We have met over 90 colleges during the past 5 months and presented Kloodle to a range of staff members; from careers advisors to personal tutors, from senior tutors to principals. The college staff we meet provide invaluable feedback, and we often return to the office with an array of ideas to bake into our product.
We used the summer holidays to overhaul our product and include these ideas. We started to write these features on top of our existing site, but it became apparent that a restructure was in order. Our development team sat down together in early July, gathered around a whiteboard, markers in hand, to thrash out the features we wanted include into the new Kloodle.
This early meeting was important, as it allowed us to plan the structure of our new site. Lots of features were applicable to our many user types; it was important that we made these features as generic as possible to avoid repetition and making mistakes. The old version of Kloodle was built like a house with an extension; our basic structure had features tacked on top, which had led to repetition and bulky code. Our new product would be much sleeker and developer friendly.
The initial version of Kloodle was coded from scratch in PHP. When discussing our overhaul, we discussed whether this was the route to follow again, or whether to use a framework. We decided the flexibility and ease of use a framework offered was important, as well as that new developers to our company could quickly orientate themselves with the code. We selected Laravel as our framework of choice, and proceeded with gusto.
The initial task was to acquaint ourselves with Laravel. Our developers had not worked with this framework before. We bought subscriptions to Laracasts and had a week’s worth of orientation, building basic Laravel sites as part of the learning process. Our Google search count went through the roof, as we explored the limits of our knowledge. Laravel proved to be a fantastic choice, as we got up to speed extremely quickly. Laravel’s easy structure, community support and fantastic packages provided great assistance throughout this phase.
The main development phase started towards the end of July. We set ourselves 3 weeks to build the bulk of the features, with 2 weeks’ worth of testing to iron out any kinks. The initial task was to build the ability to post content. Kloodle allows users to post statuses, blogs, photos, videos and documents to their profile. This is a generic feature to all account types. Laravel made it simple to get this up and running.
The complexity of Kloodle comes from our different user types: we have organisations, staff, students, employers, colleges, and universities. All these relationships needed mapping out properly and determining how each user type would interact. This aspect of development caused the most headaches and we spent the majority of our time in these early stages fathoming out how this worked.
Once this was elucidated, we hammered away at the various features: profiles, personal statements, action plans, connections, following, messaging, groups, newsfeeds, and finally our registration and login system. Development of the new version ate into nights, days, free time and weekends. Each new feature perpetuated our enthusiasm as we wanted to see the fruits of our labour working and being used.
We conducted our first test day with the staff at Northampton College. The site was basic at that point, and we were eternally grateful for the staff’s ability to use leaps of their imagination and stick with our train of thought.
We ironed out the kinks uncovered during that testing day, and forged on towards the next testing day; a “pizza and code” day at our offices with students from Holy Cross college. The day started with stress, as GitHub decided to play shuffle with our merges, rendering at least some aspect of each developer’s work deleted. After fighting with rollbacks and file merges, we got the site to a testable state, and proceeded to uncover a raft of improvements. We are much obliged to the students for their thorough testing; each one was sufficiently reimbursed with pizza for their troubles.
After the testing day, we had 4 days to finalise the site for our first cohort of signups. These 4 days tested our resilience, as long hours, breakages, deletions and stress were regular occurrences. We pushed the final polish to our main servers at 12:50am, and our first sign ups took place at 10:30am that day.
To turn around a complete site in such a short time frame was a Herculean effort. Our development team were maxed out, and their efforts have contributed to what is a sleeker, more effective and all-round better Kloodle. We look forward to users reaping the rewards!