In the beginning (1997), there was SIF, which now touts itself as "the first, largest and most implemented open global standard for seamless, real time data transfer." I think "for schools" is implied. It was an oddball pain in the ass that required a lot of work for application developers to integrate and for most of the past 15 years you'd be paying a commercial vendor to set up a "zone integration server" for you. Google tells me there are finally a number of established open source options and that Pearson (surprise!) bought out the main SIF vendor. Importantly, SIF crowded everything else out through a period of revolutionary change in the web and web services, stunning successes for a few open standards (HTML, CSS...), and a number of partial ones (XML, RDF...).
I tuned this whole scene out for a while until it sorted itself out...
I think the turning point was the Common Educational Data Standards, sponsored by ed.gov in some way, which got rolling in 2009. It is a refinement of the existing core models for educational data and the core of all that follows. It is just "a set of commonly agreed upon names, definitions, option sets, and technical specifications for a given selection of data elements."
Since then, things have gotten cooking.
Ed-Fi is an expanded data model based on the CEDS and some dashboards and other tools and software I don't think anyone will really care about. It is sponsored by the Dell Foundation.
inBloom has a data model based on Ed-Fi, which probably expands it. It also importantly lays out a REST web services API for moving the data around, which is generally speaking the way I'd want to do it. They've also written some open source tools for getting data out of their giant repository, which is also supposed to be open sourced Real Soon Now. The creepy part is just that they really want ALL the data in THEIR inBloom repository. That's backed by Gates, Murdoch, etc.
OK, then we get to the startups. Clever and LearnSprout (probably there are more). They are kind of like a simpler version of SIF where they define a simple REST API, hook into your SIS and host the integration server. They get key student and enrollment data out of your SIS automatically, park it on their server, and other applications your school is using can easily pull it. It is the 80% of what schools want and need in this area, and I suspect they do a good job. It certainly makes sense for smaller schools -- charters -- at least. I would like to think larger districts can and should handle this themselves. If this is a hot startup it isn't saying much for the startup scene though, because I can't see it as more than a niche market and a transitory phase.
The best part is that we should end up with "rough consensus and running code" on some basic REST API's and underlying data model, which is a big deal.
So... what am I getting wrong in the above?