Brutally Honest March 2005 SchoolTool Update ============================================
By Tom Hoffman, Project Manager This is a good time for a comprehensive update on SchoolTool development. During the past month, much discussion has taken place in a series of face to face meetings between me and our developers and partner schools, and the conclusions from these discussions hasn't systematically filtered out to the mailing lists. My communications tasks have become more complex now that I have a set of "customers" to keep in touch with as well as the developers. Also, there are about six months of development in the SchoolTool trunk that has not yet been seen in a serious public release, so the difference between the SchoolTool that a person has running on their Ubuntu server and the current state of development has become quite broad. Our Roadmap and Where We Are In It ---------------------------------- Let's review the roadmap as of the fall of 2005. We were aiming for a complete beta in April 2006 could be sufficiently tested and refined through the spring and early summer that our 5 partner schools (in the Northern hemisphere) could "throw the switch" at the beginning of the 2006-2007 school year and rely completely on SchoolTool for contact info and demographics, attendance and grading. Even if we achieved that goal, it would be difficult to anticipate how many other schools would be able to download and use SchoolTool in the 2006-2007 school year without the additional support and customization we'll be providing our partner schools. Regardless, we are a little behind. Our developers have stayed on schedule: POV and Stephan Richter have done good work on time for us. The primary thing which has thrown us off is the length of time it took me to replace our Etria developers, after Etria unexpectedly dissolved itself last fall. It was definitely a case where having a globally distributed project managed by someone with limited experience in software development (i.e., me) cost us a few months. I was indecisive, but I do think I made the right move in hiring Infrae to join our development team, and they're now working on their first SchoolTool contract. We are going to fall a bit short of having a complete beta in April. It would be stupid to try to push our partners into doing a full production deployment of SchoolTool without sufficient testing time in the spring and summer, so as a result we need to reframe what is going to happen in the fall as a round of beta testing. In fact, one could make a good argument that we should have looked at it this way all along, and that it was unreasonable to think that a school should turn over all its mission critical data to a new application like SchoolTool without a more gradual break-in period. What became apparent as I visited our partner schools and discussed their needs in more detail was that each one didn't need to start using all the components in the fall. As a result, the most sensible plan seems to be to have each school test one or two components in the fall, and to do it in a safe way that will minimize the risk of losing data if we run into problems. Since the schools have different priorities and needs, we can still test each of the components in at least one school in the fall, turning on additional components in each school as the year goes on. Partners -------- A quick review of our partners: * High Tech High - a group of well known, innovative schools in San Diego, California (actually including an elementary and middle school now). They currently use PowerSchool, which is a complete student information system marketed by Apple, but they've had a longstanding interest in a more customizable open source alternative. * Science Leadership Academy - a new high school opening in Philadelphia, PA. Their authoritative attendance and report card data will be in their district's SIS. They will be using Moodle. At SLA, SchoolTool will have a slightly different role. We will try to use SchoolTool to aggregate data from the district systems as well as locally generated data, and make as much of both accessible to teachers, students and parents as is appropriate. SLA hopes to have a laptop for every student in their inaugural class of 9th graders in the fall. * St. Ives School - a tiny independent girls school in the London suburbs. They don't have any integrated student information system. Some things are done on paper, or using some fairly sophisticated MS Office hacking by Miles Berry. Miles also does a lot with Moodle and Elgg. * La Futaie - a public elementary school in Brussels. By far the least technical infrastructure of all the schools. The pace of their implementation of SchoolTool may depend on how quickly they get ethernet drops and terminals into their classrooms, a project planned for the summer. * Feinstein High School - I hadn't been planning on working with my former school here in Providence, mostly because I was so frustrated with the draconian controls over the school's network by the district. However, I do have unique access to the school and a long history of working closely with them, so it does make sense to work with them as a partner. We just need to set up an off-site server for them. * Smith Leadership Academy - I added this Boston school to our initial list of partners, however, upon visiting the school, it became apparent to Stephan and I that their administrative staff didn't really understand the project and would probably be very demanding and difficult for us to handle with our limited manpower and resources. I hope they will be happy SchoolTool users someday, but they didn't seem like they would be happy beta testers. * Southern hemisphere partners - I'll discuss this issue separately below. CanDo ----- I should also stress the importance of CanDo, a competency tracking system developed in Virginia by a band of students and ex-students led by their teacher Jeff Elkner and by the lead user/evangelist Dave Welsh. They've got more funding from the state to continue development and deploy the product in additional schools. The next versions of SchoolTool and CanDo will be tightly integrated. CanDo paid to have their lead developer, Paul Carduner, fly to Boston for a two-man training/sprint with Stephan Richter which resulted in our core model for both graded assessment and competency or outcome based assessment. We will have two shared sprints in July with the CanDo team, one in New Hampshire and one in Virginia (at their expense). Jeff, Paul and the Cando team are currently training several new student developers in Zope 3 and