Comments(3)

Comment by Jason Kemp at 12:52 pm on 19 April 2007

A very strategic purchase with real sysnergies between the two companies.

As a background reason “blocking” Microsoft will also work for them.

There is an interesting anlaysis of this over at Baseline

“In a sign of how concentrated R&D is, the top 10 companies on our list invested $32.5 billion in product development in their 2006 fiscal years. That was almost seven in every 10 R&D dollars spent among the companies.

The biggest spender was Microsoft, with $6.58 billion. The software giant was an exception in that a lot of its R&D spending was devoted to new products aimed at businesses, including Office 2007 and the Vista operating system. But Redmond, Washington-based Microsoft also plowed significant sums into its Xbox 360 gaming console and Internet search.

That field is still dominated by Google, which more than doubled its R&D spending in 2006. Google’s $1.23 billion investment made it No. 9 on Baseline’s list, far ahead of competitors like No. 13 Yahoo ($833 million) and No. 21 EBay ($495 million). But even with its rapid growth in R&D last year, Google was able to devote just 11.6% of its revenue to R&D. The median R&D expenditure among all the companies we looked at was 15%.”

from http://www.baselinemag.com/article2/0,1397,2114821,00.asp?kc=BARSS03129TX1K0000628

The DoubleClick purchase may just push Googles number closer to the 15% level but upshot - Google is in a major leaugue of its own now.

Paul Strassman also makes some other good points about Google & Microsoft over at “Microsoft has an enormous war
chest that can be matched only by Google, which has more
cash on hand, with $11.2 billion in the till.
Microsoft’s position, however, is vulnerable to attack by
software firms that bypass its stranglehold on client devices.
For instance, offering a software-as-a-service substitute for
Microsoft’s Office Suite or SharePoint, instead of a shrinkwrapped
product, promises huge cost advantages for customers.
It all boils down to the question of whether a thin client,
with only a browser, can compete with a fat client with more
than 50 million lines of Microsoft code. I have priced out the
hardware and support costs for installing thin clients. The payback
is less than one year, and the expected life of a thin client
will exceed the useful life of a Vista installation.”

See http://www.strassmann.com/pubs/baseline/2007-05-a.pdf





Comment by John Clegg at 4:20 pm on 19 April 2007

I wanted to add that CNET has more really good videos from thethe Web 2.0 expo in San Fran.

Check out :

Yahoo competing against Google
To flip or not to flip.





Comment by JimW at 11:12 pm on 25 April 2007

Now, Google has recently got into LSI (also known as LSA) (Latent Semantic Indexing or Analysis) document search, which I came across here:

Google Latent Semantic Indexing Technology

This makes Google’s search engine more powerful in combining link-analysis-search type algorithms which is what their PageRank algorithm is, and content-search algorithm as LSI. Google now, can capture search terms via popular links as well as document semantics (documents with similar concepts). I see that DoubleClick are already using Text Mining for marketing search engine and LSI is one of those Text Mining algorithms. Most web site of today are still using Literal search terms (key-words) and Boolean search. They are fine when the user know exactly the words to type in to search for. However they fail miserably when the user try to type in some words which are vague and imprecise.

I am surprised that Google has just incorporated LSI search capability into their search engine recently , where LSI had been around for over a decade. There are lots of LSI publications on the internet , but developers can download this paper with detail description of the algorithm:

Using Linear Algebra for Intelligent Information Retrieval

LSI is based on an algorithm called SVD (Singular Value Decomposition) and there are open sources implementation of SVD in different languages available from the internet. But how to use SVD for search is described in the paper above.





Name:(required) Email address (required, will not be published): URL (optional): Comment:

Notify me of followup comments via e-mail