No doubt search has come a long way in the last decade or so. But frankly, it's still hit or miss, and pretty clumsy. Really finding what you are searching for can be a grueling task at times. Sometimes we get lucky. Things like Ask's Smart Answers, Google's Onebox, MSN's Instant Answers, and Yahoo's Shortcuts can be really useful, but are still limited to only what the search engines have deemed worthy of these types of answers.
I've been reading about Powerset lately which promises natural language search. An article I read today uses the following examples to explain how Powerset plans to outdo Google:
Powerset is notable because it says it can improve on Googles search engine by understanding the meaning between words. Take, for example, the phrase, Who did Dick Cheney shoot?"? Powerset will give you results with references to Harry Whittington, Cheneys hunting partner who Cheney accidentally shot. However, if you type in Who shot Dick Cheney?"? it will give you a response Sorry, no results."? Google, on the other hand, cant distinguish between these two phrases, and gives very similar results for both.
Similarly, if you type in Who acquired IBM?,"? Google will give you lots of results about companies that IBM acquired, even though thats not what you asked because Google cant understand that the difference between an object and subject. Powerset, on the other hand, will give results of the various companies that acquired IBM units, including Lenovo, and AT&T which is a better answer to the original question.
Now, see, this is the kind of thing that would make search a much improved product (assuming it works as described). And I don't know about you, but I'm more than ready to spend less time searching, and more time finding and using what I've searched for. But even this level of search isn't the be-all, end-all of what search could be. Let me give you an example of a question I've long wanted answered, but have never been able to find.
I would love to find the PERFECT spot on earth - in terms of climate, that is. Perfect for me, of course, as the perfect climate would likely be different for each person. I would love to be able to type the following into a search engine:
Find the top 10 locales on earth that have an average high temperature of 75 degrees Farenheit, an average low temperature of 65 degrees Farenheit, an average of 3 rain showers per week, and almost never has extreme weather of any kind.
And I'd like to be presented with a list of those locales, in table form, showing each of the parameters I've searched for.
Sounds impossible? Why? Surely all of that information exists on the web. Why couldn't a plain old ordinary general search engine be able to dig up the information and present it to me? Too time-consuming for the engine? Maybe. If so, it could let me know that this would not be an Instant Answer, but that it would be happy to email me when it has the information ready.
This is only one of many types of searches that users might want but cannot currently get from the main search engines. Perhaps there is a niche search engine out there dedicated to exactly that type of query, but if there is, I wouldn't know where to find it. And while niche search engines can be very useful, we would need to have bookmarks of thousands (millions?) of such sites at our fingertips.
So often, we praise the likes of Google for making search such a wonderful thing, but in reality, it is still so rudimentary that it is sometimes just frustrating. I mean, really, this is so cliche, but I still have to say it. If we could send a man to the moon decades ago, why on earth can't we have search that goes far beyond the Googles, Yahoos, MSNs, and Asks of today?
Do you think I will get my question answered by a search engine in my lifetime?
You’re so right, Donna. I’ve always felt Search would be improved by a more obvious interaction between the search engine and the searcher. That would require a more obvious way of homing in on the relevant area by back and forth questioning. With AJAX, it would be a snip to do or you could even do it in Flash.
Instead we get only the packaging improved as in the MSN/Live Ms. Dewey. Perhaps it’s to distract us from the relevance or lack thereof of the search results.
Actually, Barry, I have to give credit to http://www.msdewey.com . I think it’s hysterical. Highly entertaining. Doesn’t improve search results, but it sure makes the process less frustrating.
Define “top 10”, “rain shower”, “almost never”, “extreme weather”, “of any kind”; does it need to have a land mass as well? 🙂
It’s really not that easy. It’s definitely easier to fly to the moon and back.
Recognition and processing of natural language queries is hard to do, even when you have a well-structured database with a few fields and perhaps 1000 entries. Expand that to the web, where data does not come in fields, content is partially copied, on-page information can be irrelevant (ads, news feeds, etc) and where you work on billions of data-sets at the same time … and expect a fully relevant answer within a second or two (however, as you mention, a tough question can take time, if the answer is guaranteed to be correct).
The other problem is that not even humans can process natural queries correctly every time. Misunderstandings are common, especially when one side doesn’t really know exactly what they are looking for in the first place :-). The human solution to that is to “discuss” until the query has been sufficiently refined… or discarded as being impossible.
Never said it would be easy, JohnMoo, but neither is rocket science. Easy or not, I still want it. 🙂