If I’m anxious about Google becoming a database vendor, it’s partly because the user experience with existing databases has been so dismal to date. On the other hand, Google’s understanding of and commitment to usability is head and shoulders above any of the other vendors in that world. Maybe Google’s completed version of Book Search will have an interface that invites rather than repels use, and has a stable long-term vision driving its design. If so, it might almost be worth it to just let them go ahead and fence off the commons, for the same reason that the consolidation of monopoly capitalism in the late 19th Century at least paid off in terms of standardization across a broad range of products and technologies.
Working on a couple of new projects, I’ve been diving back deeply into catalogs and search spaces and portals. It’s mostly been a depressing experience. Here and there, I have a satisfying feeling that something I’ve used for years has steadily improved. Our own local catalog Tripod is so vastly better in basic design and navigation than a decade ago that it’s almost startling. Another old stalwart, JSTOR, feels more intuitive in its design than it once did.
Travel across various search spaces and databases, though, and several basic frustrations arise.
1. Databases which default to an advanced rather than simple interface upon first access. Sometimes that’s because a portal points to the advanced interface, sometimes it’s because the basic interface is a hidden or obscured option.
2. Basic interfaces which are cluttered or require toggling four or more separate drop-down menus or other settings even to carry out a basic search.
3. Advanced interfaces which are really cluttered, with constraining menus, toggles or radio buttons scattered across multiple columns. Sometimes a search page looks like someone vomited up every kind of interactable object that’s ever been used in a form or UI. (Or as in the case of ISI Web of Science, with a marketing slogan at the top that’s made to look like interactable text.)
4. Diversity of interface designs. By now, we really should be converging on a common design. Instead, every vendor seems to feel an obligation to maintain a different design as a branding tool, not to aid users.
5. Constant shuffling and pointless tinkering with the UI for databases. It’s one thing to make a really big shift (say, towards an inviting basic entry-point interface away from a cluttered entry-point advanced interface) and another thing to constantly move menus around in a page layout. But the latter is very common behavior.
6. Really low standards for the quality of digitization and for searching within digitized text. JSTOR is a happy exception, but some other digitization projects are just hair-tearingly poor once you get into the nitty-gritty and start to make serious use of the resources they hold. There’s at least one company doing archival digitization where I find the type of material they’re digitizing appealing but I’m prepared to argue against ever buying anything they’re doing because the design and usability standards of their work are so slapdash.
7. Fragmentation of material. Rather than moving towards amalgamation and interoperability across databases, you really get the sense that everybody’s been busy grabbing at whatever piles of text they can lay their hands on, building the biggest little mudhill they can manage to put up, and then building walls around it. There are interstitial services that help a user “jump” from one little fragmented collection to another and portals that aim to be a “top level” to return to, sure, but we should be doing better by now.
I can’t believe this fine post hasn’t generated any comments. I tweeted about it, as did Tom Elliott.
I touted point 7, which I think is most significant long term. Tom quoted point 3.
I, too, am struck by how slowly convergence towards optimal UI has come. I wonder if there are intellectual property issues which affect design as much as the need for branding. I keep thinking that the “killer app” may not be some cool new tool, but some new user interface that finally clicks as fully intuitive for all users.