.  I'll definitely be back to this one.  We've only had the best of luck with the restaurants this week.

Fixed SplashID finally, after getting a new registration code from the company.  I still don't know what went wrong.

So, I was supposed to be on vacation all week, but spent most of it in the office at MIT.  First week was energizing.

Saturday, 05-SEP 2009

Ronee packed, we cleaned up the apartment, then took a walk over the bridge into Boston.  A friendly guy stopped us, assuming we were tourists, and gave us some friendly advice:  (a) Buy a weekly unlimited pass to the T for $7 (includes the water shuttle); (b) if you need a place to stay, rent a room at the Jeffrey House at the foot of the Longfellow bridge, Boston side - great rate; (c) if you drive into town, park at the Alewife T station for $7/day for ($49 for the week), and take the T (even if you stay more than a week, you can pick up the car, drive it out, drive it back and you're golden) - works well with the weekly T pass above.

Stopped at Fanuel Hall for lunch (at the Cheers knockoff tourist trap).  Ronee bought a couple of gifts, and we walked over to downtown crossing and took the T back.  Put up a wash, and then she sacked out.  Cab picked her up to go to the airport ( 866-45-Coach).  What will I do tonight?

So, I decided to go out for a pizza or a sub at a neighborhood place on Third.  It closed at 5 pm, and it was now about 6:30.  So, I walked up to Cambridge Street and checked out a few places, and then double backed to the Cambridgeside Galleria, where I had sushi at the FoodCourt.  Walked back tot he apartment and after introducing myself to a new TV show ("Defying Gravity") called it a night.

Sunday, 06-SEP 2009

I rented a Zipcar for 9:30.  Picked it up at 9:15, and drove around Cambridge to get my bearings. Stopped at Trader Joes and picked up a few things (some of which I shouldn't have).  Drove back to the apartment in a roundabout way to explore, dropped off the groceries, and headed to Target to pick up the things I couldn't get at TJs.  I still had some time, so I went off to the MicroCenter next to TJs.  I'm thinking of getting a printer for the apartment, but in the end decided I didn't need it thanks to MIT and to Evernote.  Dropped the car back and had lunch at the apartment.  Talked to cousin Lenny who's been ill - he sounded great.

I wish I had thought of the ZipCar idea first.

Monday, 07-SEP 2009

I spent the entire day in the apartment.  Watched a new episode of TV show I've been following on Hulu: Defying Gravity.  Really like it.  Then I decided to spend a little time on my music collection.  Managing the metadata is a lot of work.  I probably backed the collection up at home in a dumb way: I merely copied the music, leaving all the metadata behind in California.  Fixing this up takes a lot of time.  I did most of it, but still have about 50 CD's to go.  Lunch was some noodle soup, dinner was a big hot dog and some peas.  First meal cooked at the apartment.

Tuesday, 08-SEP 2009

My Office
Up early.  In the office by 9 am, having done SAP email at home.  Made myself lunch (tuna) to take with me into the office.  Caught up on some CHM Board action items. Dinner with Prof. John Williams and Abel at Amelia's Trattoria on Harvard St. in Cambridge.  I've a perfect batting average with restaurants so far.  John will introduce me to the Engineering Systems faculty working group;  I'll sit in on some of there seminars.

Phone in the apartment now seems to be fixed (but we've seen this before).

Wednesday, 09-SEP 2009

It was a very busy day, and I didn't accomplish as much as I hoped.  I really don't want to waste a minute of this opportunity.

I had to interview a couple of candidates for positions in the SAP Incubator.  To prepare, I wanted to review the resumes and print them out.  That's where the troubles began.  First, I found out the Microsoft Entourage didn't let me view, copy, or print files attached to Exchange calendar entries (where Katrin had dutifully put them, as is our normal method).  Hmm... Ok, plan B: go onto Citrix WTS, and access through Outlook.  This I did.  But to do this, I have to VPN into SAP, meaning I get off the MIT network.  Can't print from WTS, so I emailed the docs to my account at SAP.  Disconnect from SAP to check email at MIT.  But the mail must have been backlogged, because it wasn't arriving.  Plan C: fire up VMWare.  That didn't work either for some reason I don't remember.  All of this is time consuming, and I still needed to get over to the SAP office.  Well, I did, but I had to get Katrin to email the resumes to someone at SAP Cambridge and print them out and hand them to me when I arrived.  Of course, my badge doesn't work at SAP Cambridge -- don't know why.

