OpenWetWare:Feature list/Lab notebook/2007 Oct Brainstorming/Cameron

Intro

• Our ELN system is at http://chemtools.chem.soton.ac.uk/projects/blog/blogs.php, look at the 'Bio Blogs' and is in the process of being described at http://blog.openwetware.org/scienceintheopen, look for the e-notebook tag.
• This is a blog, not wiki based system but I don't believe that is a central issue. Many of the problems are the same and at the end of the day they are both just ways of interacting with a database of posts/pages.
• Can also see a presentation at http://docs.google.com/Present?docid=dhs5x5kr_543fcwr7c&fs=true with some notes on the talk at http://docs.google.com/Doc?id=dhs5x5kr_558dc7chs
• We are moving towards having the system act as a web service. The idea is that analysis would be handled by external services e.g. through a workflow that could be stored on MyExperiment that would query the Blog, pass the data to an analysis web service (probably a wrapped MatLab or similar) and then post back the results to the blog in some sort of readable form.
• Our biggest current issues are;
  • the need to generate lots of posts because of the one item one post approach we have adopted. We need a widget to autogenerate the product posts. Twitter for the experimental lab! It is after all the one Web 2.0 gimmick we haven't got a 'science version' of yet :)
  • display. We need better ways of actually interacting with the ELN. The needs of the user, the supervisor and the outside world are very different

General thoughts

• Very important to articulate what you want to achieve. Is this just an electronic version of a paper notebook? Do you want it to link materials, protocols, data etc? DO you want it to be machine readable? Should it be just a record of what happened or a polished description of the experiment?
• A lot of the differences between USefulChem and Southampton Lab Blog arise from what the original motivations were. Think about whether you want to go down one route or the other or to try and encompass both.
• Try to build something that works even if someone doesn't use it properly. If it breaks as soon as someone can't do something they'll walk away.
• Expect it to take some time to figure out what 'properly' is, we are still working on it
• Build in tools that encourage people to use it 'properly'. e.g toolbars that provide easy access to links to other posts, or insertion of metadata, drag and drop, or one button upload, of data (not just images!) is very important
• Provide templates, ideally make them part of the way the system works, provides consistency and ease of use
• Being able to clone a post is very useful
• Feel free to play in our sandpit, just pick an arbitrary 'sandpit group' name http://chemtools.chem.soton.ac.uk/projects/blog/blogs.php/blog_id/15 You can register for an account at; http://chemtools.chem.soton.ac.uk/projects/ and then email me to sort out the access

Handling Metadata

• If the system is to be powerful it must include capture of metadata. Our system has gone through several stages where we have changed the metadata that we record and the way we categorise things.
• Simple tagging may be enough but can lead to confusion, we have developed categories of post, and then divide materials and procedures into more or less arbitrary categories.
• Our most recent incarnation is in the 'sandpit' blog http://chemtools.chem.soton.ac.uk/projects/blog/blogs.php/blog_id/15
• For chemistry there has been much good discussion on the best way to identify compounds see;

http://chem-bla-ics.blogspot.com/2006/12/including-smiles-cml-and-inchi-in.html http://usefulchem.blogspot.com/2007/06/inchimatic-chemspider-and-usefulchem.html

• For biology it is less clear how to go about this - do we just need to adopt something? How do I say 'This post is about the enzyme Sortase' or 'I can ligate proteins to DNA'?
• This is important not just for machine reasoning but how people will find the lab books they want on Google. Jean-Claude Bradley has discussed how people find UsefulChem through Google based in InChi searches (see link above for e.g.).
• Chemical Blog space and PostGenomic are good models for community aggregation sites. How do we get these to work effectively on data and notebooks. See for e.g. 'Molecules' at

Chemical Blog space which would pick up a link from e.g. UsefulChem