Hacker News Reader

Comment by jgrahamc | Hacker News Reader

I don't think this is very compelling. The example given (searching for a particular Marriott) works flawlessly in Google and I can get at much more information than Open Graph tagged sites will give me.

What problem is this solving?

View on HN

What facebook is trying to build is not a search engine. Rather, I believe they're building more of a recommendation engine. As your friends (in theory) begin to "like" a bunch of stuff on the web, next time you do a search for hotels, the hotels your friends "liked" will show up.

The meta data for each site/artist, etc you like will also give them (in theory) enough data to recommend not just on what your friends like.

Whoever unlocks the magic in the data and offers a compelling, personalized experience will rule the next decade.

Problem this solves; making search more relevant & personal....

Further whose advice and opinion holds more weight for you in regards to where to stay in Hawaii, best restaurant, etc? Your network of friends/family/contacts or a search engine? Who holds more influence over you for vacation spots, restaurants, experiences with businesses? Some examples I can think of now of how this is a better form of search. Though Google wins for me when needing directions, queries on pieces of history and the like; things your friends do not know but would tell you to search Google.

Overall i don't get the whole business mantra "What problem does this solve," when creating a start-up. SOme of the biggest sites on the web (Facebook or Twitter) were not solving any problem - rather they created new ways for us humans to communicate.

I think Facebook is vastly overestimating the number of people who are actually going to go out and "like" stuff.

The effort of Liking is easily an order of magnitude less than creating a link, which is what (nominally) Google uses.

And I'd SWAG that an order of magnitude more people are capable of Liking than are even able to create a new link.

The effort to implement a like button on a site is an order of magnitude more difficult than implementing a link.

Actually, it's not. Instead of an a tag, it's an iframe where the link is passed in as a parameter.

It took me approximately 90 seconds to implement from looking at the docs to deploy to one of my sites.

Technically 90 seconds is an order of magnitude more difficult than making a regular link. Not a huge hurdle, but I don't need to read any docs or go fetch embed code to make a link.

... I guess it's also because I'm using Rails, so I'm already generating links via a helper function. So I wrote another, one line helper that does the wrapping. So that initial 90 seconds is amortized over every new Like button I add, because I use like_url rather than link_to.

Also, often like buttons are used on something like stories, so you're only adding the code to one template, and now you have tons of like buttons all over.

I chose to go the other route which takes even longer. Load the XFBML version, and include the open graph meta tags.

I did the same thing that you did, wrapped it all in a like_url helper.

I don't know, man. The NHL, for example, has seen tons of engagement related to liking things. http://scobleizer.com/2010/04/22/facebook-ambition/

That and if my friends likes are valuable to me. Considering my having to ban Farmville and the like I'm pretty sure I don't want my social graph having anything to do with my search results.

Clicking on the link doesn't take you to a Facebook page -- it takes you to the TripAdvisor page, which is chock full of additional information. That page may also tell you if you have friends that have "liked" that property, which is not available via Google.

Full disclosure: I work at TripAdvisor and I led the Like project where we were a launch partner for Facebook.

I don't think that there's anything left that I can't talk about; I'm happy to answer question.

ps. And yeah, we're hiring to make this stuff more better.

Reply on news.ycombinator.com