Cloudworks: design decisions

Juliette Culver and I have just had a paper accepted for Computers and Education on Cloudworks. A draft is available here. The paper describes the first three phases of design decisions, along with evaluation of each phase. We have just being using Cloudworks extensively as a means of live blogging the ALTC 2009 conference and plan to reflect on the experience and what worked and didn’t. Would really welcome any thoughts on the paper or on the current look&feel/functionality of the site. Below is a summary of the design decisions to date:

Design Decision 1.1 Cloud metaphor

We wanted to avoid the use of technical terms such as ?learning design? and hence choose to call the core objects of the site ?Clouds? and the overall site ?Cloudworks?. The notion of Clouds was intended to indirectly evoke metaphorical images of ?blue skies thinking?, ?thinking at an elevated level?, ?visioning and thinking creatively?. The name ?Cloudworks? also works as an acronym for ?Collaborative Learning Design at The Open University?, although it is important to stress that we do not see Cloudworks as a specific tool solely for the OU but as a generic tool for anyone to use.

Design Decision 1.2 Initial content population of the site

In order that visitors to the site did not find an empty site and had examples of the type of content expected on the site, we made the decision to initially populate the site with some content. This was done in two ways. Firstly, through trawling existing sites for good practice ? this included harvesting the 44 case studies of the use of VLE tools mentioned earlier, appropriation of learning designs generated by the AUTC Learning Design site ( and a selection of examples from other well known learning object repositories and case studies of good practice.  The criteria for inclusion was that the examples should present a good spread in terms of pedagogy, subject and tool use and should provide different types of representations from short textual narratives through to more complex visual designs, as well as being representative of the different potential types of Clouds that might be included in the site. Secondly, once we had a reasonable mix of seeded Clouds, we ran a series of five ?Cloudfests? with potential users, where participants were asked to generate Clouds for the site and where they also critiqued existing Clouds. We used the data from the interviews with teachers and the 44 case studies of the use of the VLE tools, to draw out barriers and enablers to finding, discussing and sharing learning and teaching ideas and used these to help steer the discussing in the Cloudfests. 

Design Decision 1.3 Include social features

Analysis of the design interviews with teachers and of the VLE case studies showed that teachers value the opportunity to share ideas with others; indeed for many a named contact to get further information about a particular learning and teaching intervention was perceived as more useful than finding similar information via a website. This was particularly true if the teacher knew the individual and valued their expertise, but was also because they felt there was then an opportunity to follow up with further queries if required.

The importance of socialisation in social networking is well recognised and is one of the underpinning philosophies we have adopted for the site. As a result, from the early stages of development of the site, each Cloud was intentionally social, in that others could comment on and add to it. In the initial stages of development, these social aspects consisted simply of the ability for users to add comments to Clouds and these comments then appeared in a linear temporal fashion under the Cloud. However, our ultimate aim was to build a much richer set of social functionality, drawing on observation of other successful Web 2.0 social practices, alongside evaluation of users? perceptions and use of Cloudworks.

We wanted the focus of the site to be around Clouds and associated discussions, rather than replicating more complex social networking sites such as Ning or Elgg, where the user can incorporate multiple Web 2.0 tools for aggregating content and for communication. We wanted therefore to keep the focus on objects (Clouds) about learning and teaching. This metaphor of a Cloud as a social object was a core principle of the site.

Design Decision 1.4 Tagging within categories

Instead of allowing completely free tagging we restricted the use of tags, allowing free tagging within three categories: pedagogy, tool and discipline.  The aim here was to make it simpler for people to search for particular types of content without having the constraints of pre-defined vocabularies. We felt these three categories reflected the intended scope of the site and acted as a reminder to users of the kinds of things they might either be interested in looking for or contributing. Again these categories were abstracted from the teacher interviews and case studies, as these were what teachers typically used to filter information. 

Design Decision 1.5 Low barrier to entry

One of the themes at the initial vision workshop was the tension between a low barrier to entry to encourage users to generate content verses the desire for high-quality content (the issue of reputation systems and evidence for quality came up frequently). It was also clear from the workshop that detailed information about a topic was often less important than having contact details for a person to talk to about it  (which triangulates with similar comments from the teacher interviews as discussed earlier). Each Cloud thus consisted of a short informative title, a two-line description, a more detailed account and any relevant links.

