Comparing Personalized Homepages

I have tried out the personalized home page option for the 3 major sites: Google, MSN (My MSN), and Yahoo (My Yahoo). All three do a good job to allow you to set up page(s) where you can see the information you care on a single page, and users can easily add/modify/delete content on all 3. Here I will focus on the differences among the three. I will also list recommendations on which one to use based on which criteria you value the most.

Differences

Below are features that are different among the three:

Presence of ad banner: Google and MSN do not have ad banners, while Yahoo does.

Continue reading

Microsoft introduces LinkfromDomain operator on live.com

Microsoft has introduced the LinkfromDomain operator on live.com.¬ This operator allows users to see all¬ outgoing links from a¬ top-level domain.¬ This is basically the opposite of the LinkDomain¬ operator, which allows users to see all¬ incoming links to a domain.

Whlie test driving this command, I noticed the following:

1.¬ LinkfromDomain is not available for subdomains, while LinkDomain is.¬

2.¬ ¬ The LinkfromDomain operator will give all links from that domain, regardless whether that link has a rel=nofollow designation.¬ Also, links buried in Javascript are also displayed.

3.¬ The LinkfromDomain operator is a good way of identifying outdated links from a site.¬ This is very useful if you are the site owner.

Google-Youtube deal announced

Google announced its acquisition of Youtube after the market closed today in a stock purchase valued at $1.65 billion.¬ I’ve outlined why I thought this is a good purchase for all parties involved in the previous post.

The fallout with this purchase is that Yahoo will most likely speed up its courtship of Facebook, which is the next biggest Web 2.0 site that is available on the market.¬ The Google-Youtube deal, however, does give Facebook more leverage in its negotiations with all suiters.¬ Expect Yahoo to announce its purchase of Facebook soon, and at a higher price than the $1 billion that has been mentioned recently.

Google – Youtube. A good deal?

There is rampant speculation that Google is close to buying Youtube.¬ Let’s take a look at whether this is a good deal from 3 respects:¬ Google, Youtube, and users.

Google

1.¬ This acquisition will instantly make Google the #1 player in online video, currently the hot trend on the internet. ¬ Outside of search, Google really doesn’t have any product that it can claim to be #1 in its category.¬ Getting Youtube will change this.

Continue reading

Google Code Search

Google released Code Search, a search utility to find computer codes across the web. Users can type in a snippet of code they are looking for (including regular expressions), and Google Code Search will return the codes that contain that particular snippet. People can also search by specifying the file name, package name, language, or license they are looking for.

I tried it out and did not find it to be very useful. True, if you know the exact line of code or the exact file name you are looking for, this can be useful. However, I am usually more interested in finding code that performs a certain function in a certain computer language, such as “I want to find a piece of code that sessionizes raw web log data in perl.” To be able to accomplish this type of search, people will either need to tag their code (probably not happening anytime soon), or the search engine will need to be able to decipher the information hidden in the comment field within the code (this is more likely).

Combo Search

Combo search refers to a different way of presenting search results. In the Google world, people are used to see pages after pages full of only web search results, or image search results, or new search results, etc (with the exception of some sponsored links). This is, however, not the way search results are presented in Korea.

Let’s take an example: search for “cat” in Naver, Korea’s top search engine. You’ll see a long results page divided into many different sections, including sponsored results, news results, category results, local results, book results, movie results, blog results, music results, knowledge results (similar to Yahoo Answers), as well as web results. This way, users are exposed many different types of results for that query term, and the user experience is entirely different.

This type of search results display is possible in Korea because of its high broadband penetration rate. With dialup access, this type of display would simply take too long to download.

This is perhaps why Google is not able to gain much in the Korea market. Google is noted for the simplicity of its pages, but this may work against them in Korea, where users have come to expect to see a large variety of pages.

In this type of system, SEO for web pages becomes less important because web results themselves are less important there. Even if you make it to #1 in web search, you might still be way below the fold (depending on where the search engine decides to place the web results section). In Korea, SEO thus means looking into image search, book search, etc, in addition to simply web search.

Research in Search Engine Labs

There is an article on Search Engine Watch that talks about the research going on in major search engine labs. Below is what each search engine lab is focusing on:

Microsoft adCenter Labs

  • Paid search
  • Contextual ads
  • Behavioral targeting
  • Emerging markets

Yahoo Labs

  • How to know what to believe
  • Trust models online and propagation
  • What makes communities thrive and wither
  • Tagging images and videos, and sharing these file types

Google Labs

  • Image processing
  • Fact extraction
  • Statistical machine translation

Out of all the above, the only one that gets me excited is Google Lab’s image processing. I’ve always thought that the tagging mechanism currently used for searching for an image (or video, or other media types) is only a temporary solution. The real end goal is to be able to search an image based on what the image looks like (show me pictures that look very similar to this image). The same thing with goes with video (based on what’s said) and audio (based on what’s said and the action).

There are several companies working on voice recognition and image mapping. One or two of them will make it big.

Google Webmaster Tools: Part 3, Statistics Reports

This blog talks about the reports available under the Statistics tab in Google Webmaster Tools. The 4 main sections are: Query stats, Crawl stats, Page analysis, and Index stats. Let’s take a look at each one below:

Query stats

This page displays two main reports:
1) Top search queries (queries that most often return a page from your site), and your site’s position for each query.
2) Top search query clicks (queries that generated a click to your site), and your site’s position for each query.

You can also drill down by search time (web, image, etc) or search location (google.com, google.co.uk, googl.ca, etc — note that does not represent where your user is coming from. It simply indicates which Google search engine the user used).

Crawl stats

This report shows PageRank distribution of the pages in your site, as well as which page has the highest PageRank. This is a functionality that I can do without — Google should either provide PR information on all the pages, or just get rid of this report.

Page analysis

This page has two sections: Content, which shows the type of documents Google found on your site, as well as the distribution of language encoding for the pages on your site. Common Words, which shows the words most commonly found on your site, as well as the anchor text most often found in links pointing to your site.

Index stats

This is simply a list of commands that you can use to find more information about your site. For example, site:www.yoursite.com can be used to find the indexed pages from www.yoursite.com.