Thursday, January 08, 2009

SchoolTool 2009 Plan

The recent news about OLPC has reminded me that I haven't posted the 2009 development plan for SchoolTool. I'm happy to say Mark is willing and able to continue funding the project, which is a big relief as we're just starting to get some real momentum...

Quick 2008 Recap

SchoolTool core -- We successfully released a beta for SchoolTool 1.0 on schedule in October, our primary goal for the year. It includes the core components of a student information system, plus calendaring. This has had the desired effect; we've seen a significant increase in queries, feedback and bug reports from around the world in recent weeks.

CanDo -- CanDo finally merged completely with the improved SchoolTool web interface. 2007's developer internships paid off with a strong core of student developers, led by Filip Sufitchi. An 8000 user, multi-school production CanDo server was successfully deployed in the fall and is seeing significant use and generating enthusiastic feedback from teachers and administrators. The year was capped off by an RFP by the state of Virginia for a $40,000, eight district CanDo pilot.

SLA Intervention System -- Principal Chris Lehmann and the staff at Science Leadership Academy in Philadelphia worked with Alan Elkner to create and deploy a student intervention tracking system which has worked well for them and attracted immediate interest in a larger pilot by district administration.

2009 Goals

  1. Release SchoolTool 1.0 in April as a basic, functional student information system, and support schools and teachers independently deploying and using it through the rest of the year, including a 1.1 release in October.

In this year we only need a relatively small group of users, perhaps a dozen schools. Many more might overwhelm our support capacity. Schools and systems looking toward substantial 2010 deployments will already be looking in 2009, so we will be building those relationships as well.

  1. Package and release CanDo so that it can easily be used outside of its Virginia base. Help with support as necessary to make their pilot and likely larger statewide fall deployments a success.
  2. Do fit and finish on the SLA Intervention System, package it, support pilot deployments in Philadelphia and promote it elsewhere.
  3. Raise an additional EUR 50-100,000 in outside funding. The CanDo RFP gives us a good foundation for further fundraising around CanDo. Having that project based essentially in Washington at the beginning of a new administration committed to economic stimulus, innovation and technology in schools shouldn't hurt. SchoolTool core is interesting to any enterprise that is dependent on collecting data for schools; supporting a free SIS that they can distribute to schools and receive cleaner data would be a smart investment. POV has expressed interest in pursuing EU funding. There are numerous angles to explore.
  4. Raising money implies an organization to receive money, which we currently lack. Having a US-based entity that could accept money from US schools and pay developers outside the country would be helpful. I don't think it makes sense to do this via the Shuttleworth Foundation. I could create a US-based SchoolTool Foundation as a non-profit. My initial research indicates that joining the Software Freedom Conservancy may be the best bet, as it gives us a means to receive and disburse money as tax-exempt donations without the administrative overhead of starting and running a non-profit corporation. Canonical may have a role to play eventually as well. Regardless, this will require further discussion and research


We benefited from a stable development team this year. Ignas Mikalaj┼źnas has really taken ownership of the code and done great work this year. You generally can't hire a BDFL, but Ignas and I are forming a pretty good two-headed approximation of one. Alan Elkner has given us the consistent US developer presence we lacked, working closely with our partners in Philadelphia and Virginia. When Alan split off periodically to work on the CanDo payroll, we picked up Justas Sadzevicius at POV, who has quickly picked up the SchoolTool codebase and efficiently dispatches bugs and adds features for us.

This team and their pay rate and bonus schedule will remain the same this year. We expect Alan to work about one quarter time on CanDo's budget. We aren't budgeted to support CanDo internships this year. That amount could easily be eaten up by exchange rate swings in Alan's salary.

Primary Developer Responsibilities

Ignas > SchoolTool core releases, packaging. Alan > Intervention System, CanDo liaison.

Other Expenses

Travel: I've reduced the allowance slightly because we won't be sprinting at PyCon and EuroPython this year, which will save the expense of longer hotel stays and conference fees. We will sprint at the Arlington Career Academy in early February and probably again there during the summer.

Servers and Sys Admin: Brian Sutherland will continue to serve as sys admin for the near future, but we are increasingly able to shift the complicated bits to Launchpad. We also stopped hosting a demo instance of SchoolTool. It is a better use of time to make it easy for people to apt-get their own instance. So overall, we will spend less on this line.

No comments: