acroyear | on open-sourcing the cloud

Open source personal health record: no need to open Google Health - O'Reilly Radar:

Predictably, free software advocates say, "make Google Health open source!" This also misses the point. The unique attributes of cloud computing were covered in a series of articles I put up a few months ago. As I explain there, the source code for services such as Google Health is not that important. The key characteristic that makes Google Health and Microsoft HealthVault appealing is...that they are run by Google and Microsoft. Those companies were banking on the trust that the public has for large, well-endowed institutions to maintain a service. And Google's decision to shutter its Health service (quite reasonable because of its slow take-off) illustrates the weakness of such cloud services.

I'm inclined to agree. What *should* be opensource is any library that is used to access a cloud system (which Google does for almost everything), particularly if there are security elements because more eyeballs looking at security source code improves it rather than makes it more vulnerable (an attitude Microsoft continues to fight, and lose as more and more viruses attack their software while other more open works prove more resilient).

But the client library is useless without the server, and if nobody will host the server, then there's no point.