|
WebWord Weblog Posting Posting Date: March 21, 2003 37BetterGoogle (37Signals) -- "Our idea for 37BetterGoogle was to go beyond Google’s standard search and offer alternatives based on the words you originally entered." (MadMan comments: How does knowing the number of results tell you anything about whether those results are relevant? Is 7821 results better than 5479 results?)
Reader Comments...
I see your point. Only you know what you are searching for, but maybe this feature can help those that are not exactly sure how to get what they are looking for. I don't know if the number of results will make any difference to your decision, but the variation list may assist you in tweaking what it is you are after. Posted by: JB on March 21, 2003 12:49 PM
No, but 250 results are better than 2 and definitely better than nothing. Google's spell suggestion feature catches my typos all the time, and I imagine it would help even more if I were a poor speller as well as a poor typist. A stemming suggestion algorithm could similarily 'fix' my poor choice of an infrequent variant phrasing for a more common, more useful one. Posted by: need2know on March 22, 2003 12:41 AM
The idea behind 37BetterGoogle is that it's very difficult to know the best way to search for something. How can you know if "african coffee trade" is a better search than "africa's coffee trading" when you're doing research? So, we're proposing that Google add an optional feature that will suggest alternate searches based on slight modifications to your original search. At the very least it provides suggested alternate searches if you couldn't find what you were looking for. And, of course, this is just an idea. Posted by: Jason Fried on March 22, 2003 09:56 AM
I like it. Automatic Word Stemming = Dumb Lack of Precision = Bad
What Google needs is a Buzzwordometer / Buzzwhack ranking = Modern Fogg Index Posted by: (the other)JS on March 23, 2003 12:20 PM
I wish I could remember where I read this, but I thought I had heard that Google had something like this in the works? I think I was reading an article on their internal system for gathering feedback from employees for new feature development. It was talking about how they recognize employee contributions even when they didn't have the technology/time to build out a really great idea. It seems like a logical outgrowth of the user-assistant features they have now. It's funny, but it works so well as a spell-checker and thesaurus that I probably only actively search on about 40% of my visits in a day. Posted by: Lydia on March 24, 2003 03:42 PM
I understand the concept, I appreciate the concept, but I don't like the execution. In each of these examples, one has to choose within a specific search expression where they only vary in their grammatic construction. Think about what would happen when you click on any of the sublists which are (hypothetically) provided — the user would go to the first-ranked google result in that specific expression's subset. Whether or not this subset has the most listings is somewhat irrelevant. To me, a far more preferable execution would be for google to construct a search string based on "OR" strings; ranks set up by the interlinkages of any of those terms. Users would not have to choose between subsets... Or, for those who think the combinations could occasionally be so important that it should not be automated, offer check-boxes of related phrases to include in the search terms. (Now that I think about it, google labs was testing an algorithm which would suggest related terms, which could ostensibly be used to round out a search.) Posted by: Frank on March 24, 2003 03:47 PM
Home | Moving WebWord | Cool Books | Hot Web Sites
URL: http://webword.com/weblog/ ©1998-2005 by WebWord.com. All rights reserved. |