The first challenge for Fabrika was to
create the design for a social media platform
based not on visual input but rather audio.
Since the idea was revolutionary, it was decided to combine it with a look and feel that would be familiar to users, so the overall app flow and user experience was set to mimick Instagram. However, in Poetiv images were not meant to play a role, so the poems at first were indistinguishable visually one from another. This prompted the need to create a visual cue related to the poems, to make them easily recognizable and visually diverse. We thus designed an amplified sound wave out of the audio and an image that would fill the empty space in the sound wave, creating a background gradient based on the image in the post tile. User feedback showed that this special effect had the intended outcome - every post had a very unique look; users were able to easily identify their favorite poems, to differentiate between postings when scrolling, and the visual helped users focus while listening to the poems.
The iOS team had a challenging task creating the poem upload and player feature, which
would allow recording, uploading sound, instantly creating a sound wave and finally uploading
the sound wave fill image with a background gradient.
The first version included audio recording only, but it became clear that most users also wanted to upload pre-recorded audio. iOS 11 featured a new component, which enabled audio file downloads from the users’ iCloud. The app then had to instantaneously create the audio wave image and image upload. The whole process took place on the local device and once all assets were ready they were uploaded to the server in the background while the user was sent back to the home screen. This generated a very sleek and seamless upload experience.
The app also works in the background and offline, so users can listen to poems anywhere and anytime.
Fabrika decided to build a backend in a very scalable
way so as to cut out the need for DevOps in
the medium-term and to decrease server costs. We wanted to challenge ourselves; so we built the
first social network using a Serverless approach, with AWS Lamdba functions.
The biggest issue facing us with the use of AWS Lamdba were cold starting functions which results in delays. To avoid this issue in the beginning when there were not a lot of users and some functions might have been left idle for an extended period of time, Fabrika built a script to “call” each function after a certain time interval to avoid it having to start with a delay for users. The results were indistinguishable from from using dedicated servers and continue to have low costs for the client.