Ravi Pandya
ravi@iecommerce.com
www.iecommerce.com
+1 425 417 4180
vcard

syndicate this site

ARCHIVES

2007 11 10

2004 10 09 08 07 06

2003 04 02 01

2002 12 11 10 09 08

2001 11

ABOUT ME

Ravi Pandya
Architect
Server & Tools Business
Microsoft
ravip at microsoft.com

03-	Microsoft
00-02	Covalent
97-00	EverythingOffice
96-97	Jango
93-96	NetManage
89-93	Xanadu
88-89	Hypercube
84,85	Xerox PARC
83-89	University of Toronto, Math
86-87	George Brown College, Dance
95-	Foresight Institute
97-	Institute for Molecular Manufacturing

DISCLAIMER

The opinions expressed here are purely my own, and do not reflect the policy of my employer.

Twitter

I'm posting mostly on Twitter now - follow @ravinpandya

06:32 #

Orleans: Cloud Computing for Everyone

http://research.microsoft.com/apps/pubs/default.aspx?id=153347

Our paper on Orleans was accepted to the ACM Symposium on Cloud Computing in Cascais, Portugal, October 26-28. If you're going to SOSP or SOCC, I'll see you there!

06:45 #

When not to think different

Sadly, if Steve Jobs had listened more to his rational side in this case, he might well have lived many years longer:

Steve Jobs Succumbs to Alternative Medicine

"Most pancreatic cancers are aggressive and always terminal, but Steve was lucky (if you can call it that) and had a rare form called an islet cell neuroendocrine tumor, which is actually quite treatable with excellent survival rates - if caught soon enough. The median survival is about a decade, but it depends on how soon it's removed surgically. Steve caught his very early, and should have expected to survive much longer than a decade. Unfortunately Steve relied on a naturopathic diet instead of early surgery. There is no evidence that diet has any effect on islet cell carcinoma. As he dieted for nine months, the tumor progressed, and took him from the high end to the low end of the survival rate."

21:13 #

In memoriam

Two roads diverged in a wood, and I-
I took the one less traveled by.
And that has made all the difference. *

Thank you, Steve, for everything you have created.

My first computer was an Apple ][+. I mowed lawns to buy it, and sold greeting cards door-to-door to expand it to 48k and buy a 140k disk drive. Adjusted for inflation, it is the most expensive computer I have ever owned. And worth every penny, for what I learned from it and created with it.

05:57 #

Black Diamond Long Course Triathlon (70.3)

This race doesn't start until 9:00, so it's easier to manage getting there in the morning than most races. I arrived at 6:45, and was early enough to get an onsite parking spot. It was shaping up to be a beautiful sunny day, like last year - pretty lucky for Seattle in late September. I had plenty of time to set up in transition, rest, eat, and do short warm-up run.

See more ...

19:37 #

Ironman 70.3 Lake Stevens

This was my second triathlon, after Black Diamond last year. Overall, it was a good race, though I had a couple of problems. I had a flat on the bike course, which took me an unbelievably long time to fix. And I was short on salt & hydration, so I cramped up during the run.

See more ...

05:32 #

The Practice of Triathlon: Three Meditations

Slicing through water
Line from palm to arch of foot
Half-glance back to breathe

Pedals spin like silk
In a compact aero curl
Wake streams out behind

Feet strike light and quick
Earth rotating underneath
Branches pass above

09:11 #

Orleans: A Framework for Cloud Computing

For the last year, I've been in the eXtreme Computing Group (XCG) in Microsoft Research, working on the Orleans project. Our goal is to make it much easier to write reliable, scalable, distributed cloud+client applications. I'm really enjoying it, and it's great to be working on something that I can talk about publicly.

We now have a video on Channel 9 with the whole development team discussing the project:
http://channel9.msdn.com/Shows/Going+Deep/Project-Orleans-A-Cloud-Computing-Framework

We've also published a technical report describing the overall design and architecture of the system:
http://research.microsoft.com/apps/pubs/?id=141999

There's also a brief summary of the project here:
http://research.microsoft.com/en-us/projects/orleans/default.aspx

Feel free to email me if you have any questions or comments. (Or if you're interested in working on this kind of thing - we've laid a great foundation so far, but there are plenty of interesting problems still to work on, like better programming language support, or large-scale data- and compute-intensive applications.)

