Twitter has become fairly central to the way I inform myself, whether it’s tweets from news organizations, links provided by the people I follow or the 140-character-at-a-time questions and observations that get thrown out.

And while there’s something neat about the way it is all — the silly and the important — mixed together, coming in at a flow of five, seven, 10 or more tweets every five minutes, I’m thinking there must be a better way, particularly as my list of sources grows.

I’m not smart enough to create this, but here’s an idea for a Twitter client:

1. Allow me to tag every individual Twitter source with freeform descriptors. I could tag CBC “news.” Ryan Sholin I could tag as “journalism,” “newmedia,” “freaking_great_ideas,” etc. Users could offer their Twitter feeds with baked-in, suggested descriptors, which I could accept or override.

2. Allow me to establish a filter for each of the tags and set the parameters for how often those feeds are checked. I may want to have the client pick up my news tweets every two or three minutes, tweets from friends every half-hour, and tweets from journalists every hour.

3. Build on-the-fly Twhirl-like panels for each of the categorized feeds. Rather than a single window onto the world, I would have my news panel (which would pop up ever five minutes, and which I could call on to review, re-read), my friends panel, etc.

4. For following live-tweeted events, allow me to set up temporary, time-limited feeds, which would expire after a set time.

A client like this would not only offer me a level of control over how I interact with Twitter, it would allow me to follow many more sources without feeling overwhelmed by the resulting in-flow. And, to top it all off, I would probably pay a reasonable amount of money for such convenience.

As I wrote, I’m not smart enough to do this, or even smart enough to know whether it can be done. But it would make Twitter even more valuable than it has already proved to be.