Harry Chen Thinks Aloud

Today I received a newsletter from my mortgage lender, in which it says two things:

1. experts predict that home-price in the US will continue to rise
2. now it’s a good time to take your family on a cruise by taking a home equity loan

Is it really wise to take out a new loan for vacation when the price of your house soar? I think not. I believe it’s a bad idea to take out a loan to pay for leisure expenses. If you can’t afford it, don’t go borrow and pay for it.

I think many mortgage lenders are pitching the wrong information to the consumers. Not only this unsound advise is not helping the consumers financially but also encouraging them heading into a wrong direction.

–
Posted by Harry Chen to money at 10/28/2005 08:36:15 AM

The idea of tagging web information has been around for awhile. Given that many popular web applications (e.g., Gmail, Yahoo! My Web) now support the use of tags. I believe that in the future more information on the Web will be tagged by some kind of text labels one way or the other.

So what does that mean in the context of the Semantic Web? Will the emergence of information tagging help to speed up the Semantic Web development?

I believe information tagging will bring us one step closer the Semantic Web. For a long time, Semantic Web researchers have been asking the question that given the vast amount of web information that is already existed, how are we suppose to annotate each and every single piece of this information?

I think tags can help to solve this problem.

Let’s assume that we can build automatic or semi-automatic computer programs, called pre-processors, to tag non-semantic web information. Once we have created such programs and run them against different non-semantic web information, we will be able to collect summaries of this information in the forms of tags.

Now, let’s assume that we can control this pool of possible tags that are produced by our pre-processors, and we can define ontologies (using RDF and OWL) to express the explicit semantics among these tags. For example, an ontology may define “blogging” is a subclass of “personal journal”, and “personal journal” is a subclass of “historic document”. All “personal journal” instances must have one and only one property called “author”, and its value must be a type of “person”.

Let’s say we run our pre-processor programs against my personal blog posts, and the set of tags produced by the program includes “blogging”, “harry chen”, “semantic web” etc.

Let’s also assume that people on the Web can make explicit statements about various tags. For example, as a part of my personal profile on blogger.com, it has an RDF statement says tag “harry chen” is a type of “person”.

Say some person Bob on the Web is interested to find all historic documents that are written by some person, and these documents are about the subject “the Semantic Web”.

To answer his question, Bob would define a query similar to this, “find all URI X that are instances of the historic document class, X must have one or more author, and the document that each URI X references must contain the ’semantic web’ tag.”

Because my blog posts have been tagged with “blogging”, so they are instance of the “blogging” class. Since “blogging” is a subclass of “personal journal” and “historic document”, all my blog posts are also instances of these two classes. Because I’m the author these blogs, these blogs have at least one author.

Given the above reasoning, my personal blog posts will be one of the many answers that match Bob’s query.

What do you think?

–
Posted by Harry Chen to semweb at 10/27/2005 12:01:09 AM

I believe in the next couple years the US market will not outperform the rest world like it used to in the 90’s. I think inflation will increase more than we have expected in 2006, and US productivity will remain flat. When real estate cools down, consumers will slow their spending for psychological reasons.

Recently, I read few articles that discuss about how people can protect their investments in 2006. Few recommendation is to reduce asset allocation in the US market and increase asset allocation in the foreign markets. I believe China is a good place to put your money.

• IHT: Around Asia’s markets: Optimism is shifting from India to China
• BusinessWeek: Betting on China’s Banks

–
Posted by Harry Chen to money at 10/25/2005 05:16:00 PM

I downloaded a copy of the SemanticWorks 2006, an ontology editor released by Altova. Here I report my 30 minutes user experience with this software, without reading any Help documentation.

The installation of the software was easy. The GUI of the application looks very nice and professional. It looks and feels very much like the XMLSpy — an XML editor by Altova.

Some shortcoming of the software:

1. It doesn’t seem to understand the OWL imports construct. Or, maybe this function is in the software, and I just haven’t figured out how to do it yet.
2. I like to build modular ontologies, and often I need to defined cross-referenced ontology concepts. This software lacks good editing support for ontologies that are stored in multiple files. For example, the software doesn’t allow me to create an instance X of some class type C, if C is not physically defined in the same file as X.
3. When editing ontology in the graph view, few wrong clicks can cause the software to crash.

After all those negative comments, I think I still like SemanticWork 2006 because it’s the first commercial ontology editor. It seems like a good RDF editor, even if it’s not a good OWL editor. Maybe because it’s the first release version of the software, many of its functions are still rough.

If you have good OWL ontology editing experience, please let me know.

–
Posted by Harry Chen to semweb at 10/25/2005 04:53:39 PM

I know I’m a tech-savvy person, but I would never guess that the county that I’m living in is also a tech-savvy county.

That’s according to USA Today and the Claritas marketing search firm, which set out to find the U.S. counties with the most technologically advanced households. The survey ranks Howard County second and Montgomery County third. Broomfield County, Colo. topped the list.

See this Washington Business Journal article.

For awhile I wasn’t sure whether I should have both 401K and IRA. The more I think about the issue, I come to the following understanding:

• If you just have an IRA account, even you max your contribution, it’s not going to be enough for your retirement. Let’s assume that you save $4000 each year in your IRA, annual yield is 8.5%, and annual inflation is 4.5%. In addition, let’s assume you desire monthly withdrawal of $2500 (today’s value) when you retire. Based on this calculation, the total amount after 30 years will only last you no more than 6 years.
• If your company matches part of your 401K contribution, then you are actually getting a raise from these matched employer contributions.
• While contributions that you make toward your 401K can reduce your income taxes, but you are required to pay taxes when take money out of the account. However, IRA (Roth IRA) works just the opposite. IRA contributions can’t reduce your taxes every year, but it’s tax-free when you take money out. If you bet that income taxes rate will rise in the future, which I believe is likely, an IRA account is a good shelter place.

–
Posted by Harry Chen to money at 10/23/2005 03:31:16 PM

It took me awhile to figure out how to add a “category” feature to my blog. My blog is hosted on Blogger, which doesn’t have built-in support for “category”. I found a tip described by OldCola — Blogger Categories, and it proved to be pretty useful. The procedure is a bit complex, but nevertheless it works!

Given the continuous economic booms in the East, some people begin to think that in next several decades, countries in the East will take over the U.S.’s role as the world’s economic leader. There is two way to think about this issue. One, this situation is not going to happen because back in the 80’s there were talks about how Japan would become the world’s leading economy, and that did not really happen. Second, the U.S. will actually lost the economy battle because things are different this time around. Many developing countries are following the same recipe that the U.S. had adopted to develop its great economy — e.g., invest in new scientific research and encourage innovations.

If the first were true, there is not much we have to do. However, if the second case were true, then the U.S. government should wake up now and take this issue seriously. Why? According to this IHT article “Economic View: U.S. paradise lost?“, in recent years the U.S. government has lost interested in funding new basic research programs, and has been discouraging smart people to come the U.S. to stay and work. I think this puts the U.S. economy in a very dangerous position.