Very frustrated with SAP IT.  It needs a major overhaul, especially compared to the ease of use of the MIT network.  Sent an introductory email to our new CIO suggesting my help.  How to lose a lot of time due to the perceived need for tight security at MIT.  So you have to resort to very insecure workarounds.  Duhhh....

Interviewed the candidates, spent some time with Wayne Haubner, head of the lab, and did some email, talked to Katrin, talked to Ronee, checked up on my sister, etc.  Went back to my MIT office, but by now it was late in the afternoon, and I hadn't started on my list of things to do.  Since I have to drive to NY tomorrow, and have a busy morning scheduled, I had to review my travel plans, update my maps, etc.  Then, finally, I could finish the Sussman paper.  But wait!  Apple released new iPhone firmware, and a new version of iTunes.  So, I wanted to do that on the fast network at MIT, not the slower one at the apartment.  That took a long time, and in fact, I aborted the backup which was taking too long.

Dorothy Curtis stopped by.  She's been very helpful.  She suggested I review all the research groups on the web site.  I'll start that tomorrow.

So now, it's about 7:30, I'm still in the office, restaurants closed, and it's chilly, and I wore a short sleeve shirt and no jacket.  So, I didn't finish the paper, went back to the apartment, and heated up the leftovers from last night, and made some more food, and finished the paper.  I don't think after reading it, that it's an important paper.  It potentially has application in parallel processing, but I think the key ideas could be implemented in other more conventional ways.  Still, there are some interesting performance characteristics exposed in the Scheme code.  At one point, early in the paper, Sussman seems to be rediscovering the value of assignment statements in non-Lisp-like languages.  But maybe that's unfair.

By the way, the paper is "The (Abridged) Art of the Propagator", by Radul and Sussman.

Thursday, 10-SEP 2009

Had a call with Lisa Reeves, whose VC fund is raising a new round, and is using me as a reference.

Went to the Faculty/Staff bi-weekly lunch at CSAIL, and was introduced by both Victor Zue and Anant Agarwal.  Several people want to meet with me as a follow up, and I'm setting up meetings. Had discussions on multicore software with Anant and his colleagues.  Also had a conversation with K. Krasnow Waterman and Gerry Sussman on Sussman's paper on propagators.  K wants to follow up on some of her research on "Roles", and is very interested in SAP's approach.  Due to travel we've agreed to meet week after next.  Randy Rettberg (BBN Butterfly and Apple ATG) wants me to buy him coffee and in turn he'll tell me why I should get involved in biomedical engineering.

Then I had to hustle over to Harvard Sq. to pick up my rental car for the drive to the IEEE Computer Society Industry Advisory Board meeting at TJ Watson Research Center in Yorktown Heights NY.  Stopped by Peets at Harvard Square for coffee.  Ahh... Peets.

Had a conversation with Wes and the marketing folks about the "innovation" parts of Léo Apotheker's corporate presentation.  While on the phone and not paying attention to my driving, I think I may have neglected to pay the iPass toll leaving MA and entering CT.  I wonder if they'll come after me.

The drive to NY was mostly uneventful.  Found the hotel, checked in, but then had difficulty finding the restaurant in White Plains.  Managed to misread a parking sign and got an expired meter parking ticket which I'll pay on Monday.  Finding my way back to the hotel was a nightmare.  No GPS and no navigator on dark roads can be a challenge.

Friday, 11-SEP 2009

Drove 30 minutes to IBM.  Short moment of silence commemorating 9-11.  Nice touch.  That day is still with me, and listening to the news that morning was emotional for me, even eight years later.