Design Decision 1.6 No private content

Another tension from the initial workshop was between the website being open and issues such as rights clearance and student access. Here following the Web 2.0 principle of harnessing collective intelligence resulted in the decision that, in order to gain critical mass for the site, all the content should be open and no private content would be allowed. We felt that in order to capitalise on Web 2.0 practices the site needed to be open and also that existing tools behind institutional firewalls (such as password protected forums, blogs and wikis) already provided adequate mechanisms for sharing and discussions within distinct groups. Openness allows for serendipity, for a Cloud created and discussed within one community to be discovered and re-appropriated in another context. However we also needed a means of validating users and hence anyone can view content on the site, but to add content or comment on existing Clouds the user needs to register on the site. 

Design Decision 1.7 User Profiles

As discussed earlier, sharing and discussing experiences is a core facet of teacher practice and hence we recognised that the information on the site about individuals needed to be informative, to enable others to gain quickly an overview of that individual?s expertise and interests. Hence the user profiles, in addition to having user-generated information (such as name, institution and interests), also included an automatically generated stream of the clouds that user has created. This helps to differentiate users within the site; so for example it might be inferred that users with a lot of Clouds have some degree of authority ? although in the initial stages no peer reviewing or voting of Clouds or individuals was included, this is certainly one of the more advance features we are interested in exploring. The aim is to not only provide a listing of users within the site, but an indication of their interests and expertise.

Design Decision 1.8 Cloud types

The core aim of the site was the intention for it to be a place to share and discuss learning and teaching designs and ideas. At an early stage of the conceptualisation of the site it was decided that these designs/ideas would be described as ?Clouds?. In the first version of the site there were five types of Clouds.

Design Decision 2.1 Amalgamate cloud types

The initial five categories of Clouds were amalgamated, so that now the sole object in Cloudworks is a ?Cloud?. This decision was made because it became clear that it was difficult to categorise clouds into the types suggested. For example, it is not clear if a site containing a number of designs should be included as a ?Resource? Cloud or a ?Design? Cloud. Likewise, the distinction between a tool and a resource was not always clear-cut.  Nonetheless, the types of Clouds which could be included remained the same, i.e. a short description of a learning and teaching idea, a more detailed learning designs or case studies of practice, a question or issue a user was seeking advice on, or information about particular resources or tools and how they can be used to support learning and teaching.

Design Decision 2.2. Increase social features

It was clear that the site was not being used socially. We were generating the majority of the activity on the site, either in terms of the creation of new Clouds, or through use of the site in workshops. As well as retaining the social element of being able to have a comment around a Cloud, in the revised site, new content and discussion was made more prominent on the home page, with a list of new clouds in the centre and new comments on clouds on the left hand side. The intention was to help make the site appear more dynamic and to highlight site activity to encourage further activity. 

Design Decision 2.3 Cloudscapes

A new feature ?Cloudscapes? was introduced to address the issue of focusing on community engagement. Clouds can be aggregated into ?Cloudscapes? associated with a particular event, purpose or interest. Example Cloudscapes include: conferences, workshops, projects, research interests, types of pedagogy, course design team spaces, tool development spaces, or course-specific Cloudscapes. 

Design Decision 2.4 Following functionality

As discussed earlier, the ability to comment on Clouds was seen as the first step to mimicking some of the practices around the use of other Web 2.0 tools. Another practice, evident in many social networking sites, is the idea of indicating who you are connected to ? the concept of connecting to friends and following their activities is prevalent in many sites such as Facebook, Ning, Elgg, Linked-In and Twitter. We were particularly interested in the way in which the microblogging site Twitter ( has been appropriated over the last year or so as a lightweight mechanism for engaging ideas and sharing and were struck by the way in which this matched our criterion for low barrier to entry of use of the site as discussed earlier.  In Twitter posted messaged (tweets) are constrained to 140 characters and tend to be a mix of light hearted and professional comments. Users ?follow? others and can be ?followed?, anyone following you will see your tweets and vice versa. In the e-learning community we have seen an uptake of Twitter as a mechanism for providing a community back chat of discussions around e-learning issues and research. We wanted to explore how such practices might be replicated in Cloudworks, as a result a ?follow? feature was added to the site. Users can follow both people and Cloudscapes. A list of who and what they are following then appears dynamically on their user profile, helping to enrich the picture of an individual?s interests and expertise discussed earlier.

Design Decision 2.5 My Cloudstream

Another feature evident in many web 2.0 sites is some type of activity stream. This shows activity of relevance to an individual such as: who has recently connected to whom in your community network, new posts added, comments made by others etc. To mimic this we introduced the notion of a ?Cloudstream?. An individual?s  ?Cloudstream? includes a temporal listing of any new Clouds a user creates, as well as Clouds from any individual or Cloudscapes they are following.