16:21 #

UW Chamber Dance Company

http://depts.washington.edu/uwdance/performances.html
Their performances are always worth seeing, but this year's program is particularly good. Dance is such an ephemeral art form - books and music can sit onthe shelf waiting to be picked up, but a dance takes weeks or months of rehearsal to bring it to life. Every year, Hannah Wiley at UW chooses historically interesting modern dance works and sets them on the grad students. It's a great opportunity to see works you might not otherwise have the chance to experience - from Oskar Schlemmer's minimalist Bauhaus Dances to Alwin Nikolais' colorful and vibrant Pond.

06:13 #

BrickCon

I went to BrickCon at Seattle Center with Gavin today. It was a fun but intense total immersion in Legomania. Some interesting links:
First Lego League of Washington
The Great Brick Contraption - a modular system for building Rube Goldberg contraptions
HiTechnic - new and interesting sensors for Mindstorms. One of the guys at the booth had built a two-wheeled balancing robot like a Segway!
Blakbird's Technicopedia
And we came home with an awesome Mobile Crane set from the Math'n'stuff booth.

16:43 #

Large-scale Incremental Processing Using Distributed Transactions and Notifications - Peng Dabek

http://research.google.com/pubs/archive/36726.pdf
This is a description of the "Percolator" technology behind Google's new incremental "Caffeine" search index. It produces ~100x lower latency than the previous MapReduce-based system, at about a 2x cost in resources. It's a system for dependency propagation using observers on BigTable columns. It also leverages BigTable transactional semantics to provide snapshot isolation and other transaction management services like straggler elimination.

07:42 #

Black Diamond Half-Ironman Triathlon

This was my first triathlon, so I had tried to prepare carefully. I had written up a race plan & checklist, packed my bike and gear into my car the night before... and realized in the middle of the night I had forgotten my bike shoes. I was up at 4:40, had a big mug of black tea and peanut butter & banana on toast, and headed out to Nolte State Park. I was there a little after 6:00, and was one of the earliest arrivals. I was the first through body marking, and set up my bike and gear in the transition area. I drove over to the parking area, took the shuttle back, and had a couple of hours to wait until the 9:00 start. I had OJ & some oatmeal with yogurt and applesauce (from our own trees), took my P3-SL over to the Enumclaw Sports booth to have the chain lubricated as it had gotten rusty after commuting in the rain, called my family, and gathered my thoughts for the day.

See more ...

06:09 #

Almost a Dynabook

If only the iPad ran Scratch and Squeak. Maybe Alan Kay can get Steve Jobs to make an exception and allow them into the AppStore... In the meantime, there's Dan Ingall's Lively Kernel.

05:29 #

In Search of Memory: The Emergence of a New Science of Mind, by Eric Kandel

http://www.amazon.com/Search-Memory-Emergence-Science-Mind/dp/0393329372
This is an engaging and inspiring scientific autobiography. Kandel chose a deep and complex problem to study, the physical basis of memory formation, linked in the book to his vivid childhood memories of Nazi-occupied Vienna. He then spent decades studying it in the simplest possible organism, a sea slug, and eventually teased out various ways in which memory formation results in physiological changes in nerve cells and connections. Interestingly, one of them has to do with prions that act as a bistable latch.

07:28 #

Brains, Meaning and Corpus Statistics

http://www.youtube.com/watch?v=QbTf2nE3Lbw
Fascinating - I listened to this while I was unpacking my office yesterday. Applying machine learning to fMRI images and web word proximity data, they were able to distinguish which of two words a person was listening to - without having trained on either the words or the person's fMRI data. Tom Mitchell has a number of other interesting papers on his website:
http://www.cs.cmu.edu/~tom/

05:39 #