Access to Memory is a brilliant archival description management system, written in PHP and available under the CC-BY-SA license. The installation documentation is thorough… but of course I just want to get the thing running and so didn’t bother actually reading it all! I mean, where’s the tldr</a>?!
So here are the essential bits (for a more-or-less bog standard Ubuntu install with Apache, PHP, and Node.js), running as a normal user and installing to a subdirectory.
To start, clone the repository from https://github.com/artefactual/atom.git and check out the latest stable branch (e.g. stable/2.2.x).
Then make the config, cache, and log directories writable by the web-server user: chgrp -R <em>www-data</em> {cache,log,data} (or whatever your webserver runs as, of course).
Now change into the /plugins/arDominionPlugin directory and run make; this will build the CSS files.
Navigating to the installation now will redirect to the installation system, and probably throw up a bunch of errors. Probably to do with missing dependencies, or permissions; sort these out (e.g. sudo php5enmod xsl and you should be good to go.
(Good to go to the next step, that is.)
Now install Elasticsearch. It’s easier than the AtoM docs admit: just do it the normal way with sudo apt-get install elasticsearch. (I’d submit a change to the AtoM docs to remove the “it’s not in the Ubuntu repositories” line, but I’m not quite sure how yet.)
Start Elasticsearch (in the background; the -d switch) with sudox /usr/share/elasticsearch/bin/elasticsearch -d and carry on with the installation procedure. The rest seems to be fairly straight forward.