Introduction

The collection “Photography Unbound” was assembled by the Getty Research Institute (GRI) to research the application of computer vision. It consists of around 30.000 pre-1925 photographs organized in about 300 bound albums from different institutions and a variety of themes. They range from photograms of plants over travel photography to portraits, often with different early photographic techniques. The images differ in their layout as well. It might be one or multiple photos on a page (and thus image file) and sometimes they are cropped to their actual dimension. This variety and album organization initially got us interested in this collection.

Digital representations of albums are nothing unusual. We still use this organization into subsets (less with a free layout) today on every phone. But as we analyzed the aspects of photo albums we noticed that an integral component of experiencing albums was absent: Social Presence. This formed our main research question on how to integrate means of social presence into digital collection interfaces. Combined with the research happening at the GRI where one focus is the presence of people in the photographs an interesting duality emerges. They are looking into the images and we attempt at looking back from them.

In the following sections we will look further into how social presence is manifested in real life. We will lay out related work as a short history of this concept in interface design. To get a deeper understanding of what could be possible and also exists in different contexts already we will propose a framework for social presence in digital interfaces. Based on those components we will present our own screen designs and a prototype for a “Photography Unbound” collection interface.

Social Presence in Real Life

As the collection is explicitly one of albums and their photographs within we took the physical album as a starting point. Albums have no overview, we have to view them page by page. They are mostly organized by a theme with no subsections. Further information about the photographs might be found on the back of them or notated on the page. The layout might be important as well and rearrangement is usually not done or even possible. Experiencing an album often means more than just looking through it. It is a social occasion as well where we sit together with friends or family. We have to collectively agree when to turn the page; someone might point out a detail on an image; another person might be telling a story and goes backwards to bring a reference to mind again. All this forms our understanding and experience of this specific kind of collection which should be taken into account when designing one in the digital space.

This social behavior is integral to other real life collection spaces as well and mostly missing from their digital counterparts. A visit to the museum is always connected to social cues we send and receive, voluntarily and involuntarily. Based on those cues we might avoid a packed crowd in front of an object, or go there just because of it, as this object has to be amazing. We might adjust our pacing to some people next to us, or speed up because they are annoying. Maybe there is someone explaining the object to friends which we overhear and decide to go back because we previously did not quite get it without context. A person standing nearby might be wondering about a detail and we point out a similar object in the previous room.

Related Work

In the 1945 article “As we may Think” Vannevar Bush [1] proposes some of the earliest ideas of information organized through associative linking to create “trails” which can be shared with other people. This article directly inspired the onset of Human-Computer interaction and hypertext, fundamentals of the World Wide Web, through Douglas Engelbart [2] and Ted Nelson [3] in 1962/63.

Navigational ideas of the early web before the proliferation of search engines in the mid-90s mainly revolved around users sharing their personal link lists directly via E-Mail, message boards or on their own webpage to close contacts, people with similar interests or the general public respectively. Another aspect would be collaborative filtering which consisted of recommendation systems based for example on popularity or ratings which are still very relevant for today’s browsing habits. This lead to the differentiation into three navigational models by Dourish and Chalmers [4]: semantic, spatial and social navigation. The latter further developed by Erickson [5], Dieberger [6], Höök, Wexelblatt and Maes [7] among others in the late 90s and early 2000s.

Since that time the navigational aspects haven’t found much attention although the technical possibilities developed further. The main applications of social principles have shifted to collaborative work and recommendation systems with a parallel continuation of message boards in form of modern social media. More research also went into online learning (e.g. Social Presence Theory [8]) closely connected with different social sciences.

One other field of research is computer-supported collaborative work (CSCW). This is very much at the foundation of creating the technology involved which allows multiple people to work together digitally. A great paper summarizing a lot of aspects of this field is by Lee and Paine [9]. Although CSCW shares integral parts with social navigation they define it as „[…] working towards a particular goal through one or more overlapping fields of action.“

The use of social presence especially in collection interfaces is currently quite limited despite technical advancements and its inherent emergence in the real world as evidenced in the previous section. We identify a gap in research, prototypes and bold applications between more commonly used recommendation systems and simulations of museum visits in virtual reality.

Means of Social Presence in Digital Interfaces

We differentiate social collaboration from social presence by the former having an agreed-upon goal which is supposed to be reached within a group. Social presence is for us a far more loose concept and coexists with potentially existing individual goals and habits. But especially in context of cultural collections its main idea is to aid browsing and open up a new dimension of serendipity. From this of course social collaboration might emerge.

We consider the information flaneur [10] an inherently social being; walking on paths which others created, being influenced by others in the same space, and of course themselves creating and influencing.

