The IRC bot project has become stable if not a little useful. There is much to do w/r/t writing new behaviors but the framework appears to be mostly in place. Hooray for that! Most days I’m in front of a computer you can find me on Freenode’s #fezchat channel. I tend to run the bot under the nic citadelOfJerrys.
My new project idea stems from an article I read recently (but cannot find now) regarding the one-way direction of hyperlinking. The thesis being that content creators can hyperlink but content consumers cannot, short of becoming content creators themselves. Not everyone has the means to do this; nor the inclination.
So what does consumer hyperlinking look like? I’m not certain but it did occur to me that I’d like some kind of automated journaling of my web browsing. Some kind of browser plugin that saves off the DOM for every page I hit over the course of the day with a little bit of annotation/metadata rolled into each record.
From there I’d like to build some kind of TFIDF index for searching. This could also be leveraged to generate topical heat maps. In this way, I can get a better idea of what my consumer hyperlink following behavior looks like. At least from a research perspective, I could publish my own richly linked topical webs. I’m sure there is a more broadly accepted term for a collection of links in this manner.
And this is kind of what I anticipate a consumer-based hyperlinking system might look like. In short, a cross between the Yahoo! curation model and Google’s Page Rank model. All of this lands on some open forum where consumers could publish personal webs and users could subscribe to webs or publishers they find useful/interesting/what-have-you.
That gets really pie-in-the-sky looking but I think the exercise of writing the browser plugin and data mining/presentation interface would be a fun project.
At any rate, as part of the FircBot exercise, I’ve built out a build system that publishes packages to a web repository. It checks these boxes:
- Leverages Jenkins v2
- Integrates with GitHub
- Publishes to Bintray
- Includes an automated versioning paradigm
- Discerns branch/continuous builds from master/published builds
- Makes extensive use of Jenkinsfiles in each branch/project
My only remaining wishlist item is change-detection and automated builds based on that. In the grand scheme of things, this isn’t a huge need.
Anyway, that’s what I’ve been up to as of late.