SchoolTool. Also, in an unexpected twist, we discovered that Virginia has contracted with Zope Corp. to develop their online repository of competencies, so we're talking with them to make sure that their competency repository talks to our competency tracker. Whether or not any other synergies (or rivalries) will develop remains to be seen. Components ---------- Calendaring +++++++++++ The record on our calendar component is decidedly mixed, due to my worst decision of the year: underestimating the difficulty of timezone support. Steve Alexander warned me against it, but everyone else either thought it would be a lot easier than it really is or kept their mouth shut. For a school, timezone support is superfluous, but for a general purpose calendar server, which we had decided we wanted to create as well, if you don't have timezone support, you're pretty much a toy. So we tried to implemente timezone support last summer, but it came out rather badly. We've incrementally cleaned up the bugs while being forced to primarily focus on the other components of SchoolTool. We'll have a much more solid calendar server by the summer, but right now, we kind of have a bad taste in our mouths about it. High Tech High and Science Leadership Academy will both be testing the calendar component in the fall, particularly in connection with the gradebook component, so that assignments, tests, etc. entered into the gradebook will show up on the student (section) calendars. These schools will emphasize the public facing side of SchoolTool. Demographics/Contact Info +++++++++++++++++++++++++ Development of this component is now underway at Infrae. It is fairly straightforward stuff, but requires well-designed forms for secretaries who may be registering students by hand by the dozens or hundreds. This work will bring several external Zope 3 components into SchoolTool including zope.formlib, zc.table and zc.catalog, which should subsequently speed development of the gradebook and other components. This component should be done around the end of April. All the schools will test this to some degree in the fall. Gradebook and Grading +++++++++++++++++++++ Stephan has completed a basic web gradebook, and he is currently revising the term system to manage end-of-term events, grading, and various types of progress reports. This will be done next months. Subsequently, Infrae will do additional UI work and other refinements to the gradebook through the spring and summer, particularly working with teachers from High Tech High. This is, however, one reason that we definitely won't have a complete beta in April. I discovered in visiting our partners that generating report cards was consistently a serious problem, particularly in trying to integrate standards or outcome based assessments and comments with traditional grades. St. Ives was the only school which had a reporting system that seemed to genuinely fit their needs and desires, thanks to some heroic MS Office hacking by Miles. I think this is an extremely common problem right now, as many different streams of international ed reform create more complex reporting requirements. We will be generating report cards in the fall for La Futaie, St. Ives, and Feinstein. I actually used Stephan's basic gradebook and the pdftemplate library to make standards-based report cards for Feinstein at the end of the fall term in January. It worked quite well. ReportLab and pdftemplate will prove to be significant competitive advantages for SchoolTool, and I can imagine schools adopting SchoolTool primarily or solely to generate complex report cards. Attendance ++++++++++ POV is soldiering away on the attendance component. It is going a little more slowly than I'd hoped because more additions to their original attendance model are necessary to fit the American system of attendance than I originally realized. They'll finish a second iteration of work in April, but there will still be several weeks of loose ends to tie up later in the spring. Attendance will be tested in the fall by St. Ives. Southern Hemisphere and Developing World ---------------------------------------- Helen King, from The Shuttleworth Foundation, and I had a meeting in Paris with several members of UNESCO's International Institute for Educational Planning last month, which proved to be very informative. Ken Ross, who runs their training programs related to monitoring and evaluating the quality of education, made the case that the killer app for schools in the developing world would be a free system to facilitate the collection of accurate census data. In most countries in the developing world, we were told, each school must complete detailed biannual censuses of their students, staff and physical plant. These are the main source of information for the national administration, and they are infamously inaccurate. Facilitating this task could easily put SchoolTool in every school in many developing countries. It would make sense to make custom distributions of SchoolTool that only include components relevant to the census, making the application as easy to install and use as possible. South African schools have a census requirement, so Helen and I have been discussing testing this approach with schools in South Africa via The Shuttleworth Foundation. More discussion is needed, but it seems like a very promising angle. Platforms --------- It is becoming clear to me that for the 2006 releases of SchoolTool, we're going to have to drop the pretense of supporting Windows and MacOS as servers. We don't have anyone who really knows how to make SchoolTool run properly as a Windows service, which makes it not viable for a mission critical server like a student information system. If there is demand for it, we should probably pay Enfold Systems to implement this next spring, since they seem to do the most with Zope on Windows. We have a kind of sketchy installer and preference pane for MacOS, but as the server has gotten more complex, the development version of SchoolTool has developed serious performance issues. Generation of certain pages has slowed to a crawl and some features have stopped working completely. Sample data generation just fails to complete, for example. Perhaps some of these problems could be cleared up with some Mac-specific profiling, but there's speculation that there are filesystem related issues issues as well. I don't know of any "Zope on Mac" experts we could turn to for this problem. _______________________________________________ Schooltool mailing list [email protected] http://lists.schooltool.org/mailman/listinfo/schooltool