The meeting was well run.  We "retired" the first project (the one I was chair of with the able help of Shel Finkelstein) and declared victory.  Of course there is still ongoing responsibility to help Sumi Halal find writers for the IEEE CS "View from Industry" column.  I'll contact him today since the deadline is so close for the next issue.

That evening I drove to Richard and Lydia Sack's house in New Milford. Michael Rothman met us for dinner.  Spent the night at Rico's and Lydia's house.  Michael, Richard, and Lydia are old college chums, and it was like 40 years vanished in a flash.  I hadn't seen Michael in 18 years, but it didn't matter. 

I think the thing with old true friends is that you can simply relax and sink into your most comfortable persona on which you built the friendship in the first place.

Richard and Lydia Sacks, Michael Rothman, and Ike
 Michael's comment seeing Ike at Richard's computer

Saturday, 12-SEP 2009

Drove three hours back to Cambridge.  Stopped in Cambridge to do some grocery shopping before returning the car to Harvard Square.  The Hertz office closes at 4 pm on Saturday, but they had an "Express Return" box.  Hopefully that works out well.  Made some dinner back at the apartment, read, did some emails, and called it a day.

Sunday, 13-SEP 2009

Laundry, email, reading.  Cleaned up my music collection, now completed.  Managed hockey tickets.  Worked on the social calendar.  Going to the Red Sox game on Wednesday, Vicky and Dick's and Suzy and David's for Rosh Hashana.  Didn't get to work on the Firetide financing however.  Will review the docs on Monday.  Reading "Netherland" by Joseph O'Neil, recommended in an interview by Barak Obama.

Monday, 14-SEP 2009

It feels like a wasted day when it came to research.  I was up early, did my SAP email before breakfast, and was in the office by 9.  I didn't leave the office until 7:30 pm, but didn't get to do any research today.  Busy with SAP email, CHM board work, Fellows, some personal email, and a 90 minute lunch with an old friend, Ron Creamer.  Ron and I built a disk drive business for the Macintosh in the 1984-1986 time frame.  He has a company called Pageworks, which was a desktop publishing company that go into printing, and then real estate.  We grabbed a fast food lunch, and headed over to the Charles River Yacht club and ate our lunch at some picnic tables overlooking the Charles River.  Delightful.  I won't go into the CHM Nominating Committee deliberations, but it took most of an hour. 

I came back to the apartment by 8 pm, turned on Monday night football, and watched an exciting come from behind victory of the New England Patriots over the Buffalo Bills.  Talked to Alex several times during the game since he was watching also.

Early tomorrow morning I have to work on the Firetide financing docs.  Should be straightforward.

Tuesday, 15-SEP 2009

Got a lot more done today.  Reviewed the Firetide docs and got them signed and faxed.  Then I went through each major project on the CSAIL site, and sent out emails to the ones that looked most relevant to me and to SAP.  Unfortunately, after waiting 30 minutes on hold for an HR meeting at SAP to start, I gave up. Got on the Data Base Group mailing list.  Went to an interesting talk entitled "Algorithms Meet Art, Puzzles, and Magic" by MIT's Erik Demaine, concerning the fusing of mathematics, art, sculpture, and magic (as in magic tricks).  Also had a concall with Sumi Helal of U. of Florida.  Sumi is an editor of IEEE Computer Magazine; we discussed various topics to the magazine's "View from Industry" column and I gave him several suggestions.  Very tired last night, but had a bunch of personal stuff to do at the apartment.

Wednesday, 16-SEP 2009

Met with Julius Akinyemi, visiting researcher at the MIT Media Lab, and former SAP customer at Pepsi.  Introduced me to Andy Lipmann, the director of operations at the Media Lab.  We tried to get into see Frank Moss, but he was busy.  Julius is trying to put together funding for a group involved in new commercial technology for emerging economies.  I introduced him to John Williams work at MIT that John did jointly with SAP on water management. 

