Amethyst Reese

For the last couple years, this site has been managed and updated using the open source static site generator called Poole. It is an excellent and simple system, comprised of two Python files — the Poole source code and a site-specific macros.py — a simple page template, input documents and static content. This is more than enough for a small static website, but more complicated sites, like blogs, require quite a few macros in order to generate things like archives, tag pages, or even RSS feeds. Over the 18 months that I worked with Poole, my macro file had gotten a bit disorganization, and the limitation of working from only a single template file was starting to strain on what I could do with it.

Now keep in mind, I still like many of the features that drew me to Poole in the first place, such as the use of a document-centric build process, with YAML headers for defining page metadata rather than an inline, proprietary format used by projects like Jekyll. I also enjoy the simplicity of the content representations, but the lack of an extensible content pipeline is my biggest complaint. Supporting formats other than Markdown requires adding yet another hook to the macro file; generating archives and tag pages required hacking in a multi-phase build using os.exec(); and inlining another page’s content post-render was just not possible, resulting in Markdown named-link collisions when rendering multiple posts to a single page.

I wanted something new. So, for the past few months, I’ve been working on Nib.

The resulting design is based heavily on the concepts of Poole, but built around a primary goal of producing a proper content pipeline that is simultaneously aware of the differentiation between resources and documents, and defines multiple stages where plugins can hook into the process and alter or generate page contents at build time. Indeed, most of the actual functionality of Nib is contained within a handful of plugins, while the main module merely defines a framework for the content pipeline.

In effect, adding support for more content or resource formats should be as simple as adding a new plugin attached to the appropriate file extensions. Advanced content manipulation, generating “virtual pages”, and aggregating pages or documents into multiple locations are all possible as well.

The Markdown plugin is 14 lines of code; the LessCSS plugin is 13; even the blog plugin — which generates the archive pages, tag listings, and Atom feed — takes only 86 lines to do a better job than the old Poole macros that required double the effort.

Today, I’ve thrown the switch. While it looks nearly identical to the old site generated by Poole, the page you are now reading has officially been generated by Nib. The news feed sees the biggest makeover, having changed formats from RSS to Atom, and now offers full content posts. Link-style posts, a format that I’ve been wanting to experiment with, are also a new option. Both features would have required more effort on Poole than I was interested in expending, but the new architecture has allowed me to indulge myself.

Nib is still in an extremely early, and unstable, phase though. It works for the needs of my site, and does come with some basic documentation and a sample site to start from, but it’s far from complete. Near term goals include adding support for an intelligent menu, as well as support for more content and resource formats, like reStructuredText or SASS. Contributions are always welcome though, even at this early stage. Nib is liberally licensed, and I would love to hear feedback from anyone trying to use it. Hopefully it will be useful for someone other than myself.

Asked of me on Hacker News:

What do you prefer about Android?

I prefer Android because it allows apps to do more things for the user, and allows them to better integrate with the system as a whole.

I can replace the on screen keyboard with one that has a full five-row keyboard for times when I SSH into a machine. On a similar vein, when I SSH into a machine, I can actually leave the SSH session running in the background while I switch to another app, without fearing that the OS is about to kill my SSH app while I’m looking something up or responding to a text message. I can also leave my IRC client running in the background without it constantly needing to reconnect when I switch back to it.

Intents in Android, especially in combination with the global Share mechanism, allow any app to receive arbitrary data from any other app, meaning apps don’t need to know about specific apps or services in order to integrate with them. Clicking on a URL allows you to choose which browser (or set a default) to open the link in, allowing you to use alternate browsers (or alternate email clients, SMS apps, dialers, etc); tapping Share in the browser allows you to send the current URL to any application that can receive a URL, making apps like Instapaper, Pinboard, and 3rd party Twitter clients have the same capabilities as first party applications.

Sideloading apps means I’m not limited to installing programs from the Android Market/Play Store, and can do things like buy apps directly from the Humble Indie Bundle and install them on my own.

Proper background service support, and allowing apps to affect things outside their sandbox, lets me run programs like Locale that can monitor the phone’s status, location, etc, and modify the phone’s settings automatically based on a set of conditions that I’ve pre-arranged. My phone automatically silences itself at night time and while I’m physically at work, turns on my Wifi when I’m at home or work while defaulting it off when I’m out and about, and more.

That’s just some of the reasons I like Android better than iOS.

The past two months have been intense. At the beginning of May, I was let go from BioWare and Electronic Arts in a round of budget cuts for my studio. Since then, I’ve played far too many video games (and beaten multiple titles), slogged through innumerable interviews, redesigned my website from the ground up, and celebrated my seventh wedding anniversary in some of the best summer weather I’ve ever experienced.

Oh, and I got a job at Facebook.

On May 3rd, BioWare informed me that my position in the studio was being eliminated. My terrible luck had struck again. But this time was different; within a few days, a recruiter from Facebook found this website and my open source work, reached out to me, and encouraged me to apply for an engineering role. Four weeks later, after phone screens and an on-site interview, I was offered a position as a Production Engineer. My luck had not only reversed, it had gone off the scale in the opposite direction.

