Dear Energy Disaggregation folks,
Over the past few years we have kicked around some ideas for setting up a wiki for energy disaggregation information. e.g. lists of NILM datasets, data recording hardware, NILM software, NILM companies etc. We're hoping to put something into action some time soonish. If you're happy for us to just get on with this on our own then, of course, feel free to stop reading now :). But if you have an interest in how the wiki is set up then please do continue reading and chip in with any thoughts or experiences...
As far as I can tell, an excellent wiki software system is
MediaWiki (the code behind WikiPedia and many other wikis). I think there are ways to connect MediaWiki to Wordpress (which runs
nilm.eu at present) but that probably isn't necessary unless there's a good use case for it.
But, of course, a lot of the content we've spoken about for the NILM Wiki is content which can be presented in a table (e.g. a table of NILM datasets, listing various parameters). It would be great if whatever wiki software we used could 'understand' the data in the tables so the data can be searchable / sortable / re-usable / machine readable. It's possible that we can do this just by using MediaWiki's '
infobox' feature. But I'm not sure if that allows us to aggregate multiple 'infoboxes' (each of which is basically one
row of data) into a
table of data.
We have discussed using a MediaWiki extension called
Semantic MediaWiki (SMW). SMW looks like a 'normal' wiki (just like wikipedia) but has the added benefit that SMW can 'understand' some of the information on the pages. So, for example, to create a table of datasets we'd create one wiki page per dataset where we'd use a simple markup to denote various properties of each dataset (sample rate, number of houses, etc). Then SMW can automatically generate a table listing all or some of these properties in a sortable and searchable table. So it makes it easy to search through the database, to sort and filter the database etc. And also makes it possible for other software to programatically query the database.
Actually, after doing a quick bit of googling, it looks like perhaps we should use a new MediaWiki extension called
Cargo (instead of SMW) which looks like a modern, simpler version of SMW. Here's a page discussing
the differences between SMW and Cargo.
From the quick bit of googling I've done now, my hunch is that MediaWiki combined with Cargo should work for us. Not entirely sure though!
Just for reference, here's the (long) blog post (with a link to an even longer Google doc!) from 2013 discussing some of our original ideas about the wiki.
Thanks loads,
Jack