Also went to a talk by Prof. Harold Thimbleby of Swansea U. entitled "Killer applications? Simple but widespread programming problems that are usually ignored".  Interesting.  In this talk he calls for programmers to be far more diligent about software and gave examples of poorly designed software embedded in medical devices.  These kinds of errors can kill people.  Continued to set up faculty meetings.

I realize this diary should really be a blog, and in reverse chronological order.  When I have some time...

Had lunch with Andy Lippman of the Media lab. Andy and I discussed a project proposal I helped to introduce around redefining sports.  Joining us was Joe Paradiso, also of the media lab; he works on immersive environments, smart items and such. 

Met with Daniel Jackson, PI of the Software Design Group at CSAIL.  We talked about Alloy, "Alloy is a lightweight modeling language for software design. It is amenable to a fully automatic analysis, using the Alloy Analyzer, and provides a visualizer for making sense of solutions and counterexamples it finds."  The goal is to find flaws as early as possible with a set of automated tools.  In particular, we discussed the possibility of modeling transactions, and other ordering relationships.  He gave me a couple of papers to read on Alloy and making software development a more robust discipline.

Left the office earlier than normal.  Will do my reading at  the apartment.

Went to the Red Sox game with Marie and some of her friends.  Shayna and Bob backed off at the last minute.  I'm worried about Bob's health.  But the game was great, albeit cold.  I left after 6 innings, and returned the apartment to watch a bottom of the ninth exciting finish.  Could barely stay awake.  Red Sox won!

Thursday, 17-SEP 2009

Reading up on the Scala programming language.  Scala's goal in life is to be a better programming language for components, which is very very admirable.  My reading started off well, but it possibly includes too many advanced concepts for wide acceptance.  But, they appear to do a nice job on generics, and I like the syntax better than Java, I think.  The real test is to write some code.  But you'll only find out about how scalable Scala (pun intended?) is when you write large systems.  I read "An overview of the Scala Programming Languages (second edition)" and "A Scala Tutorial for Java programmers".

Met with Anant Agarwal, who invited me to his group meetings on Tuesdays, 3 pm.  He went over Tilera, and his new effort for DARPA around Angstrom and FOS.  I'll read some of the FOS papers.

Started to explore the MIT Library.  Downloaded the 25th Anniversary edition of CACM, which has many important early papers, such as Codd's on RDB, and Denns & Van Horn on capabilities.  I'd like to reread them.

Friday, 18-SEP 2009


Met with Steve Ward and had a wide ranging discussion on respective views on Cloud computing, over which we're not too far apart. He calls it "Fog Computing" which brings the cloud down to earth.  I talked about my ideas on synchronization, and existing cloud apps like Evernote and Twitter.  He also described his ideas on goal oriented computing, and sent me a paper to read.  I sent him some readings on Scala to get his views.  He will put me in touch with his student Justin who wants to work in the area of cloud computing.

Invited Renee James to the Fellows dinner as a guest of SAP, and reached out to Pat Gelsinger to come to MIT for  visit in his new role at EMC.

CHM got a $1M donation form Max Pavlevski, founder of SDC.

Made an appointment with Saman Amarasinghe.  Discussed the idea of using graphics chips (raised by Dean Jacobs) with Sam Madden who pointed me to, among others, Saman.

Turns out that Jack Dennis lives in the building next to mine.  I invited him over next week for a glass of wine.

Went to Doug Crockford's interesting talk about JavaScript which is simultaneously the most popular programming language on the planet and the least popular programming language on the planet.  ECMAScript 3.1 spec should be released by the end of the year.  I'll have to look into this also.

Met with Hari Balakrisnan about some of his wireless ideas, and his enterprise software ideas - near real time data warehousing and retrieval.  Hari co-founded Streambase with Stan Zdonek and Mike Stonebreaker. Need to get back to him in a few weeks.

Tonight starts Rosh Hashana.

Saturday, 19-SEP 2009

Went to synagogue with Vicky and Dick, then to holiday dinner at the Schullers.  Great fun.

