Comments on: from google app engine to google search sandbox http://thepoliticsofsystems.net/2008/06/02/from-google-app-engine-to-google-search-sandbox/ Thoughts on Power and Software Thu, 22 Jul 2010 16:48:54 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: bernhard http://thepoliticsofsystems.net/2008/06/02/from-google-app-engine-to-google-search-sandbox/comment-page-1/#comment-200 bernhard Mon, 02 Jun 2008 17:08:51 +0000 http://thepoliticsofsystems.net/2008/06/02/from-google-app-engine-to-google-search-sandbox/#comment-200 Hi Ryan, True, SearchMonkey is interesting and handy (played around with it a little bit some time ago) but it really doesn't go that much further than the old Google SOAP API (I don't think they ever went REST, no?) in the sense that you cannot bypass the company's ranking. These services just allow for application level access to the normal search interface. The SearchMonkey <a href="http://developer.yahoo.com/searchmonkey/smguide/sm_overview.html" rel="nofollow">overview</a> states: "SearchMonkey does NOT enable you to reorder results on a search page. You can use SearchMonkey to change your search result display so that they are more attractive and useful, but SearchMonkey does not change algorithmic rankings." Sure, you can use AND NOT operators to exclude certain terms or rerank results (I used that in a <a href="http://procspace.net" rel="nofollow">tool</a> I built for my thesis) but you cannot directly access the index of crawled sites without passing through a ranking layer. What would be an interesting compromise though would be to have access to thousands or tens of thousands of results with one API call (through REST or SOAP or whatever) so you could do some serious reranking. But neither Yahoo nor Google allow that :-(. For the Google Sandbox I have in mind, you'd have to run code on Google's servers and that's why I find the App platform interesting. But sure, currently it's definitely not about opening up search. Just wishful thinking... Hi Ryan,

True, SearchMonkey is interesting and handy (played around with it a little bit some time ago) but it really doesn’t go that much further than the old Google SOAP API (I don’t think they ever went REST, no?) in the sense that you cannot bypass the company’s ranking. These services just allow for application level access to the normal search interface. The SearchMonkey overview states:

“SearchMonkey does NOT enable you to reorder results on a search page. You can use SearchMonkey to change your search result display so that they are more attractive and useful, but SearchMonkey does not change algorithmic rankings.”

Sure, you can use AND NOT operators to exclude certain terms or rerank results (I used that in a tool I built for my thesis) but you cannot directly access the index of crawled sites without passing through a ranking layer. What would be an interesting compromise though would be to have access to thousands or tens of thousands of results with one API call (through REST or SOAP or whatever) so you could do some serious reranking. But neither Yahoo nor Google allow that :-( .

For the Google Sandbox I have in mind, you’d have to run code on Google’s servers and that’s why I find the App platform interesting. But sure, currently it’s definitely not about opening up search. Just wishful thinking…

]]> By: Ryan Shaw http://thepoliticsofsystems.net/2008/06/02/from-google-app-engine-to-google-search-sandbox/comment-page-1/#comment-199 Ryan Shaw Mon, 02 Jun 2008 15:07:10 +0000 http://thepoliticsofsystems.net/2008/06/02/from-google-app-engine-to-google-search-sandbox/#comment-199 While well short of full access to the index, Yahoo's new <a href="http://developer.yahoo.com/searchmonkey/" rel="nofollow">SearchMonkey</a> platform is a step toward the kind of open search platform you describe. Google's App Engine, on the other hand, appears to me to be geared toward easily scaling up (and wedding to the Google platform) standard database-driven websites, not opening up search. In fact, Google's trajectory has been to gradually provide *less* access to search data, as for example when they replaced their REST search API with a less flexible but more easily controlled (by Google) Javascript API. (The REST API is now back, but with very restrictive terms of service that essentially disallow using it for anything other than displaying Google search results on a web page.) Yahoo perceives a potential strategic advantage in building an open platform where Google refuses to be fully open; time will tell if they are correct or not. While well short of full access to the index, Yahoo’s new SearchMonkey platform is a step toward the kind of open search platform you describe. Google’s App Engine, on the other hand, appears to me to be geared toward easily scaling up (and wedding to the Google platform) standard database-driven websites, not opening up search. In fact, Google’s trajectory has been to gradually provide *less* access to search data, as for example when they replaced their REST search API with a less flexible but more easily controlled (by Google) Javascript API. (The REST API is now back, but with very restrictive terms of service that essentially disallow using it for anything other than displaying Google search results on a web page.) Yahoo perceives a potential strategic advantage in building an open platform where Google refuses to be fully open; time will tell if they are correct or not.

]]>