Google's hanging algorithm

Hanging algorithm? What the heck does that mean?

Well, I think Google has left a big hole in their searching algorithms for email and desktop searches. Now I'm a BIG fan of both services from Google. I can honestly say that they have changed my computing habits and make my day more efficient. However, there is a flaw.

One obvious flaw is the inability to link separate "conversations" in Gmail. If you start two different threads of emails with different subjects, Gmail will not let you combine them. I hope there will be an option for this in the future. But even if that was the case, it may be annoying to manually "join" all of your conversations.

So what happens if you have a few emails (or Word Documents) talking about your trip to Cancun, the people you are going with, the time you are leaving, the hotel you are staying at, and a short itinerary. Not all of these emails or documents may have the word "Cancun" in them. So how do you search all of them? Well, some objects have the name of people, some objects have travel dates, and other objects may simply be referenced in multiple threads.

Wouldn't you like the capability to search for the word "Cancun" and have Gmail or Google Desktop Search (GDS) know that your 15 emails with the word "Cancun" may also have:
- Some of he same 10 people in common
- some of the same travel dates
- some of the same items mentioned (suntan lotion, rum, hot chics, etc.)

Google should create a whole algorithm based on "relevance " via a networked mesh of terms. (Think of Friendster and searching for your friend "Todd" and having it return Todd and all of his friends.) Maybe this hotel itinerary email doesn't actually mention the word Cancun, but it says Mexico and has some travel dates in common with your Airline itinerary to Cancun. All of these objects could come up in your search. No need for you to assign categories to emails. An entire "relevance" thread can be created, bringing in much more information then was previously searchable. Without this suggested method, a human is required to do these "network mesh" calculations and guess what other words should be searched to find relevant information.