Task: graphical scripting environment

I've been kicking this around a while, and I'm defining an explicit task. This would be an Wx GUI, a very simple one.

• The basic notion: a project, really just a group of files or directory structures. The basic window would allow you to drag files onto a fresh group, and save/load groups to a file list. Opened in a directory with such a file list, the tool would load it automatically.
• "Steps". Not sure where those are defined, but they're individual scripts to run against files or subsets of the files. Might want to toss in file filters to define subsets.
• Step lists. Macro actions consisting of multiple steps. Later with a more sophisticated language (loops and tests, essentially - probably just generic Perl of some kind, munged into a DSL).

That gives us the basic GUI functionality of Okapi, but in a highly configurable shell. This kind of thing could be wrapped around anything.

Visualizing log entries

Here's an utterly fascinating code article: parsing the SSH logs for cracking attempts and displaying them on an animated world map based on IP geolocation. Wow!

Towards a better API framework

From the server side of APIs, here's an interesting article on the past (Rails-style MVC frameworks) and the future (HATEOAS, essentially).

This deserves some thought. Maybe a plain function isn't the best way to organize code, and maybe OOP isn't quite there yet, either. Maybe something like an API is a valid level to understand code at.

I think I need sleep.

The Julia language

Julia is a fast math language. The benchmarks are really impressive, and the language itself seems quite readable.

Pretty LISP

Pretty-lisp.org is an alternative display/edit environment for LISP code. I don't care for it, but I like the general idea of a graphically oriented code editor implemented in JS on a Web page. It makes me think that Decl might usefully implement a local server for the purpose of editing Decl code, and the notion of putting different options in different places on a page has always been very attractive.

Update 2012-11-14: Darn.  I should have taken a screen shot.  The site is dead.

Google Event Tracking

Huh. Google Analytics also supports JavaScript event tracking. Fascinating.

Web text scraping

Good article on figuring out where the heck the content is on a given Web page. References this blog post elsewhere.

Graphics and their placement

A fascinating article talking about David Ogilvy and a study he commissioned about ad copy back in the misty days pre-Net, and how his results apply to content on the Web.

REST and HATEOAS

And speaking of API clients, I'm seeing a bunch of stuff lately about RESTful APIs.

• A good slide show by David Zülke
• A blog post by Steve Klabnik
• And another article by Klabnik

Kill Math

I can't remember whether I've blogged this before, but this kind of visualization stuff always turns me on. The Kill Math project is devoted to finding ways to model math interactively and intuitively, to solve problems without symbolic algebra. It's pretty neat stuff.

Task: Survey API clients on CPAN

I don't know what pathology this is, but I have the tendency, when approaching a technical task, of wanting to build the tools to build the tools to make the task easier [ob.xkcd 974]. A little while back I had the cute little idea of organizing my official tasks on blogs using the Blogger API. Well. CPAN has a lot of examples of API client access, and so I'd like to do a real survey of what's there, and compare and contrast.

If I can manage not to spend much time on tools to organize a survey, or research in general, I might manage to do something.

Poetry generation in Ruby

Slick little example Ruby program that takes today's headlines from a list of RSS feeds, then finds (quasi) rhyming couplets. The couplet metric is crude, but the results are still fun. It would be slick to do this with better couplet metrics using NLTK or something.

Berkely DB - NoSQL before it was cool

The title essentially summarizes the article. But it's a valid insight.

Bayes classification of HNN posts

I love this meta stuff. I also love the idea of something to read the news for me. Ha.

NYT on automatic book writing

The NYT has the story of a man who has self-published 200,000 books, essentially (obviously) all auto-generated from Web searches. That sounds pretty cool.

HTML in JavaScript

Now here's a really neat way to "compress" HTML going out on the wire, by rendering it in JavaScript. I like this notion a lot.

Stylometry

I remain unconvinced that stylometry can do more than identify "statistical identity trends", but it's still interesting work. [boing boing]

Getting experience with big data

A list of suggestions of how to get initial experience with large datasets.

The Great Web Framework Shootout

[github] More or less a Rosetta Stone for about twenty different Web frameworks in various languages.

Infunl API

Just as an example of how an API can be set up (and because I find logs interesting and an API about logs doubly so), here's Infunl's API. Infunl tracks user visits in the browser (apparently), kind of like Google access stats.