March 29, 2012 – 11:38 am
Okay, so the new prototype is fun and all, but what about the main service? It needed some love, too, right? It’s okay, we know — lots of yucky “Server Error” messages.
We reconfigured things on the web server a while back, and yes, we admit, we didn’t get everything set up the way we should have. But we’ve finally gotten around to taking a peek under the hood, and we think we’ve twiddled the right knobs to get things running more smoothly. In fact, we’re so sure that we’ve made things better, that we’ve raised the combined friends/followers limit from 3000 up to 8000.
If we can just get Geordi to quit hooking Data’s positronic brain into the ship’s computer, maybe we can keep this bucket of bolts flying right!
Update: We also changed the sorting of the lists so that users are listed in a more natural order. If you are looking for a particular friend or follower, it should be easier to find them, now.
March 27, 2012 – 10:44 am
You thought we were dead, didn’t you? Just another abandoned ghost service, floating along in the Internet like a mysteriously vacant boat in the Bermuda Triangle. Nope! Still here!
If you haven’t heard already, there’s a new prototype in the works (no, really — you can check out the Twitual prototype). It is very bare-bones at the moment (duh, it’s a prototype), but you should get the general idea of where things are going. My next steps are to add in functionality to let you drill into people’s details, add some real UI/design, improve handling of errors, and restructure the underlying code a bit.
Things might have been further along before now, except that I had been running into some technical problems. I won’t bore you with the details here, but I posted about it on my personal blog: Dealing with UTF in node.js
March 13, 2009 – 10:41 am
We have removed the cap on the total number of friends and followers, so you can now try to analyze any Twitter user on Twitual. We will warn you, however, that with very large friend + follower counts, the process of fetching all of the information can be slow. To give you a general idea, some recent tests were taking about 3 minutes to analyze a user with 10,000 combined friends and followers. Obviously, this makes it pretty impractical to analyze highly networked users like @wilw or @mashable (who each have over 200,000 followers).
The reason it takes so long is that the Twitter API only allows us to fetch 100 entries at a time. There are API calls to return all of the friend/follower IDs for a user at once, but they only return numeric IDs, not the screen name, which is far more useful for display. So we either need to find a way to fetch multiple batches of information in parallel, or we need to mirror the entire database of Twitter IDs and screen_names on our own server. Or maybe both. Even better would be a new API call to fetch all the screen names, but that’s not our call.
And, we realize that there is currently no feedback in the browser while Twitual is doing its thing. We’re going to be working on that, and hopefully when we’re done, we’ll have some spiffy whiz-bang progress bars to let you know what’s going on, and how close we are to being done with your request. Just don’t expect it to happen overnight — we’ll have to take our current, simple, no-nonsense code and transmogrify it into a bunch of complex, full-of-nonsense code. Those kinds of things take time!
In the meantime, we still think Twitual is a pretty nifty tool, and we hope you do, too. If you agree, keep spreading the word about us. Let us feel the love, we’re needy that way!
Over the past few days, a couple of other sites noticed Twitual:
Have you seen us mentioned anywhere else? Let us know!