About a month ago, Bill Burnham posted about what he dubbed persistent search. Basically, it is the continuous searching of content based on information given by the user. So for example, if I wanted any data about the Cubs to be sent to me, that would be a persistent search.
There are current companies who offer this type of service but many don't include all possible content as they are focused just on blogs and other data found in RSS files.
One problem that the post didn't address is the issue of displaying this new found data to the appropriate device. If I'm out and about, I don't want to get a constant stream of alerts to my cell phone especially if they are just news articles or blog posts. Instead, I want to be able to say that I only want the latest score on my phone and perhaps an hourly indication of how much other data has been found. This isn't an easy problem to figure out I don't think since having the user indicate which device they are currently using is doubtful to happen. Instead, it might just need to be a preference and some overlapping of sent data will be probably occur.
The other elephant in the room is what happens as the number of users increase. This constant stream of data is going to need to be indexed as well as alert users on this data. How would this publish-subscribe relationship best be architected? Perhaps when a user indicates the search criteria for the data, a subscription for that criteria will be created or if one already exists, the user will be added to the correct queue.
If the infrastructure is created, a company could definitely make an impact offering persistent search especially if they focused on indexing data as quickly as it was available.
Technorati Tags:
search, persistent+search