Hidden web standards behind Google "custom searchEngines"?
- by Hoàng Long
Today while playing with Google Chrome Omnibox, I notice a strange behavior. I guess there's some "hidden" web standard behind it, but can't figure it out. Here's how to reproduce:
Go to http://edition.cnn.com/
Use the search function at the higher right corner, Search a random keyword, for example: "abc"
Close the tabs.
Open a new tab, type until Chrome reminds you about http://edition.cnn.com/, then press "Tab"
The Omnibox now shows "Search CNN.com"! And when you type "abc" and press Enter, it uses the CNN search function to do the job, not Google!
I also tried it for several different sites. To some it won't work. But to some sites, like CNN, vnexpress.net, it works after I use the search function of that site once.
I also learnt about chrome://settings/searchEngines (type it in your chrome box and you will see), and learnt about you can add custom search engine in chrome. But the question is, why Chrome can realize the search URL automatically to some pages, and not others?
It's not because some site subscribe to Google service, because I can do the same method for my site (http://ledohoanglong.wordpress.com), and I'm sure that there's no subscription. So I guess there's a method to "expose" the search function of a site, so that Google Chrome can catch it (after I call the search function of that site once, of courses).
Does anyone know about how it works behind the scene?
