Our SIS now includes a fairly mature special education module. It's not perfect, but it's well-integrated with the rest of the SIS and gives teachers immediate and easy access to the accommodations they need to provide for students.
From a teacher's perspective, it really rocks. Instead of trying to find paper copies of a student's IEP or 504 plan (for those of you outside the States, these are official documents that describe teachers' responsibilities, among other things, for students with special needs in the classroom), they simply click a link beside the student's record in their grading and attendance screens. Pretty slick.
Our special education teachers think that piece is great, too. Regular ed teachers can no longer claim ignorance of accommodations for special ed students in their classes. Where this falls apart, though, is with power users of our current special education application. Known as ESped, the web-based application is developed by folks who live, eat, and breathe special education and have a really solid understanding of the workflows inherent in dealing with the extraordinary bureaucracy of special ed. As one of our special education staff put it, they know about changes to state laws, forms, and data requirements long before we do, and have already implemented these changes before we run into problems.
Our SIS vendor, on the other hand, employs a lot of really great programmers who have a brilliant understanding of database modeling and programming (the data model for X2 is complicated because of the volume of data, but appears much cleaner and more normal overall than that in ESped). These same programmers simply haven't lived in the special education world and lack the expertise of their ESped counterparts in this particular area. It's not really a failing of X2; it's simply not their area of specialty.
So here's the question: Do we stick with ESped, manage data concurrency issues between the two systems, and find some other way to provide teachers with special education information easily and transparently? Or do we give up some of the features of ESped and move to a single, unified, normalized database that meets basic requirements but lacks some outstanding features (including specialized support from special education-trained individuals) that set ESped apart? Anyone out there had to make a similar choice?