Sunday, 20-SEP 2009

Rented a Zipcar, went shopping.  Cleaned the apartment, did laundry, cleaned up after the suds overflowed the washing machine.  Made dinner, sliced my thumb.  Stopped the bleeding.  Read the NY Times.  Had better days.

Monday, 21-SEP 2009

Attended day 1 of the 2 day MIT Sloan School CIO Summit.  Mostly worked in Enterprise 2.0 issues with Andy McAfee and Erik Brynjolfsson.  Met a bunch of good folks (Salesforce, State Street Bank, Commonwealth of Massachusetts, etc.) as well as a bunch of folks from the Sloan School.  CEO Léo Apotheker joined us in the afternoon, and gave a good speech after dinner.  Met Mathieu, his son, and a student at Sloan.  Got a copy of Erik's new book (and an MIT coffee mug).  Frankly, I would rather be reading the papers I wanted to read as part of my CSAIL research.

Tuesday, 22-SEP 2009

Day 2 of the MIT Sloan School CIO Summit was much better than Day 1,  IMHO.  Good talk by Erik B.  Good talk by Sandy Pentland.  I need to get together with Andy McAfee and with Sandy offline.  David Verrill wants me to come over and meet with Erik and him.  I think it would have been more appropriate to invite our CIO rather than me now that we have new one.  But I haven't met him yet.  I'll discuss with Erik.  Stopped off to run a couple of domestic errands on the way back from the Le Meridien Hotel where the conference was held.

Read "The new, Faster Face of Innovation" by Brynjolfsson and Schrage, and "How Social Netwokrs Network Best" by Alex Pentland.

Talked to Murray Spork about SAP internal transfer.

Read "The Case for a Factored Operating System (fos)" by Wentzlaff and Agarwal.

Wednesday, 23-SEP 2009

Started the day clearing out my SAP work, as usual.  Read the second paper on fos, and composed an email to Wentzlaff adn Agarwal.  Met with Randy Rettberg who is doing biomedical engineering.  Fascinating stuff.  He also went through his reasons for leaving the Sun Storage group as CTO, to come to MIT and do something more significant with his time.  Met with Charles Leiserson on two topics: Cilk and "cache oblivious algorithms".  The Cilk stuff was more interesting than when he presented it to SAP commercially.  Cilk was recently sold to Intel.  Did a bunch of SAP email at the end of the day after talking to Katrin.  Still hard to get on the SAP VPN sometimes, and it's a real PITA.  Stopped off at Dominos for dinner - didn't want to cook tonight.  I invited Jack Dennis (generally credited as inventing "capabilities") over a few days ago, and he came over tonight to share a bottle of wine and talk about capabilities.  Even though he's "emeritus" and theoretically retired, he's still doing work and has an NSF grant to work on "tree memories" (which are capability based).

Tomorrow looks like an extremely busy day.  I'll start off at SAP/Cambridge to interview someone for the GBI.

Thursday, 24-SEP 2009

Spent the morning at SAP/Cambridge going over the details of a new GBI project and interviewing the lead.

Went to the faculty/staff lunch and had lunch with the VP of Engineering, Mike Schroepfer, who was here to give a talk on Scalability of Facebook, which I attended.  They have a lot of different technologies that go into the site, distributed in several data centers around the world.  They have modified Memcache significantly (UDP, etc).  Seeing a general  trend, they use MySQL as a big disk, PHP for their web server (basically) and Memcache as their caching infrastructure.  It's an interesting problem, when you think about it.  The FB hompepage for each user has got to be built essentially from scratch.  There isn't a lot of locality ("MIT users have friends outside of MIT").   FB is young, "hip", and high energy.  To contribute: http://developers.facebook.com.

Went to Dawn Song's (UCB) on BitBlaze.  Interesting problem.  If you have a  program P, and a public patch is issued, apply it to yield P'.  Then compare P and P'.  You know know where the vulnerability is, and might be able to analyze what it is (e.g. integer overflow).  Then using various techniques, you can generate an exploit of the vulnerability much quicker than people will patch the vulnerability.  Microsoft has "patch Tuesdays".  The exploitation rate goes up almost immediately.  She bascially ran out of time and I had some trouble understanding her at at times.

