Document Actions

Cosent Blog

Co·sent: shared perception; joint knowledge; collective intelligence.

Social intranet sprint Berlin

Sep 15, 2014 | permalink | intranet · plonesocial · plone

A week of intense collaboration in Berlin has significantly accelerated the development of a new Plone-based social intranet platform.

Social intranet sprint Berlin

The Plone social intranet sprint in Berlin at the Humboldt University was attended to by about 16 participants across two tracks, plus a social program in the evenings. Apart from all the hard work we also had a lot of fun. Berlin is a vibrant city and the Ploners who are also Berliners took us to some great restaurants.

strategy track

The strategy track saw enthousiastic responses to a plan for joint investment by ~10 Plone companies into a new Plone intranet software suite. We'll use the coming weeks to solidify the momentum for this initiative and try and convert positive intentions into hard investment commitments.

A key part of the proposed plan is a design-first process to create a compelling user experience, leveraging Plone5-compatible frontend technologies. Netsight, Cosent and Syslab have already made an exploratory first iteration with this design process, focussed on re-designing the microblogging and social activity stream interactions in Plonesocial.

coding track

In the coding track, we've taken these new frontend designs and implemented them on top of the existing Plonesocial code base. The result is, that we now have a Patternslib based Plonesocial implementation that can be installed and run in a Plone 4 installation. Because Plone5 Mockup is also Patternslib-based, the work we're doing is forward compatible and will be easily portable to Plone 5.

In addition to implementing the existing Plonesocial features, the sprint also resulted in the integration of plonesocial.messaging (private one-on-one messages) and new "reply" functionality (conversation threading).

The new frontend introduces a host of new features that are not yet provided by the backend and need to be architected and coded: file uploads, URL and file previews, "@mentions", "liking", "favoriting" etc. Also, we're already working on extending the design with a number of subtle but powerful micro-interactions in the form of shortcodes to provide a pluggable linking system.

The current development version of Plonesocial is more advanced than the last released version, but it needs more work before we can make a production-quality release - so come and join us in the next sprint at Plone Conference 2014 in Bristol.

Permalink »

Plone Intranet front end

Aug 04, 2014 | permalink | intranet · plone

Front-end development for ploneintranet has started.

While many people are still enjoying the beach, we're already gearing up to accelerate design and development for Plone Intranet this fall.

Following the Mosaic sprint and a ploneintranet summit in Gatwick we're now collaborating with with Syslab and Cornelis Kolbach. to turn wireframes and high-fidelity designs of social interactions into HTML/CSS/Javascript prototypes, using Patternslib (similar to Mockup).

This is a screenshot of some of the frontend work in progress:

social interaction prototype

We'll extend this work in the run-up to the Berlin sprint in September, which will be focussed on extending plonesocial and integrating it with ploneintranet.

Our goal is to provide a workable code base, that documents much of the design decisions we're making in a way that enables community collaboration.

The backend code we're working on is already open and published: see e.g. experimental.securityindexing, ploneintranet.workspace and collective.dynamicmosaic. We expect to fully open up the frontend code before the Plone conference this fall.

Please contact us if you're not a member of the Plone Intranet Consortium and would like to get involved and gain commit access to our development repositories.

Permalink »

Plone Open Garden 2014

Apr 29, 2014 | permalink | plone

The Plone Open Garden event in Sorrento, Italy, is reliably a highlight of the year to look forward to.

This year's edition was no exception. More than 50 Plonistas, wives and kids (and even one mother-in-law) included, gathered to renew friendships, lounge in the sun, discuss arcane technologies after midnight, and generally have a great time together. Oh we also had technical presentations every morning.

PLOG 2014 group picture

We talked about intranets and ways in which we can jointly strengthen Plone as an intranet platform. Netsight and Cosent outlined their research and development timeline for the coming year and worked with other Plone companies to maximize community involvement.

A recurring topic this year was the question, how we can modernize the page layout engine for Plone. We already have a lot of machinery to manage layouts in the form of portlets, portlet managers, viewlets and METAL macros. In addition we have the newer blocks and tiles to further complicate the picture. The discussion oscillated between:

  • Let's stick with portlets. They are a proven, powerful and widely used technology.
  • The portlets machinery is wickedly complex and too burdensome.
  • The content + slots + portlets page model is overconstrained, a responsive grid renderer would be better.

Being technologists, we did not spend much energy on the first two points which are mostly about opinion. Rather, we focused on the last point which presents technical challenges. The gist of what we discussed can maybe best be expressed by a story:

A editor opens a page. On the "display" menu she chooses "create new layout". A layout editor opens and lets her place and arrange tiles on the page. For each tile, she defines a policy of when (context, view, ...) and where (priority, position hint) to show this tile. For the layout as a whole, she defines a policy where this layout should be used (context, type, subtree, ...). She checks previews of the layout for various display media (desktop, tablet, mobile), tunes some tile placements and then applies the layout.

This is just one possible scenario and it will likely change. To explore the possibilities we will get together in Barcelona in the second week of June and sprint to create a proof of concept.

