This post is the second in a series, looking back over the changes in information management over the past decade. Three major and interrelated developments are the move to querying databases of information as opposed to loading information from individual files, the practice of tagging bits of information as opposed to filing things in a hierarchical folder structure, and the representation of information as a temporal stream as opposed to a static page. This post is about the move to databases from filing systems, and how that improves your workflow.
Image via Encrypted Illusions.
Databases vs. individual files
By 2000, the IT industry had been using databases to store and queries to retrieve large amounts of information for quite some time, but most applications that an individual had on their computer still worked by loading each individual file from the disk as it needed to access it. Recalling how you organized your pictures or your music back in 2000, you may have used a program like Winamp. To play music in Winamp, you used the computer’s file system to select the music files you wanted to play and then told the computer to use Winamp to open them. It was simple for a computer user to understand and straightforward for a developer to write programs that worked this way. It was also necessary to work within the memory limitations common to desktop computers of that era. However, as people’s music collections grew, the limitations of this approach became apparent. Not only was it hard to find things buried several folders deep (did I put that band under alternative or rock, were they late 80s or early 90s?), but when you wanted to make ad-hoc mixes, you either had to make copies in a separate folder, or use the playlist builder function of your software.
As computers got faster and memory got more plentiful, developers realized that they could provide a smoother and faster experience for users by loading the whole library (or a decent chunk of it and use smart caching strategies) into memory all at once. The picture management software Picasa was an example of this. Instead of browsing through your disk to the folder that had your pictures in it, you simply opened Picasa and everything was there, neatly organized in a manner that wasn’t tied to how the files were stored on your disk. You could fluidly re-arrange things however you liked, make ad-hoc collections and add notes, and add new things to your collection without having to rename files or move them around on your hard drive. At first I personally found this very frustrating, because I was used to rigid hierarchies for finding things, but eventually having the freedom to just dump all the files in one big pile and let the computer sort it out won me over.
Of course, once your personal image database grew to thousands of pictures and you had added all sorts of information to pictures and collections in a way that folders and filenames just didn’t allow, the whole concept of browsing the library as if it were a set of files in hierarchical folders began to break down. Well, Picasa came from Google and Google is good at search and user interfaces, so they solved this problem by allowing you to use search as the interface to your data. You could now ask your collection things like “show me all the photos taken in 2001 and 2003 that I’ve starred”, without you having to first file the pictures in folders or give them unique names. In short, here was a case where the computer really started to deliver on the promise of being your digital assistant, as opposed to just a digital version of your filing cabinet.
At Mendeley, we’ve learned from these lessons on how to provide a great user experience. We want to free people from the tedium of manually naming, organizing, and filing research papers. Just dump your papers into a folder (call it “inbox” or something), and let Mendeley handle it for you. I’d also like to encourage you to think about using search as an interface to your Mendeley library. The searchbox in the upper right of the desktop client offers tons of useful features, without cluttering up your toolbar and menus with rarely used options. Pretty much any field from the details display of an item can be used to filter your search results. Try adding “abstract:[keyword]” or “year:[YYYY]” or “author:[name]” to a search and see how precisely you can slice and dice your library to show only the results most relevant to you right now.
Mendeley Blog 
That’s right Mendeley is changing the way how scientist will organise and retrieve their electronic documents!
However, while current version of Mendeley let the user to search within ‘author’ and ‘year of publication’ entries. It still does not let him to search within the abstract. Hope that this essential function will be soon added to the next release of Mendeley desktop.
Yes, I do agree that this is a indeed the ‘modern’ and appropriate approach.
However, one thing I am still missing in Mendeley is that while focusing completely ‘on search is the interface’, the file name has been neglected too much in the process (save from the auto-naming feature), and thereby almost /forcing/ the user to use Mendely to access and search for papers. This is not the case with e.g. Picasa; you can still easily reorganize and rename pictures in a file browser /alongside/ the Picasa interface.
I give a concrete example: I have appended to each filename a unique identifier (something like P063, P098, etc) which allows me (a) to reference a paper very easily (I simply write down the unique identifier), and (b) to open very quickly a paper even if Mendeley is closed by simply typing the unique identifier in the windows start search field (Windows 7), or spotlight (MAC OSX).
However, in the Mendeley interface, the search box does not index the file name, so my unique identifier is useless there. (I could put it into a field, but that would be too much extra work).
Again, in e.g. Picasa, the filename is included in the search, allowing a perfect synergy between managing your pictures through the OS (Windows Explorer or Finder), or through the Picasa interface.
It also seems that if you have a document open, it seems to restrict itself to searching that document, and that’s cool. However, it doesn’t seem to search my notes – do I need to use a “notes:” prefix for that?
Micah – That’s the idea! You can search notes using the note: operator, but please be aware that sticky notes aren’t yet searchable this way, only the content of the notes field.
Simon – Mendeley Desktop does indeed search the abstract, and it even prioritizes results having the search term in the title or the abstract vs. the body of the paper. However, abstract isn’t yet available as a search operator.
Ward – Mendeley was designed to make it unnecessary to manually rename files. Just enable the File Organizer and let Mendeley do the renaming for you, then you can search by meaningful qualities, such as author or date, using either Mendeley’s internal search or your OS’s filesystem.