Attended a meeting of the Database group meeting (Sam Madden, Mike Stonebreaker).  Discussed the work the Ph.D. students were doing.  Also discussed the Facebook talk earlier by Michael Schroepfer.  Stonebreaker doesn't believe FB will scale much further, no more than double.  It's a hard problem when you think about it.  You have to forge a lot of other people's data that is arriving randomly.  I suggested maybe a different UI might help.  The students also discussed the automobile application they've built, which is similar to SAP's CarWeaver.  I offered to put them in touch with Keith Klemba.

Had dinner with Rick Schantz, with whom I attended grad school.  Rick has had the same email address (schantz@bbn.com) and the same phone number and desk for 36 years.  BBN is currently being sold to Raytheon.

Friday, 25-SEP 2009

Talked to some recruiters to give a reference for a friend.

Met with Saman Saman Amarasinghe.  He has the MIT Compiler group.  They're running 4 projects:
Met with K. Krasnow Waterman, who is an attorney, was in intelligence in the government, and is a computer scientist in CSAIL.  She is a colleague of Tim Berners-Lee.  She is working on software that may be suitable for an academic research contract for SAP/GRC.  I'll ask Paul Hofmann to take the lead.  She has software that will analyze contract language, understand it, and point to relevant statutes.  Once "understood", a user can ask questions about it (e.g. like "Why?").  It could be interesting.

Met with SAP employee Paul Wagner (who is reading this diary).  He's asked me to review a paper he wants to write, and wants to explore attending Sloan.

Thinking about what I want to cover in my meeting with Vishal on Tuesday.

Flying home tomorrow.

Saturday, 26-SEP 2009 through 04-October

Away from Cambridge for a week.  Went home for Yom Kippur (took a day off), spent a day in the office (Staff meeting, review of Stratus project), then headed over to Tucson for the ABI Grace Hopper conference and ABI Board of Trustees meeting.  Returned to Cambridge Saturday October 3.  Had dinner with David and Suzie Schuller on Sunday October 4.

Monday, 05-October

Did a lot of catch up.  Starting reading collecting and reading papers on synchronization.  Attended a graduate student microkernel reading group and joined it, and collected a nubmer of microkernel papers.  Gave a short talk to the group on the history of microkernels (RIG -> Accent -> Mach (Encore) -> MacMach -> Mac OS X.  (I hope I didn't come on too strong.)  Now to finish the synchronization papers and the microkernel papers.

Met with Michael Krigsman, ZDNet blogger covering SAP.  He claims the conversation is off the record (but we didn't get into anything sensitive).  He mostly is focused on his company which consults on analyzing and preventing software project failures.

Tuesday, 06-October

Met with our SAP's new CIO Oliver Bussman over the phone.  Offered help, but mainly wanted to discuss the need to help with innovation but taking a more modern view of IT infrastructure.  In particular, sent him and went over the points made by the Apple representative.

Met with Victor Zue who gave me an update on DARPA, and told me about the university mailing lists, and President Hockfields (and Obama's) trips to China.  Offered help on both DARPA and China.  He will send me a near final draft of the MIT China report when ready.  Will attend the MIT China Group lecture tomorrow.  Victor wants me to meet Frans Kaashoek.  [There is a 7th floor systems group, and a 9th floor systems group, with the database people in between on the 8th floor.  Is there a competition between the 7th and 9th floor?]

Met with Todd Glickman of MIT's Industrial Liason Program.  Outlined the 5 points that caused us to not re-enroll in the ILP:
1.  Difficulty in organizing an MIT ILP trip to Silicon Valley
2.  CSAIL is closer to our fields of interest than the more general and more encompassing MIT research agenda
3.  Sloan School Center for Digital Businesss  better helps our thought leadership and marketing efforts
4.  We have enough contacts to get the meetings we need to get at MIT without the need to engage ILP
5.  The annual ILP conferences were, from a technology perspective, not specific enough for us to justify travel to Cambridge.

