Hello! Thanks so much for taking the time to post this question.
We agree completely that this is important information that we need to document. We will be publishing a documentation article on this subject shortly, and linking to it from here (and also from the download page you mentioned) once it is ready. Stand by!
Until then, I’ll try to provide some clarity.
As mentioned in the text above, the key functionality that was available in Server that has subsequently been implemented within Interviewer is:
- The ability to export data directly into CSV or GraphML file formats.
- The ability to manage (essentially, list and delete) interview sessions directly on the interview device.
The “data warehousing” you refer to can be accomplished by exporting the data from Interviewer, and saving it wherever you wish - such as a centralized remote location!
On Windows/macOS/Linux you are prompted at the start of of the export process to choose the location to save the data, and remote locations such as cloud storage or network shares are fully supported. For Android, the same can be accomplished via the “share” menu, which pops up once the export is ready to save.
To go into further detail, some users are successfully using secure cloud storage solutions (such as Box, Dropbox, Google Drive, and so on) for this. Many of these services have advanced access control features, such as allowing certain users “write-only” access, which can be particularly useful for ensuring security. Institutions often have cloud storage subscriptions that are IRB approved for these uses. Alternatively, you might consider using a VPN to access a shared folders on university IT infrastructure directly.
Is Server Obsolete?
I just want to talk briefly about if Server is actually ‘obsoleted’ by the above. The idea behind Server was that the two most difficult to coordinate parts of data collection could be simplified by consolidating them within a single app (with these being protocol deployment, and centralized data management). There is no doubt that the new workflow has more steps, and is in some senses more complicated. This is also true of the protocol deployment workflow. However, we still think it is an overall better solution.
The reality is that Server has seen very little uptake, largely because studies large enough to benefit from it are not able to use it due to requirements we did not anticipate. The data store is not encrypted in any way, and has no access control functionality. Despite being called “Server” the app is not headless, and is not suitable for deployment within a transitional data-center setting. Although the data transport is encrypted, implementing this technology added disproportionate amounts of complexity to the Interviewer app, which makes maintenance harder for our very small team.
We might have better phrased this change as the “depreciation” of Server, since Server may continue to be useful in some narrow circumstances. However, we do believe that the safest and most practical way to use the software right now is without Server.