Friday, October 28, 2005

The United States still has a large lead over China, but if the current sad state of affairs with respect to education and research continues here in the US, it's only a matter of time before China catches up.

When Andrew Chi-chih Yao, a Princeton professor who is recognized as one of the United States' top computer scientists, was approached by Qinghua University in Beijing last year to lead an advanced computer studies program, he did not hesitate.

[JIH - Andrew Chi-chih Yao won the Turing Award in 2000]


China has already pulled off one of the most remarkable expansions of education in modern times, increasing the number of undergraduates and people who hold doctoral degrees fivefold in 10 years.


In only a generation, China has sharply increased the proportion of its college-age population in higher education, to roughly 20 percent now from 1.4 percent in 1978. In engineering alone, China is producing 442,000 new undergraduates a year, along with 48,000 graduates with masters' degrees and 8,000 Ph.D's.


The Transportation Experience

One of my friends just published a book on transportation, entitled The Transporation Experience.

While much of the transportation systems in Europe and the United States are mature (if not senescent), the rest of the world is still planning, developing, and deploying new systems. The accomplishments and mistakes of places like the United Kingdom and the United States, then, can teach us lessons that may be applied to places where transportation remains nascent or adolescent. The Transportation Experience seeks to understand the genesis of transportation policy in America and the UK, along with the roles that this policy plays as systems are innovated, deployed, and reach maturity, and how policies might be improved. The work presents case studies of particular transport experiences in rail, road, water and air (with a special emphasis on railroads), and then finds commonalities in all of these experiences with thematic analyses that are often bold and unconventional. The book is predicated on the idea that the story of transportation policy can tell us what transportation, is, does, and might do in the future, and at an even broader level, how society has learned to create, deliver, and operate large, complicated systems. It should appeal to students and researchers in a broad array of fields, including geography, civil and environmental engineering, and public policy.


Thursday, October 20, 2005

Thursday, October 13, 2005

[Ubicomp] What is ubicomp?

One reader asks about ubicomp and its relationship to supply chain management.

[We think ubicomp is] simply the philosophy of using computation to augment the ability of entities in a non-invasive way. If our definition is correct, then would it follow that technologies which can vastly improve transparency across a supply chain without requiring significant process changes be one valid application of ubiquitous computing? If so, then wouldn't any and all supporting technologies (data storage facility, barcodes, etc.) be part of ubiquitous computing technologies within that context?

My response:

Well, on the one hand, the term ubiquitous computing is sufficiently general that it could include lots of things, sort of like personal computing. On the other hand, if you look at the specific research that calls itself ubiquitous computing, you only see certain kinds of things. For example, mobile computing, sensor networks, large displays, etc. Lots of areas with previous work done (ex. databases, hardware architecture, AI) may be doing work that is highly relevant to ubiquitous computing, but often don't call themselves that, probably for cultural reasons.

From my perspective, what is and isn't ubiquitous computing is debatable, and ultimately not that useful.


Ubiquitous computing also suffers from the same problem as AI, which is that the bar keeps getting raised. Things that might have been considered ubicomp a few decades ago (ex. TVs and barcodes and LEDs) are no longer so.

Currently, the main characteristics of ubicomp technologies would one or more of the following:
- Natural interaction (speech, sketching, pointing, etc)
- Mobile, off the desktop
- Awareness of the physical world (location, identity, activity)
- Integration with the physical world (smart cars, smart tables)
- All wirelessly networked

The off the desktop part is important too, as that is part of integrating computation and communication into all aspects of our lives, rather than just a single place in a single machine. I would categorize things like barcodes and data storage as mostly fitting within the desktop environment.

On the other hand, there are emerging technologies for supply chains that I would also consider ubicomp:
- RFIDs for real-time tracking of inventory
- Wearable computers for coordinating employees
- Location-based services for tracking vehicle fleet

So I'd say it's a question of degrees here. While supply chain and ubicomp aren't mutually exclusive, there are some technologies that exhibit more of the characteristics described above than others.

Monday, October 10, 2005

[HCI] Martin Wattenberg Visualizations

Some really amazing information visualizations by Martin Wattenberg. Includes

  • QuerySketch, sketching for retrieval of relevant stock graphs
  • Tree map layout of the stock market
  • History Flow, modifications to wikipedia
  • Site X-Ray, modifying a page in place to see log analysis data

Thursday, October 06, 2005