Archive for the ‘filering’ Category

Technorati watchlists need work

Friday, February 9th, 2007

Watchlists are an important part of the future of RSS feeds. Unfortunately the Technorati implementation is quite poor though in fairness so are all the other watchlist systems I have used.

An example will illustrate this best. For ego and practical use I have a Technorati watchlist with the following query; “Paul Watson” -pirate -captain -whale

Looks odd but there is a good reason for it. My name is quite common and in particular there is an infamous Paul Watson who receives many blog and news entries per day. Without the -pirate -captain and -whale I’d get a lot of irrelevant results.

Recently though Japanese and French bloggers have started blogging about Captain Paul Watson and my watchlist is once again becoming useless.

Technorati either needs to let me specify “English only sources” or translate my query as best it can (unlikely). They already have the former functionality with their main search engine but I haven’t seen the option for watchlists.

There are other problems with Technorati watchlists. One clever blog spams by subtley updating it’s feeds so fooling Technorati into marking the RSS item as new or updated when to my eyes there is nothing new about it. It is very annoying and I eventually unsubscribed from that watchlist. You also cannot edit a watchlist once you have created it. You have to recreate it with your query change, unsubscribe from the old one and subscribe to the new one. There is no blacklist system to block out sources you are not interested in. Nor is there a whitelist system which would be useful too. Duplicates are a problem too. Please, anyone working with RSS needs to implement deduping, it is a basic and fundamental requirement in my opinion. Clustering would be useful too, such as TechMeme does.

Programmers out there will tell me I can do all of this by refiltering my watchlist output using some other system or a bit of unix piping or Yahoo! Pipes etc. etc. No thanks. Watchlists need to be fast and easy, they need to be flexible and usable.

So, I’d really love Technorati to get going on improving its watchlist system (maybe spare a developer from the WTF team? I don’t use WTF.)