To help the understanding and context-appropriate use of social means we propose a framework for characterizing digital social presence with four key properties: Contribution, Granularity, Timing, and Persistence.

Contribution
Active: Users intentionally interact socially or contribute information
Passive: Users contribute as a side-effect of the interface’s features

Granularity
Individual: Users can be differentiated
Collective: Users’ interaction is statistically evaluated

Timing
Synchronous: Users only see other users’ presence live or with a small delay / time average
Asynchronous: Users see a history or statistical evaluation over time

Persistence
Limited: Users leave traces which get hidden or change over time or by other criteria
Permanent: Users’ traces are forever documented and/or reproducible

The following is a none-exhaustive list of digital means of social presence. The respective property values are not necessarily fixed but more a result of applications we observed and characteristics we deemed sensible for them. But of course they are rather implementation-specific. For example, we mainly see individual Co-Presence because it is mostly applied to online co-working tools, this does not make some sort of collective version completely unfeasible but again referencing the real world, an individual granularity seems more logical.

Let’s take a look at some examples from e-commerce, social media, collaborative tools and cultural collections.

Amazon uses two main social features: Indirect Recommendations (Passive, Collective, Asynchronous, Limited) and Reviews in form of Ratings and/or Comments (Active, Individual, Asynchronous, Permanent).

Apart from View Counts (Passive, Collective, Asynchronous) and Likes/Dislikes (Active, Collective, Asynchronous) YouTube recently integrated an interesting visualization of Popularity (Passive, Collective, Asynchronous) within the video itself.

Google Docs as one of the major tools for collaborative work is based on Co-Presence (Passive, Individual, Synchronous) by displaying the current text cursor position of all users and their real-time edits. The cursors are color-coded and if the users are unregistered they get a generated animal-related name and icon.

Miro works as an online-whiteboard with an infinite canvas on which users can work collaboratively. One main feature is Co-Presence (Passive, Individual, Synchronous, Limited) by showing the current pointer positions of all users with the possibility to hide them. Those are color-coded and get generated names when unregistered as well. Other features related to this are Following (Active, Individual, Synchronous, Limited) and View Sharing (Active, Individual, Synchronous, Limited).

Below the Surface is a collection of objects found in an excavation for a metro line in Amsterdam. Users can add those objects to their own display which can be published. Other published displays can in turn be “remixed” and used as a starting point for one’s own. These constitute the means of Curating (Active, Individual, Asynchronous) and Sharing (Active, Individual, Asynchronous).

The Tyne & Wear Archives & Museums collection tracks the objects users viewed and can draw connections between them. Those Tracks (Passive, Individual, Asynchronous) in form of a mosaic view of the objects can be shared by a link although without the possibility to view their connections.

The use of all those possibilities has to be evaluated individually for every project and context. To build an interface with social presence explicitly in mind we would suggest to utilize multiple means which have a difference in at least two properties (e.g. Co-Presence [P|I|S|L] and Popularity [P|C|A|L]). Furthermore, especially for a Collective Granularity, the trust users place into the interface has be considered. After all every digital representation of Social Presence is some form of abstraction, missing most of the social cues we would use in our daily lives.

Other aspects are not to be neglected as well. In a museum we don’t mind other visitors seeing where we are going because we chose to be in this social sphere and this is a passive artifact we are used to. Digital Interfaces are experienced through phones and personal computers which can be considered parts of our most intimate sphere, thus posing fundamental questions about privacy and transparency. Active contributions to the interface (e.g. Commenting) also demand a solution for moderation which is often a hurdle, especially for smaller institutions.

Concept and Design

We wanted to explore some different means of social presence within our collection interface. Those should not be just an addon or layer on top but considered from the beginning. The means we use are Co-Presence (P|I|S|L), Popularity (P|C|A|L) and Curating (A|C|A|P) as they have quite different properties and functions.

The main takeaway from our small co-design workshop was that there was a great value in understanding the collection by laying images out on a virtual table. Our layout consists of three circular layers or rings: albums (inside), user curated albums and filters (outside) on a freely navigable plain.

The albums (original and curated) have a cover which is scaled according to their popularity determined by clicks and also the amount their images were added to curated albums. This creates visual anchor points for orientation and more generally a collectively driven dynamic layout. Curated albums use a generated color with typography as a cover which consists of a generated title from the images’ meta data. Images from the albums are clustered around them as a force-directed graph with the more popular ones having a stronger force towards the album’s cover. Than there are filters (including search) which get added to the outside ring this allows them to be more plenty and don’t disturb the rest of the display. Filters which don’t get used by others fade over time. They are as Dieberger puts it forest paths: Only visible as long as they are being used. Any of those three entities will itself be positioned according to their popularity and will move towards or away from the center. Those changes will be evaluated and changed globally over a longer period of time as not to disrupt the usability of the interface.

