Comments

Finally, I decided to post as regularly as possible :) : I have been commenting on various websites on various topics. I thought I'd rather post those comments in a collective manner in my blog (sometimes more elaborately) rather than just leaving my blog for the dead.
So, here goes a comment from this week:
Jack reported about trends in search engine for embedded, FPGA, and quantum computing in his editorial here.
===== snip =======
My thoughts: "Embedded" keyword should be rather indirectly derived from searches like "iphone" or "portable mp3 players" or "robotic control system" etc. I think that most of the embedded products today are so metamorphosed into various forms for average consumer that "embedded" or "fpga" or "firmware" is a keyword that may be directly used rarely.
Also as programmers, we have learned to search not just "embedded program" because there is so much of cross-platform development tools that a "c program code " on any platform becomes more or less suitable for an embedded system.
Or a search for any kind of RTL code could be more revelant a trend than looking for search patterns on "FPGA", I guess.
So, the above results doesn't seem to be too surprising a fact for me!

I would rather suggest that we looked at "clustered" search trends like www.clusty.com (boy, they do not have the trends feature yet!) - hopefully getting better results. I am just waiting for "clusty labs" to add the trends feature anyone listening?
Least, I expect that any trends on embedded can be predicted accurately with the search done in any of these engines - because embedded spans such a broad variety that it may even be tough to get it clustered !
===== snap =======

I have been using clusty for a while and impressed with its results ! Its good to have results clustered so that I can view what I want rather than having a mixed bag of results. For example, a simple search for "RTL" in google gives:
http://www.google.com/search?q=rtl&ie=utf-8&oe=utf-8&aq=t&rls=org.mozilla:en-US:official&client=firefox-a
where the results start from some radio network in germany which I am not interested in...
A search on clusty on other hand, gives me an automatically categorized list:
http://clusty.com/search?v%3afile=viv_817%4028%3a03JM3V&v%3aframe=list&v%3astate=root%7cN760&id=N760&action=list&sw=%7cRegister%20transfer%7c&sec=1192930712&

This can help me narrow down the search to what I want rather than just take what is given !

What do you think? Isn't a trend based on clusters better than a trend based on just unrelated "keywords"?!