I had a conversation with one of our PIs recently - she’s becoming anxious about not being able to find data after researchers have left the lab. I talked her through the usual data management advice, about using a good naming convention and a well-organised directory structure etc, but she showed me her lab’s filestore folder and it’s clear that she’s actually doing all of the above pretty well. The problem is the sheer volume of data and the rate at which it can be generated, that requires an almost cryptic naming convention to differentiate datasets. It’s starting to look like a traditional files & folders operating system will not cope well with *very* large research directory trees.

In most modern website content management systems, users deposit data (images etc) without having to see what disk they go onto, and without worrying about a directory structure. I’m wondering if the same might be possible for an entire institution’s research data, using an ELN (or other documentation system) as the management interface - so *all* discovery and access would be provided by verbose, plain English links in the description of the research, and users would not be permitted to see/operate a directory structure at all. I understand this is the basis of most data repositories - my idea is to extend this method into the labs, for management of ALL data, rather than just the tip of the iceberg which is the published work. 

This would be a huge cultural change of course, and it’d be a challenge to convince researchers that they no longer need a file browser. A leap too far for many, I expect. It’s also not completely clear that this approach would be any more scalable than a traditional file browser approach. So I’d be very grateful to hear from anyone who might have considered this idea or experimented with it already - if the idea has even tiny wee buds where legs might one day grow, it’d be good to discuss further..

Thanks,

Al