Attended Anant Agarwal's team meeting, where he introduced me to the group.  Had a discussion about the fos operating system.  There is a lot to learn.  A reading list was published, so there are a lot of papers to read and a schedule to discuss.  Learned a little about the exokernels being studied on teh 9th floor.

Not feeling great.  Went home around 5:30, but found that somehow my  iCal calendar lost all my Exchange calendar entries.  Thank goodness for my backups.

Wednesday, 06-October

Talked to Languilli at Stony Brook this morning.  He asked me to be on the Computer Science Advisory Board that Ari Kaufman is setting up.  I declined, but will recommend some other computer scientists.  He wants SAP representation however.

Made my lunch today, but then left it in the refrigerator.  Not impressed at all with the lobby cafe.

Met with Liami Youseff, a postdoc from UCSB, who was advised by Rich Wolski (Eucalyptus).  Small world.  Discussed with her the problem of dynamica allocation in an fos environment.  I still maintain that to design this system you have to make a call on shared memory vs. no shared memory.  Need to go back and read the paper to see if in fact that call is made.

Attended a "university lecture" by Professor Xu Kuangdi, President, Chinese Academy of Engineering .  He gave a good talk about lowering the carbon footprint in China over the next 40 years.  I would have like to have talked to him, but he was wisked away immediately after the Q&A (some of which got political).

Read what is probably a seminal paper, "Concurrency Control in Groupware Systems" by Ellis and Gibbs (1989).  Sent an email to Irene Grief at IBM asking about the theoretical basis of groupware. (In the cloud, generally, objects are shared and persistent.)  I just realized that the old distributed applications like NeWS, HotJava, and Bell Labs Blt terminals are precursors to what today we'd call Ajax!  Also, I read a follow on paper that imposes some simplifying assumptions, "High-Latency, Low-Bandwidth Windowing in the Jupiter Collaboration System" by Nichols, Curtis, Dixon, and Lamping.

Thursday, 07-October

Talked to Orestis Terzidis in Karlsruhe this morning.  Discussed my work at MIT.  He wanted to hook me up with Daniel Scheibli.  I sent him links to my talk and asked Wes to send him a link to Hasso's video.

Stayed in the apartment to have an SAP IT Security Council Emergeny telecon.  Unfortunately, found the phone system could not make any outgoing calls, and had to do the call on my mobile phone.  I was pretty angry with the apartment rental company since I've had ongoing reliability phone system (VOIP) problems.  After some very heated discussions,  they agreed to change the phone system.  It was installed Friday.  Had to "hack in" to figure out the voicemail system, and wound up writing an email to them to explain how it works.

Went to the faculty lunch, where there was a short talk on flying robotic, autonomous wing-flapping mechanical birds.  It's pretty interesting.

Went to the database group meeting (Madden, Stonebreaker) and heard a talk about invertable relational database schema mappings.  The key idea is to be able to transform the queries that were targeted to the old schma to the new schema.  Seems limited to me.  There was a discussion at the end about the inability to handle stored procedures, which was deemed crucial to the work.  Since SAP uses stored procedures extensively, I agreed.

Met K. Krasnow Waterman on the street going back to the apartment, and we decided to have dinner together at the Marriott sports bar.  Good conversation.

Friday, 08-October

Had breakfast with Alan Kay.  It's always interesting.  We discussed several things including:
A business contact recognized me there, and later Nicholas Negroponte walked in and joined us briefly.  It's really interesting who you meet at the Kendall Sq. Marriott.  If I had time, I'd hang out there more often.

Had a telecon with Rainer, Thomas, Petra F., Gunther S., and Jason Y. to go over a new GBI initiative.  I think it's worth pursuing.

Had lunch with one of my old Apple employees, John Hotchkiss.  Good to catch up.

Had a really good meeting with Frans Kaashoek, who gave me some papers to read.  He expressed some concerns over the fos approach, and we shared some others.  One of his projects is particularly worth exploring: DProf, a data profiling tool.  I'll look into this.