Permalink »

Designing an open source social intranet

Mar 26, 2014 | permalink | intranet · design · plone · open source

Cosent and Netsight are designing an open source social intranet platform.

Netsight have invited Cosent to collaborate on designing a complete social intranet software suite, to be developed in collaboration with the open source Plone community.

A "Plone Intranet" summit in the wake of the 2013 Plone conference listed user experience, that is: design, as the single most important challenge to tackle if we want to strengthen Plone's attractiveness for the intranet market.

As everybody knows, design is not a problem one solves in a committee. We're using a hybrid model of collaboration styles that allows us to combine the design strengths of a core team with the scaling capabilities of an open source community.

The plan is to donate the work we've been doing to the Plone community, so come and join us at the Plone Open Garden 2014 in Sorrento to get involved.

So, what have we been up to?

Last winter Cosent published the Digital Workplace Technology Roadmap.

Following up on that, we've been analysing the competition, both in terms of the user experience their platforms offer but also in the kind of problems they solve, i.e. what markets they're in. We see significant market potential for a Plone-based solution.

Additionally, we've analysed dozens of cases studies of award-winning intranet designs and have clustered hundreds of intranet screenshots to understand common functional areas, or landing pages, in intranets, mapping those against the model provided by the Digital Workspace Technology Roadmap.

Last week, Netsight and Cosent have been sprinting to turn the insights gained from all of that into actionable designs, that can be used to guide software development.

We selected three types of landing pages in intranets for deeper investigation. For each of these pages, we brainstormed specific functions that users would want to use and card-sorted those into families of similar functionality.

Matt and Lewis at the whiteboard

We then picked a single landing page to work on and created several epics with short scenarios about a typical sequence of actions a user would execute to obtain a specific outcome. For example, one of our epics is:

(Team Member) Wendy receives an email from Peter with a list of questions and data that need to be collated before the next meeting of the project board. She forwards the mail into the intranet, where she flags it as a todo for next week on Project X, tags it as "board meeting", adding a note with some initial ideas and could @marcella maybe share her thoughts on this?

For each epic, we created a diagram that sequenced every function invoked as part of the scenario, and then expanded each function step into a full-fledged user story. For example, one of the steps halfway the above epic is the following user story:

Team Member can mention other Team Members in the note (using '@' syntax).

Don't shuffle the stack

Fleshing out those user stories was a lot of work, and involved detailed discussions about our assumptions and choices regarding security architecture and overall strategy. This was done by part of the team, while the other half worked on wireframing possible solutions for the epic. That was a bad idea. They had the same discussions, with different conclusions.

elements of user experience

Moving from epic to wireframing involves jumping a level up the design stack in the Garrett five-level model of the design process. When we brought the finished user stories together with the wireframe sketches we had some major inconsistencies. This appears to confirm the model and indicate that you need to get your foundations right before moving to more concrete designs. In this case, you really need to define your scope in detail, before wireframing solutions.

After re-syncing our minds and merging our work, in the final day we ventured into wireframing territory not for a whole page, but for exploring a set of micro-interactions that form the core of a cohesive social intranet experience. We also deepened our understanding of user needs and elaborated on the personas we're using to drive the design.

All in all, we feel we have not only made significant progress towards valuable design outcomes, but also have prototyped a repeatable design process that tackles very complex design challenges in a systematic way.

We plan to have another design sprint in a few weeks to prepare for Sorrento, and look forward to sharing our work there. See you in lovely Italy!

Permalink »

User Centered Re-Design

Mar 13, 2014 | permalink | design research · design

A website redesign can greatly benefit from combining User Centered Design with an analysis of the existing web site.

User Centered Re-Design

User Centered Design is a methodology Cosent uses to systematically involve end users in the design process. It improves the quality of our designs by maximizing our understanding of the intended users of systems, what their goals are and how the site we're building fits into their lives.

In a recent project, we found that this method is especially powerful when you're redesigning an existing web site. Users who have actually been using the old site develop strong judgments on what does, and especially what doesn't work for them.

You'd like to drop dead when you open this site.

Sentiments like these are very helpful when a client needs to be steered away from imposing bad design choices on a web site. You can show the client how that didn't work in the past.

Combining user centered design with an analysis of the old web site design highlighted some major problems: inconsistent navigation, too much navigation, too much graphics, noisy page layout. It's not that the previous designer did a bad job. When working on the design for the new site, we found that the design patterns of the old site were a "natural" consequence of the features requested by the site owner. We could've easily used those same patterns our new design.

Except, our user research told us forcefully we really needed a different approach.

As a result, we focused on simplifying the navigation and visual layout. We moved from a complex multi-level categorization to a simple category menu, augmented with free-style tags which are shown only in the content area, not in the navigation bar. We consolidated multiple web sites into a single consistent system. Instead of showing lots of small thumbnail images, we used varying image sizes to grab attention and structure the page. Finally, we used an accordeon in-page navigation to reduce visual clutter but at the same time enable web site visitors to quickly orient themselves in the site.

The result are these twin cultural agendas: UIT Sittard-Geleen and MaastrichtNet.

Permalink »