Alongside Facebook, all my other leads paled in comparison. I wanted to stay in gaming, but here was the opportunity of a lifetime, to join one of the biggest and best engineering companies in the world, and the chance to learn from some of the brightest minds in the industry. The position is similar to the role I had at BioWare: writing and maintaining software to manage and automate the vast array of infrastructure that supports the front-end applications and engineers. It was an offer I couldn’t refuse.

Monday was my orientation, and I was immediately impressed by the level of passion, vision and dedication from the entire team. Boz discussed the culture of “hack”, and told a story about building a loft in a war room when it couldn’t fit everyone in just two dimensions. Chris Cox gave an inspiring talk about where Facebook is heading, and I got a surreal sense of wonder when he mentioned our long term goals or the level of impact we have on global society and the way people communicate. It feels amazing to be part of a company that truly wants to change the world and improve the lives of hundreds of millions of people.

Even the training process embodies those same ideals. The first few weeks, every engineer goes through “bootcamp”, attending learning and development sessions, sitting among other bootcamp engineers, and working with veteran teammates on real tasks to get familiar with the infrastructure, tools and codebase. Everyone is given the chance to progress at their own pace, with as much or as little help as needed. I’ve already learned so much in a few days, and I’ve also had the chance to share my own knowledge and experience with my fellow newcomers. And yet there’s still so much left to cover.

Wednesday was my first day at a desk. Taped to the monitor was a sheet of paper printed with bold, red lettering.

What would you do if you weren’t afraid?

I don’t have an answer for that yet, but I will soon.

I’ve been spending time migrating my server from Ubuntu 10.04 to Arch Linux, and in the process I thought very deeply about every PHP application installed on the old server. Five out of the six PHP apps were only there to support the one that really stood on its own: MantisBT, my long-standing bug tracker of choice.

Now, I’ve been a core developer, and de facto release manager, for the project for many years — since I was still in university and getting paid by my then-employer to contribute features and plugins that they wanted to use for their engineering team. Those plugins, like Source Integration, wouldn’t be free without me fighting to license and release them for the community. And I wouldn’t be where I am today without the experience and help I received in turn from the very same community.

I know I haven’t been as involved in the project these days as I would like to be; there is an endless list of features and improvements to make to both the core system and the array of plugins I’ve created for it. Some great community members have stepped up and filled my place at times, while I have at least tried to stay active on the mailing lists and in the IRC channels. I’ve still guided and cut the last couple releases, but I haven’t played a part in shaping the future of the project.

There are multiple competing visions charting new paths for the aging project, with conflicting goals and revision histories, and it really needs a stronger leader to take the reins and guide the project to its next milestones. I’m unfortunately not the person to fill this role, for many reasons. Maybe a few years ago it would have been better timing.

Lately, I’ve come to the realization that I can no longer bring myself to work with PHP for personal projects. I don’t like the syntax, I don’t like where the language is heading, I don’t like how much memory and CPU it requires to run on a web server, and I just spend the whole time wishing I was writing Python code instead.

This blog hasn’t run on PHP or a database for just over a year now, and with my MantisBT install being used mainly for projects I don’t have the time or will to work on, it just seems to be dead weight. Github can serve my needs well enough for the few remaining projects I work on, and without needing a complicated setup on my end. Turning off MantisBT means I no longer need MySQL, PHPMyAdmin, APC, or even mod_php at all. The remaining apps can easily be replaced with external services.

So basically, this is me announcing what I’ve already been practicing for many months now: I will no longer be a developer for Mantis Bug Tracker, but I will remain involved as a mentor for other core developers, or for those seeking some advice on my plugins or creating their own. I won’t be maintaining any of my plugins, but I will look at and accept pull requests until someone else wants to step up as maintainer. I will be removing MantisBT from my site, but will keep a database dump in case I ever need to reference it in the future.

This is not me withdrawing from open source; I have many other projects that I’ve been working on, most of which are written in Python or C++. I find them more enjoyable to deal with, and most importantly, they allow me to break out of the realm of writing web applications. IRC, as old as it is, has been my point of intrigue lately, and is at the core of my current “pet” project.

Regardless of language, you can still find me on Github, where all my toys are available for the public to point and laugh at. And as always, I will answer questions on Freenode as “amyreese”, or via email, although there may sometimes be a long delay before I can reply.

Thank you to everyone who’s contributed to MantisBT or its plugins, and thank you to everyone who helped me on the way to where I’m going.

Hey man,

Just wanted to say thanks for the ZNC Push Module you made. I’m using it with Notify My Android and it works great! This should be a standard module in ZNC imo. Stellar work my friend.

Best,
~Lucien

It makes me smile every time I hear from someone using my software, knowing that I can make an impact on peoples daily lives, no matter how small it may be. Thank you to everyone who has taken the time to thank me for my work.