December

Goobi viewer Digest for December 2020

Coming soon 🚀

  • Adjustments to the IIIF interfaces

  • Themes with right-to-left reading direction

  • Record based hints with formatted text

General

In the last days of December there have been no significant developments in Goobi viewer. The focus was on cleaning up, bug fixing and closing the year. That is why we do not report on any new developments in this digest. Instead of looking back at the last twelve months, we would like to give a preview of what is definitely coming in 2021 and what we have planned for 2021.

What is to come

IIIF

In the Coming Soon section, the item has been there for half a year now. In January, development is finally to begin. In mid-2020, the IIIF community approved the specifications for version 3.0 of the Image and Presentation API and declared it stable. The Goobi viewer will also support these specifications shortly. In the course of this, the other APIs already implemented in this context (IIIF Search and IIIF ChangeDiscovery API) will also be updated.

We may also see the implementation of the IIIF Authentication API in 2021. However, there are currently unresolved challenges with announcements by various browser manufacturers to discontinue support for "third-party cookies". The background to this decision is that this particular type of cookie is used heavily by the advertising industry for tracking and this should be stopped. However, the third-party cookies are an essential part of the IIIF Authentication API, indeed it is based on exactly this functionality. Therefore, the community has been in contact with the W3C Web-Advertising Working Group to get involved in the decision-making process to either ensure the functionality of the API or to work on the development of an alternative. Depending on the outcome of the further discussion, the result will also significantly influence our decision to implement the API or not.

Right-to-left themes

As announced in the last digest, we will be making extensive changes to the Goobi viewer core in the coming months. Specifically, we want to implement a theme for a project whose target group is equally divided between Arabic and English as their mother tongue. However, the languages have different reading directions. Accordingly, the entire theme should be mirrored when the language is changed. To achieve this goal, XHTML structures have to be adapted and CSS instructions have to be changed to newer concepts such as CSS Grid and Flexbox. Concepts that are considered obsolete, such as floats, are removed. Furthermore, individual solutions must be developed, for example for absolutely positioned elements or arrows. An example of an absolutely positioned element is the flag "Add to watchlist" above a thumbnail in the search hit list. Arrows can be found, for example, in the full-screen display at the top right to exit or below to expand and collapse the sidebar.

A positive side effect is improved responsiveness for mobile devices.

Record based hints

The Goobi viewer will get the possibility - depending on the theme in the sidebar or in the content area - to display record based hints. For this purpose, there will be a new link in the widget "Changes to the record" which will then point to a new area in the backend. There, a title and a text can be added for the activated languages. The text can be formatted with the familiar WYSIWIG editor.

Citation recommendations

The Goobi viewer will get support for different citation styles. The supported formats must be stored in the configuration file. The format can then be selected in the widget "Citate and Reuse". A citation recommendation is then displayed directly in the same widget. Typical examples would be the citation styles American Psychological Association (APA), Modern Language Association (MLA), Harvard, Chicago, ... .

Further

Without going into too much detail, further visual changes for the display of multi-volume records, developments in the context of maps or the completion of the operating concept in the admin backend, etc. are in the pipeline. More about this in due course.

What we have planned

We have set ourselves one big goal: We want to make the search better. It's been a few years since the last major redesign, and we keep noticing things or getting feedback that could be more intuitive and better.

We need the help of the community to achieve this goal! The search covers a wide variety of areas and all of them want to be considered in some way. For example:

  • The user interface for simple and advanced searches

  • The user guidance after a search to further refine the search query

  • The visual design of the search hit display

  • Performance!

But the most important thing is:

  • The algorithm for determining the right search hits.

This is precisely the point where we need help. We are collecting use cases in the form of user stories. These are written in colloquial language and describe a way of working from a user's point of view in order to achieve a goal. For example,

On the home page, I click on a title in the New Additions widget to open this work.

In search, we collect user stories that are a little bit more detailed. For example:

I am in the Goobi viewer at the address https://viewer.example.org. There I search for the word house. I find hits like glasshouse but also like houshold. For me, this means that an automatic link and law truncation takes place. This also happens when I put the term in inverted commas, but for me this indicates a phrase search. I would expect no truncation to be done in a phrase search in the simple search.

Another example of a user story is:

I am in the Goobi viewer at the address https://digital.bibliothek.at. There I search in the simple search for the following sentence in inverted commas: "The discovery of the world". This is the title of a digitised monograph. I expect the hit to appear in the first place, because it only exists once in this form. However, I find many other full-text hits for works in which the words discovery and/or world appear. The actual hit can only be found at the very end of the hit list.

With these user stories, we can then track behaviour with a defined dataset. From the third quarter of 2021, we would then like to go into detailed planning for the revision and improvement of the search. Ideally, we can discuss the whole thing in the autumn (at a user meeting??) and talk about what the implementation will look like.

There will also be a direct call for user stories in the Community Forum next year. There we will publicly document and discuss the stories.