The startpages is an introduction to the collection with the possibility to follow a quick tutorial explaining the various features.

The main view has all the images layed out as described earlier. Cursors show the presence of other people. The lines connect duplicate images which are in the original as well as in user albums.

Clicking on an image brings the user to the album view. In the center is the image which was selected within a horizontal caroussel which is the original album (depends on the context of selection). The caroussels above and below are albums where the image is also included. By scrolling this context of the image can be changed.

Prototype

To test some aspects of social presence on this collection and let that potentially feedback into our designs we decided to realize a prototype with Miro supplemented by this questionnaire. This offered us the possibility to do a relatively easy functioning prototype because Miro has some features we were interested in already integrated and also decent performance with this large dataset.

We reduced our design back to a basic mosaics as Miro is not made to support our idea of dynamic layouts. This was not so much of a problem as we were also interested in what people might do with this collection given total freedom.

To our disappointment the board remained completely unchanged since we published it through various communication channels. We also didn't observe anyone ever being on the board. Thus we can't draw any conclusions from this prototype. Maybe the topic of this collection was just not interesting enough in itself for a broader audience? Or our communication was just not exciting enough? We avoided any sort of gamification  for a reason but it might have helped here. We don't really know.

Technical Background

The technologies for various implementations of especially snychronous social presence have been around since the early 2000s. The basic idea is that the action which one user might do (e.g. moving the cursor) is sent to the server which sents it further to all other active clients where the screen might be updated with the new coordinates. This requires a constant communication between the server and its clients. HTTP was not designed to do this, it only covers „user requests - server answers“ but the server also has to sent the updates without the user requesting them.

The earliest implementation was through Comet which is a collection of different methods supporting two-way communication. For over a decade now this has been succeeded by WebSocket which was standardized and has a wide support among browsers. It opens up a communication channel between the server and client where both can sent data freely. This allows for live updates and thus collaborative work and social presence to work pretty seamless in the web of today.

Conclusion

We consider the framework and research into social presence the most important outcome of our project as it opens up a whole bandwidth of possible implementations. It is not specific to the Photography Unbound collection or not even to cultural collections in general but it wouldn't have emerged in that way without the analysis of the real world characteristics of photo albums.

As a suggestion of what might be possible we demonstrated an approach we very much created from thinking about social presence by letting the complete layout be driven by social means.

What we are sadly missing out at this point is an analysis of how different people with different goals (or none at all) would interact with such a layout and with the possibilities given. Our failed prototype or the very minimal co-design workshop didn't really provide any answers to that question. This would obviously be required before thinking about further refining designs by any means.

References

1. V. Bush: As we may Think, The Atlantic Monthly, Volume 176, No. 1, pages 101-108, July 1945
2. D. C. Engelbart: Augmenting Human Intellect: A Conceptual Framework, Summary Report, Stanford Research Institute, on Contract AF 49(638)-1024, October 1962
3. T. H. Nelson: Complex information processing: a file structure for the complex, the changing and the indeterminate, ACM ‘65: Proceedings of the 1965 20th national conference, pages 84-100, August 1965
4. P. Dourish and M. Chalmers: Running Out of Space: Models of Information Navigation, Proceedings of HCI’94, Glasgow, August 1994
5. T. Erickson: From Interface to Interplace: The Spatial Environment as a Medium for Interaction, Proceedings of Conference on Spatial Information Theory, 1993
6. A. Dieberger: Supporting social navigation on the World Wide Web, Int. J. Human-Computer Studies, Volume 46, pages 805-825, 1997
7. A. Dieberger, P. Dourish, K. Höök, P. Resnick and A. Wexelblatt: Social Navigation: Techniques for Building More Usable Systems, Interactions, pages 37-45, November/December 2000
8. G. Cui, B. Lockee and C. Meng: Building modern online social presence: A review of social presence theory and its instructional design implications for future trends, Education and Information Technologies, Volume 18, No. 4, pages 661-685, 2012
9. C. P. Lee and D. Paine: From The Matrix to a Model of Coordinated Action (MoCA): A Conceptual Framework of and for CSCW, ACM CSCW, 2015
10. Dörk, M., Carpendale, S., and Williamson, C.: The information flaneur: A fresh look at information seeking, CHI ’11: Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, pages 1215–1224, 2011

Team

Tilmann Finner: Concept, Data Handling, Research/Framework, Paper
Linus Langkabel: Screen Designs
Armin Ajoori: Concept, Prototype