Design Decision 3.1 Add RSS feeds

In line with increasing the Web 2.0 functionality associated with the site, RSS feeds are now available for Clouds, Cloudscapes and people. This enables users to flag only those aspects of the site they are interested in and means rather than having to go to the site, the information can be send to them as an RSS feed and incorporated into their chosen personal digital environment. 

Design Decision 3.2 Integrate streams from Web 2.0 sites

A common Web 2.0 practice, particularly evident in the blogosphere, is the ability to integrate dynamic content from other Web 2.0 sites, often using a ?cut and paste? embed code. Dynamic Twitter, Flickr and Slideshare streams are now possible for both individuals and Cloudscapes. In each case an agreed ?tag? is used as a means of identifying appropriate content for inclusion. For example, if a conference has an agreed Twitter tag #conf09, use of this on the conference Cloudscape will dynamically incorporate all the tweets including that hash-tag. 

Design Decision 3.3 Merge the tag categories

Evaluation of the earlier versions of the site and how tags were being used on it, indicated that users were confused by having three different categories of tag-clouds and in fact were not finding these distinctions helpful, particularly when creating Clouds associated with workshops or conferences, where tags associated specifically with the content of the Cloud and the name of the event were emerging as more natural tags. As a result the tag-clouds have been merged so there is no longer a distinction between pedagogy, subject and tools. 

Design Decision 3.4 Make the home page more visual

Jelfs highlighted in her usability report that the homepage was too busy and not very engaging. Analysis of other feedback indicated that users were not always clear about the scope of the site and what it contained. As discussed above, the newly added Cloudscape facility provided a useful means of engaging specific communities, particularly at workshops and conferences. We wanted to highlight this and hence featured Cloudscapes were added to the front page of the site. We felt this offered the dual purpose of highlighting current, active communities and as a means of illustrating the range of different types of Cloudscapes that could be created.

Cloudworks Future Development 

A second user design was commissioned in April 2009 and a new design based on this was launched in July 2009. As part of this the site was completely rebuilt in CodeIgniter ( The new design provides a much cleaner look and feel and a simpler, more intuitive navigational structure. Initial feedback on the new design has been very positive.

Further enhancing the social aspects of the site is the key driver for the next stage. The success of the use of the site for conferences and workshops is encouraging; nonetheless the site is still not being used in the spontaneous way we envisaged in the original vision statement. We therefore intend to work with a few specific communities in-depth, to articulate their needs and evaluate their use of the site over a number of months. Potential communities to work with that we have identified so far include a cross-institutional community interested in e-learning, a group developing and deploying OER, a pedagogy and research group interested in enquiry-based learning and a support network for careers work and innovation.    

By adopting a reflective approach and not tying down the site in terms of tight specifications a number of surprising patterns of use have emerged. For example, we could not have anticipated at the start of the project the success the site would have in terms of acting as a shared live blogging space.

This post has described a set of design principles which have shaped our development of the site. We have argued that these principles have been derived from our original vision for the site and the associated theoretical perspectives it draws on and that we have used findings from our evaluation work to progressively improved the functionality of the site. We will continue to incorporate further Web 2.0 functionality, trying to pick up the best of social networking practices and appropriate them within the site.

Conferences offered time-bounded events where people are bought together around a shared interest. Cloudworks provides a simple to use back channel to capture and archive the conference discussions. Similarly it works well as a mechanism for capturing discussions during workshops. It is also proving useful as a mechanism for aggregating and discussing resources for a particular community of interest. For example a Cloudscape has been set up to support a group of learners on a language course. We are beginning to explore how the site can be used to support other types of community, as well as looking at ways in which such community engagement can be initiated and sustained.

However the broader vision of a site, where it acts as a conduit for sharing learning and teaching ideas and designs, where teachers upload ideas as a matter of course, and as a back channel drip feeding new innovations, has not yet being achieved and is a much more ambitious and difficult thing to realise. Barriers to this are social and cultural as well as technical. Technically we intend to continue to incorporate and test out Web 2.0 type functionality. We will continue to run activities and events using the site and intend to set up further evaluation studies to tease out the social and culture barriers. We also intend to work with specific ?champion? communities to explore how the site might be used to meet they needs. 


One Response to “Cloudworks: design decisions”

  1. » Blog Archive » Cloudworks: theoretical perspectives Says:

    […] the paper I presented at the Ascilite conferencelast year into a journal article. Whereas the Computers and Education paper we have just published concentrated on the design decisions we have made to date, this paper […]

Leave a Reply