In 2012 Brill implemented a new platform called Brill Online, intended to integrate various contents including journals, e-books, and primary source materials. The platform was developed by Publishing Technology plc (developers of ingentaconnect) and is based on a customized instance of pub2web). An examination of the platform shortly after trial access came available in late spring, 2013, indicated that the platform currently fell short of functionalities offered in a number of licensed digital collection interfaces. The publishers have confirmed that some recognized bug fixes and several revisions are “underway.” But the platform and underlying metadata will require considerable upgrading in order to meet common expectations. The following observations include the developers’ notes on the status of work underway as of early fall 2013.
Search
The search feature offers both simple searching (all fields) as well as an advanced search. Any query will search both the collection and document metadata as well as full-text documents. Search results indicate where the match is found (in Title or Imprint, for example, or in the text itself). Results can be sorted by relevancy or title.
Fixes and improvements underway
- Result citations need to show the publication title in addition to issue and date.
- When a publication citation is displayed, either as result of browse or search, the interface presently defaults to the “Periodicals by Country” tab, hiding the list of issue dates. The user must know to select the “Contents” tab.
- Back buttons are needed to return from the document reader to search results and full collection.
Functions needing further improvement
- Filtering of results is very primitive: there are few options and the results are confusing.
- Filtering seems to apply to metadata at the journal title level rather than issue or item. For instance: when place was selected as a filter option, the place choices did not display, but a search result of 10,366 citations was reduced to one citation.
- Filtering by “date” included a choice of a fixed date (1909) and an apparent open date (1909-), but both returned citations which included issues published after 1909. Also, the date filter option included the date 2013, taken from the metadata for the overall digital collection.
- Another filter option, “type”, only indicates PDF as a format, which is the format of all of the online content.
- Navigation to search results needs much improvement.
- There is no excerpted text for contextual use within the results, forcing a user to click each result to see if the result is relevant.
- When selecting a search result, rather than taking a user straight to the page containing the full text, the interface first takes one to an issue description page with metadata about the title. A user must further click the document reader to open up the issue, which opens to the first page of the issue. Developers have indicated this is a design choice. They will consider adding links to the first location of results.
- While it might appear that the user must perform the search again within the issue to find the term, the search term hits are in fact displayed in the left-hand navigation bar once the document reader is opened. A user has to select “go” for contextual snippets and hit-term highlighting to appear. A fix is underway so that the document reader will open at the page containing the first search term.
Browse
Browsing by country appears relatively straightforward, producing an alphabetical title list of all publications from each country (with place and dates of publication, where available). Clicking on a title results in a fuller bibliographic description and a preview window for the title (in the case where only one issue is present in the product) or a list of available issues.
At the same time, clicking the tab labeled "content" produces a list of titles in no particular order, with no ability to sort the titles.
Fixes and improvements underway
- Clicking the tab labeled “Contents” will sort results geographically or chronologically, using additional metadata. (Chronologically by start date of the serial publication?)
Functions needing further improvement
- On clicking a title with only one issue available, it is not immediately apparent from the interface what issue is available, nor that it is the only issue available. The bibliographic metadata (in the “Note” field) displays an issue number and date, but for consistency’s sake, it would be helpful to have the single issue listed in a “Contents” tab, as in the case of titles with multiple issues.
- The “preview window” automatically displays the fifth image scan (as an estimated preview of the content), rather than being tagged to display the first page of the title/issue. This does not provide a user with a clear sense of what the title/issue might contain.
- Closing the full-screen reader takes a user back to the page for the particular issue, with no means of returning to the list of available issues. A user must click the browser’s “Back” button a least three times to return to the title-level metadata containing the list of available issues. Linking the issue back to the title page (via the breadcrumb navigation, for example) would be helpful.
Display
The full-screen reader allows for easy navigation within an issue, and the ability to search for words within a particular issue. The left-hand menu offers both a “thumbnail” of each page as well as a search feature with Hit Highlighting.
Search results vary depending on the quality of the image. Images are presented in 300 dpi grayscale, captured from microfilm. As such, the page images are dependent on the quality of the hard copy originals as well as the original film quality. In general the reviewer found the images to be eye-legible and free of film defects, though some titles were filmed at a high reduction ratio, affecting image legibility and, as a result, OCR results.
The default display size of each title (on a 1680 x 1050 resolution screen setting) was often too small to read without zooming (it is set on an "average"page size). The reader allows image rotation and page zooming (“+/-“). Additional built-in features for resizing images (such as resizing to page width) would be useful.
The publisher states that they strive for 99.5% OCR accuracy.
Fixes and improvements underway
- Downloading pages and full issues as PDF.
- Possible bugs in printing full page images, experienced by the reviewer.