Wordbot

July 3, 2008 5:03 PM

Wordbot
Wordbot is where you go to check on drafts that you're about to post, get contextual information from across the web on the topics you've written about, or quickly get considerable extended information about text stored at a URL. In short, it's a set of power tools for the readers and writers of the Internet.

This is the first public announcement of this project to the world. It's in "early alpha" state, whatever that means, but I've achieved enough stability that I'm comfortable announcing it to this discerning community of Internet dwelling nerds. The software may be a bit touchy (especially the term extractor) and it certainly is young, but it's running on a dedicated server and I've tested it thoroughly on my own, so it's ready for the wild.

Additionally, I've provided much of the PHP library which powers Wordbot as a download. It's licensed under the MIT license.

The goal here is to provide as much help as possible in the content research and creation process to readers and writers, especially editors and reporters. With that in mind, where should this go next? What's useful here now?

Writers: what could I add or change to make this a must-have part of your Internet existence rather than just a novelty?

posted by netaustin (4 comments total)

This is a good idea (and I know it's a very first version) but right now, it doesn't work very well at pulling up stuff that is actually useful to a journalist or blogger. I just tried with a post I did recently for Huffington about a new bill to ban child abuse in boot camps. Google would have more easily found other coverage of the bill because this highlighted general words, not the phrases that are most important. It didn't even highlight the bill number, which is the most important term to search if you were looking for other coverage.

I have no idea if this can be made to work that way...

Would love to see where this goes though...

Also, the URL lookup should include whois and other harder to find information that isn't just from the site itself. If it does this now, I didn't see it.
posted by Maias at 9:43 AM on July 6, 2008

Since I believe the interface is the product, my primary suggestion is jump right to the results. Given it is a research tool, I expect results, not highlighted content. In fact, my gut argues (yes it does) to either remove the view altogether or make it a tiny and unimportant widget. It's interesting for the computer scientist in me, but less useful for the writer in me. I want results and to be able to manipulate them.

This leads to my next suggestion: the black/red text widget is a powerful idea, though the tag cloud is probably not the strongest IA pattern, let alone it is for all intents and purposes useless on the stats page -- it doesn't do anything. In general, I find tag clouds only useful when there is enough deviation between tags to pop. My gut (again) suggests a weighted list (top down) would be more useful. Right now, I have to think and calculate which tags are weighted what. A vertical list with a sparkline/background bar graphic would lend itself to better scanability and also give some transparency why the first item is flagged on (red). I could be persuaded that an alphabetical cloud is useful, but haven't found a scenario while I've been using it.

I've been writing this response in the Wordbot textarea and researching it. I discovered I use "yes" more than I should.

More to come on the results and their utility.
posted by pedantic at 10:21 PM on July 8, 2008

Love it. Am actually using it, occasionally. Thanks!
posted by imperium at 4:31 AM on July 31, 2008

Just FYI, the robot is superimposed right on top of the interface boxes, in my browser. (IE7)
posted by beagle at 11:11 AM on August 1, 2008