devlog

For the last year, I had an almost permanent migraine, triggered by light sensitivity. I pushed myself through on ibuprofen and Chinese herbal remedies; recently even those stopped working. I had to stop looking at the laptop screen for a few days; the migraine went away, and I was able to go out in brightly-lit places again.

When I start looking at the screen again, the edge of the migraine kicks in after about 5 minutes. So I’m trying to drastically reduce my online commitments - unsubscribing or switching to digest mode on about 100 mailing lists - reduce the load of volunteer work i’ve been doing for OSGeo and OKFN - prioritise all my actions and communications according to time-sensitive requirements for other people.

I’m somehow enjoying being forced to filter information more. With less input, i have a lot more thinking time. Obliged to work with pen and paper, I’m having fun ideas about paper-based backup systems. Obliged to meet people over VOIP rather than IRC, my voice (quiet, hesitant) is strengthening, and I pick up more of what people need, from their tone. I’ve been having fun talking with Der T about voice-based systems.
But I need to change my work practises completely, and find paper-and-voice based ways of doing what I’m good at. I make a living out of coding, systems analysis, and writing. I can’t do any of this at the screen without causing myself the physical pain which says “now back away from the keyboard”.

I’m finding that it’s both possible and fun to do software systems analysis on paper. The last couple of days I’ve been walking through, on paper, the blabla codebase with T. I have the latest subversion branch printed out, read it through and sorted it into four piles. I call the four piles “Please Make It Stop”, “Laugh or Cry First?”, “So This Is Documentation?” and “I Thought These Things Might Be An Application”. Then we sat down and reinvented CRC cards, where Roles, Goals and Interfaces happen. If it’s not on a postcard, it’s not in the system; if it can’t fit on a postcard, it’s trying to do too much.’

Next comes some literate testing; right before i burned my eyes, i was going through a process like this for the Plex, the software aspect of the 24 days project and the ongoing ESP initiative. I wrote a set of pseudotests to be fleshed out by application developers. These expressed the models in and interfaces of the system.

I’m enjoying casual involvement with the blabla project particularly, because it is a voice collaboration and archiving environment. A lot of the current implementation is crust, but the core design is simple and beautiful. It’s in that semi-stalled state that a lot of software projects get to where they work well enough for the core developers to base work on, but are hard to get into for a newcomer and also suffering from the “second system effect”. The accretion of new features and utilities over an original rapid prototype. Now is a good time to step back, clean up, take some pride and focus on opening up contribution paths to new people. The problem is that the core developers are busy doing commercial work to keep themselves alive. Some kind of sponsorship for release candidature is needed.

What else can i do with myself? I find that writing isn’t as bad for my eyes as reading, so increasingly the internet is becoming for me a write-only medium. I’m interested in finding projects where i can print out a lot of material, do paper analysis notes and perhaps compose summaries and critiques by voice and pay if necessary to get these things transcribed. Late last year I started reading the OGC specifications for the godawful CSW/ebRIM set of standards for expressing metadata about geodata and carrying it around, and making a set of implementation notes. I did this in a fit of pique after being accused of not having a right to slag it off if i wasn’t familiar with all the OGC material, which is actually quite reasonable.

This is the sort of thing i would like to find a syndicated way to work on. A friend expressed interest in contributing support to a research study on geospatial metadata standards and appropriate software, and next-generation / distributed ways of passing data around and re-using it. One would run such a project on a completely open source basis and be able to do it as “sponsorware”, somewhat like copycan works.

This is written partly as an explanation of my status - those people who see a lot of me online will have been aware of my popping up and wailing about screentime aversion for several weeks. If i’m not writing back to your email, it’s because i need to focus my time online as much as i can, and i generally have way too much communication to deal with and mutual promises to others that i am trying to restructure and cut down on. This is also a kind of plea for interesting offers - my situation is a bit unusual, and self-support feels like a whole new game for me now.