New User Experience
New user setup is critically important. Making it as easy as possible for people to jump start their profiles is hugely important. Any existing user data that we can pull in at signup time is very important.
I would see the new user experience working something like this:
- Click Signup, get taken to a Sign Up - Step 1/3 page which asks you for your First Name, Middle Initial, Last Name, Email address.
- Displays an AJAX spinner, looks you up on PubMed
- Presents you with the matching Author’s publications from PubMed, asks you to confirm it’s you.
- If it’s you, import all the publications, set the cauthor_uid from pubmed
- If anyone else in the system also has these publications linked to them, add them to our new user as collaborators.
- Start a queued task which looks up everyone else who is also on those papers. If they don’t exist, create skeleton profiles, add them as collaborators to our new user.
- Take them to Sign Up - Step 2/3
- Their new profile page has a tour and also prompts them to import their profile data from LinkedIn, CommonCV and maybe Facebook.
Data Sources to Import User data from
I’ve spent some time looking into Common CV and discovered that you can both export data from it and import data into it. This is good news for CollabFinder, but comes with some caveats:
- As far as a user of the site is concerned, this is a manual process - you can export or import your individual data, to an .XML file. There is no publicly documented way that I can find to do this in bulk - for obvious data protection & privacy reasons. There also doesn’t appear to be a sensible way to automate this for an individual - they have to do it manually.
- The format of the XML file that one would need to create is undocumented, as far as I can tell. There’s no developer oriented documentation on the site - only help for end users. They say that they publish the xml schema, but don’t link to it anywhere that I could find. I did manage to find their xml schema via google however, it’s huge, unnecessarily complex and written by someone who didn’t understand what xml schema are actually for. Thanks Fujitsu Consulting :( It’s also not good enough on it’s own to provide all the required information to allow me to produce well formed XML files that Common CV will import.
These problems are all solvable, but mean that it will take longer to build a CommonCV import/export feature than it otherwise would. I will need to fill in a CommonCV and enter every combination of data and reverse engineer the files it exports, so that I can recreate them. I would guess that this would take a week to get right.
This could be sped up with proper documentation - does anyone know if there’s any proper developer oriented documentation for the commoncv import/export formats available?