I'm noticing a pattern at work in the library: I'm confronted with either starting something from scratch and making up standards as I go, or working with old rules that constantly require updates or work-arounds. I've been wondering broadly how to know when it is easier to make up new rules from scratch versus working with old rules.
When I think about it, there are lots of everyday examples where invention is preferred to innovation, or vice versa. Look at Wolfram Alpha (although arguably it relies on google as a standard), the QWERTY keyboard, or, in a less positive vein, Stalinism, for examples of invention rather than innovation. Office computer iconography (or for that matter web iconography) relies on pictures of already-familiar objects (file folders, recycling bins). An e-book looks very similar to a physical book, with virtual pages. Robots frequently take animalistic forms. These are examples of developments that potentially could have been radical, but instead relied on an old set of ideas to progress toward something new.
Obviously it's harder to be successful when reinventing the wheel, because it's difficult to anticipate not only what will be needed over time but also people's behavior. Objects (tea kettle, automobile, door handle) and habits (checking for messages, tracking expenses, communicating with family) frequently seem to evolve over time, rather than coming into existence fully formed.
Here are areas where I wonder whether libraries should scrap it and start all over, or persist with what we have and modify:
1) Format-bound metadata rules. When the library was only responsible for physical objects, our standards worked, and it's tough to give them up. This is not my area of expertise, but when does it become simpler to abandon standards based in a physical environment in favor of a new set of rules?
2) Lending/borrowing rather than providing access to. Integrated library systems were built on the lending/borrowing model, not the "access to" model. Why not build a new system that has both in mind, rather than having to fight systems so that they include electronic resources?
3) Reference ('how-to-use-the-library') materials. Frequently, as I revise out-of-date guides and web pages, I wonder whether it would be less time-consuming to just start from the beginning. Sometimes I spend so much time editing that I think it would have been faster to start with a blank sheet of paper in front of me.
Despite the appeal of a blank sheet of paper, invention can also be exhausting. I just wish I knew in every case which would be the better tactic. I hope this knowledge comes with experience...