Last week I attended a workshop on Blog based laboratory notebooks in Abingdon, courtesy of Cameron Neylon of the Rutherford Appleton Laboratory and Southampton University.

Cameron has summarised the discussion and findings, and there’s also a good summary of the first day written by Duncan Hull who works with myexperiment in Manchester.

Lots of discussion was generated on the first day with Cameron’s survey about lab books and their usage. I found it interesting that there was variation and a lack of common standards about how to record experiments, how to keep and archive lab books, and even whether entries should be dated or not, even amongst our small group. One of the workshop’s aims was to discover whether the ChemTools system, originally built by Jeremy Frey’s lab in Southampton, could work with experimental procedures in other labs.

The ChemTools system was designed to be open and flexible, it’s based on a blog style system with a lot of custom PHP scripting. The basic concept is that each ‘thing’ in the lab – every input and output of a procedure – has its own post. Metadata used in tagging is particularly important as that connects particular items in a procedural workflow. This can be linked via hypertext in the lab blog, and via barcodes in the physical world (see above for an example).

For a non-production system built by a couple of PhD students, it’s impressive that the workshoppers were able to generate some posts in one afternoon. The group also discovered quite a few usability issues, most of which would be fairly easy to fix. The biggest issue seems to be the time-suck of the current UI – it’s just not suited to writing tabular data, which is the most useful output style for many experiments. To some extent this issue can be alleviated by the existing templating system, and perhaps some Ajax javascripting could be incorporated to make table creation a bit more WYSIWYG. Automation of some of these post-creation processes is being posited for the not-too-distant future.

Further upcoming developments include links to environmental data in the lab (fridge temperatures and other mechanical interactions). Looking forward – keep an eye on Cameron’s blog. His post about rich RSS feeds and tools to make information connections to do things like potentially automatically manage lab stocks predicts much exciting stuff to come!

Conclusions – it’s not [yet] particularly easy to ‘blog the lab’, and requires a fair amount of thinking about one’s experimental protocols and work practices in terms of figuring out what is good metadata to record. However, the richness and connectedness of the data produced using this system does enable the scientist to quickly locate essential data when it comes to writing papers.

What about privacy, trust and levels of access to this lab data? That’s a discussion for a later post.

[Note: this post was originally published on my OII blog, which is about to be subsumed into this one, hence the double-post. Original comments on this post have been reproduced below]