Weekend, 09-October

Mostly relaxed, went shopping, etc.

Monday, 12-October

MIT is closed for Columbus day, but I went over the the MIT Startup Bootcamp conference in Kresge, all day.  The talks were all good, most were excellent.  However, there was a lot of uniformity among the companies selected - almost all were web properties.  I particularly enjoyed the talk from the founder of ZipCar.  I took extensive notes.

Tuesday, 13-October

Papers read today:
Had a meeting with the Carbon Group.  Jim Psota wants to meet with me to discuss software architectures for optical interconnects.  I'll meet him tomorrow.
Met with the fos group, but Anant was somewhat delayed, and I had a dinner appointment, so I couldn't stay for the entire meeting.  Anant wanted to discuss the differences between fos and Barrelfish, but fos hasn't yet tackled some of the problems that Barrelfish has, particularly in the area of what I'll call coordination.

Met Phil Dinsky for dinner.  Phil was my great next door neighbor and friend for several years in Framingham.  It was good to catch up.

Did a lot of SAP email tonight.  It's hard to do both jobs.

Wednesday, 14-October

Met with Tom Psota who is designing a system around a ring based optical interconnect.  Very interesting.  Wanted my thoughts on how to architect the software that will run on it.  We had a very interesting discussion.  I think this may be a good opportunity to on the one hand leverage the fos work, but on the other to solve the coordination problem I see looming with fos.  This problem needs to address the dynamic load leveling case.  Also, it may be that the circuitry for the wired mesh interconnect is unnecessary.  Using an idea that came out of my discussion with Alan Kay last week, maybe the thing to do it to cached the server code, and just have each core be "schizophrenic" - download the personality and keep a cache of personalities, throwing them away as resources and dynamic execution requires.  Too complex to go into here.  We'll make an appointment with Anant Agarwal to discuss with him.

Read the paper "User-Level Interprocess Communication for Shared Memory Multiprocessors" by Bershad, Anderson, Lazowska, and Levy, ACM TCS 1991.  They use some of the ideas Nico Habermann and I had in our early paper on Efficient Implementation of Ada Tasks.  We should have really published that paper.  I'll need to scan it in when I get home (If I can find it).  It's heavily referenced on the Internet, but I can't find it anywhere out there.

Awful headache today.  Heading home early.

Thursday, 15-October

Meeting with Michael Siegel and Daniel Goldsmith of the Sloan School.  They are proposing using simulation to do cost minimization/benefit maximization of behavior, workflow, and policies.  Will connect them with Paul.  Three specific topics: cybersecurity, software maintenance as a long term business process, and hospital/healthcare.  They use Forester's "System Dynamics" approach.

Met with Dave Gifford on long term view of academic computer science.  How to rejuvenate the excitement of the field?

Had a concall on the SAP mentoring program for women.

Attended a talk by Bill McKeeman (co-author of "A Compiler Generator" (XPL)) at IEEE CS/ACM GBC at Sloane.  He spoke about a technique to build compilers through bootstrapping grammars.  Interesting, but not very practical.  I declined the dinner following the talk.  McKeeman is a MatLab fellow and also teaches at Dartmouth.  What was interesting was that I got to see a lot of old friends and colleagues, who I haven't seen since well before moving to California 16 years ago.

Friday, 16-October

Attended a talk on "Leakage Resilient Key Proxies" by Yevgeniy Vahlis.  One of the faculty was very disruptive.  It was about constructing an crypto scheme in the presence of leakage on O(log n) in the length of the message, and how to avoid the inherent problems of leakage (sound, radio, voltage, etc.)

Read a paper: "A Software Approach to Unifying Multicore Caches" by Kaashoek et al.  Key idea is to move the threads to the core nearest the cache it needs.  Implements a new file cache for Linux called MFC to demonstrate the viability of the approach.

Read a paper: "Exokernel: An Operationg System Architecture for Application-Level Resource Management" Kaashoek, et al.  A very minimal OS with hw management in the kernel and the policies in user libraries.

Will pack tonight for California.