(besides items mentioned in my "What needs doing" pages)
When I first started volunteering, I found very little in the way of technical documentation. Where were the servers located? What function did each serve? How did different pieces of software interact? What home-grown software did we have, what did it do, and what was the right way to use it? What were the known bugs? Why was something done in a weird way — was that a clever idea that solves some non-obvious problem, or were we just following a routine started by mistake long ago?
I explored, I asked questions, and, rather than just memorizing, I took notes. Over time, this has grown into fairly voluminous documentation. It remains far from comprehensive. But new administrators can now sometimes be pointed at a section in the documentation in which to find a good explanation in answer to their question. More often, existing administrators (especially I), use it as a quick reference for facts not fully remembered.
We have several new technical volunteers. I have been spending time with each of them, helping them to grow and to learn.
Even before I got the root password, I had seen evidence that TFN's servers had suffered from neglect and needed much work. Over the past 2 years, I worked hard on cleaning up and on re-engaging technical people who had drifted away. There is still an awful lot to be done. But we have made progress, both in terms of server status and in terms of volunteer engagement.
Over the past year, much has been accomplished. We migrated from tower (desk-side) class equipment to rack-mounted hardware. This means that we consume less space in a datacenter. Thus, when we make arrangements that require paying rent to the datacenter owner, our cost is lower. When we aren't paying full price, being less of a burden improves the relationship with our benefactors.
We migrated one of our servers from FreeBSD to Debian GNU/Linux, and have been using Debian and its variants on all new equipment. We intend to switch all remaining servers to Debian. Though FreeBSD has its merits, GNU/Linux is also good and it is far more popular. Moving towards it has made it easier to recruit volunteers that already have experience with this platform.
We also did 3 major physical moves, plus many lesser operations over the past year. The Friday 13th move, last March, was terrible. We had to make the best of a bad situation, to deliver on a deadline that we couldn't really meet. But the other moves, though they all had unexpected problems (all moves encounter some problem), went much better.
I do wonder whether some of those physical moves might have been avoided. That would not only have saved us the problems we experienced in the moves, but also the weeks of advance planning and preparation — time which might have been spent working on other urgent problems.
Over the past year, I lobbied hard for increased advertising. We did get some updated posters last May. I put a box of them outside the office door. Come take some to put up in your neighbourhood. I also put the PDF on our website, so you can download it and print copies at home. Here's the PDF's URL again:
The information on our website sucks right? The instructions for becoming a member are way too long, the support forums don't work, the listed email addresses bounce, the help files are a decade old, there are broken links and contradictory information all over.
But it is slightly better than it was a year ago: I have been fixing pieces of it little by little. It is slow work because I must meet with and get specifications from the executive director, then draft new text, then wait for the inevitable amendments, then start again. It usually takes months. It is a very demoralizing process. But I persevere. I focused mostly on reducing the confusion in the membership information page.
Over the summer, I tackled two projects for which the executive director had expressed a need. The first was finding out our advertising success rate, for different modes of advertising. The other was finding out how many members we have, how many are joining or lapsing, and the popularity and fiscal weight of the different services to TFN's bottom line.
The most important discovery was that usership numbers are way down. Last spring, the number of lapsing members per month was greater than the number of new joining members. It had been thus for years.
Thankfully, by last August, between the posters and the small website improvements, the membership decline stopped: the number of paying accounts from newly joining members, per month, now equals the number of paying accounts which are lapsing. It's a start. At least our leadership now knows these numbers.
back to: main
Copyright (C) 2009 Iain Calder. All rights reserved.