Interface proposals

Part of the charge of the HerbUX project was to propose a novel interface in response to the needs identified by the community. The findings of our research suggested many directions for improving search interface; this page summarizes some of these design proposals.

The project research phase yielded many concerns, priorities, and ideas. In entering the design process, we sought to incorporate as many of those suggestions as possible, while establishing a clear scope and focus to the final product. This meant that many worthy ideas were deemed out-of-scope and/or impractical for our proposal; we have recorded them in our community needs assessment, but have left them out of our design.

The following are the issues and suggestions that came out of our research discussions:

Intended target audience

We decided that our primary intended audience for an interface proposal would be undergraduate Science educators. There were a number of reasons for this decision:

• educators need an interface that bridges specialized content and non-expert audiences (i.e. faculty and students), a particular concern of this project
• current interfaces do not support educational applications well—expert audiences are better-supported
• museums’ needs tend to be more specific to an exhibit, and have to support rapid, experiential interactions, a realm we felt would be a greater stretch for the content

Intended functional scope

From the research findings we identified a set of core general functions that address many of the identified needs. The user should be able to:

1. interact with and present specimen collections in a visually compelling manner
2. create custom sets of specimens
3. annotate and frame those sets with user-generated content
4. place specimens and specimen-sets in other disciplinary contexts
5. share specimen-sets with others (especially students and colleagues)

Design principles

In generating the proposal, we followed the following principles:

• Visual first
  Whenever possible, representations and interactions should prioritize the visual over the textual. In practice, this meant using specimen scans as the primary data representation
• Herbarium-centered
  We sought to create an interface that would not replace the physical herbarium, but extend it. This meant adopting a unifying metaphor (the workbench) that places the physical Herbarium workspace at its center
• Accessible to broad audiences
  As an ideal, we sought to make interactions easily understood by and accessible to broad, non-expert audiences. This was achieved through plain language, the prioritization of common names over scientific, and the aforementioned “visual-first” approach that showed data rather than described it
• Education first
  The primary purpose of this interface is to provide a platform for students and faculty to explore herbarium collections and create shareable presentations of the results. While this does not preclude other uses, this was placed at the core of our design approach

Walkthrough 1
Introduction and overview

Teaching in the herbarium often takes place around a central workbench. This surface starts empty, but is filled by an instructor by selecting specimens from the storage cabinets using criteria (think of this as “querying the database”), and placing them on the bench. The instructor then arranges the specimens in groupings and piles according to their learning goals. The class is brought in and gathers around the workbench to see the display. When the class interacts with the data, it is by examining and moving the specimens.

Many current search methods are centered on single transactional retrieval (i.e. submit a query; get results back), but often stop there. Searches are submitted in a single transaction, and results are returned as lists of text with thumbnails of the scanned card image. Little can be done with the results once they are acquired—they must be downloaded for further processing.

Our proposal is intended to simulate its counterpart in the physical herbarium. The following video is a simulated walkthrough of the main workflow of the application:

1. Click through a welcome screen to the workbench
2. The workbench, initially empty, can be populated with specimens by applying facets (filters). Facets allow search to be incremental, gradually applying filters in turn
3. The query results are returned as specimens laid out on the table and arranged according to various criteria
4. When the right set is found, it can be moved to a saved workbench, where the layout can be customized and annotated for presentation
5. Finally, the saved workbench can be shared

Walkthrough 2
Visualizing and filtering by time

In the first example walkthrough, specimens were filtered according to taxonomic criteria. In this example, we explore visualizing and filtering according to temporal criteria—specifically the absolute date and time-of-year in which the specimen was collected.

This is intended to demonstrate how each metadata facet can be visualized in ways that highlight the properties the data type (one obvious example: spatial data visualized on a map).

Whenever possible, the visualizations incorporate the specimens directly—as is the case here.

Walkthrough 3
Extending through linked open data

A frequent request arising from our conversations was for the ability to incorporate additional contextual data. This idea has recently been explored in the literature on the concept of the extended specimen, in which a specimen is placed within a network of connected (usually scientific) information.

In our research, what was surprising was how consistently educators sought connections to data typically associated with the Social Sciences and the Humanities—such as, for example, using collections metadata to look at the history of science through the biographies and historical contexts of the collectors.

The semantic web offers the technical infrastructure to bring in this supplementary data via linked open data. Our challenge in designing an interface was: how do we leverage the power of linked data resources, but through an interface that is easy to use and intuitive?

Our approach was to move away from the idea of a general-purpose linked data tool in favor of a simpler, more targeted tool that presents a highly curated (and therefore simpler, less error-prone, and more optimized) means to pull in supplementary material.

This curated simplicity will remove much of the complexity (and richness) of the open data universe, and we anticipate developing an “expert mode” that would accommodate a wider range of users.