Qvitoo

When we started working on Qvitoo, we wanted a mobile-first experience. So we went down that route, with a fluid layout that adapts to the viewport size.

It turns out that when we need to build many other features, actually sitting down to media-query your way to a nice desktop experience takes a bit of a back seat.

Finally, with v2 being released (with mobile-first) we've had time to circle back and improve the user experience for desktop.

What did we learn?

Mobile first was a great way to go; it's much easier to have a stretchy layout behave well on a large screen than having a constrained layout loosen up for a small screen.

In fact very little was required to turn the mobile app into a desktop app. Each insertion reported below is a line of code added, and the same for deletions.

For our CSS/stylus:

3 files changed, 13 insertions(+), 6 deletions(-)

And our JS/HTML:

3 files changed, 29 insertions(+), 3 deletions(-)

It's taken me 2 work hours from idea, to doing the above changes, testing them, and releasing them to production.

In this case, we should learn that some fruit is lower hanging than other. We hadn't prioritised the desktop feature as high as others in our Kanban, but I had a hunch it would be easy and potentially very useful. It turned out to be much easier than I expected.

Finally we've released v2 of Qvitoo at https://qvitoo.com/ with much improved everything! :)

It's been a really fun experience programming, designing, sketching, discussing and crafting this version. A few highlights:

• Brand new user experience design
• Moving from plain React to React+Redux+RxJS has made the app much more responsive and easy to write
• Moving from a micro-service architecture to two main services has cut down on release overhead
• We've replaced the backend code with an in-house stream processing engine, so we can work with time, side-effects and state explicitly, rather than implicitly.

With us you can quit worrying about your accounting, too.

Qvitoo is a bootstrapped/self-funded company. As such, the financing has been done by consulting as well as from subscriptions from users of our service.

In August we had a really tough time; the product was almost finished, to the extent that it did all the required functionality, but there was still sharp edges. Also, refactoring some parts of the UI had made some functionality unusable.

August 2018 was a month full of grind; fixing one small thing and then moving onto the next.

After finishing all the features we needed it was time to do the deployment. We're hosted on GCP, so we created an organisation and rebuilt all of deployment from the ground up. This was a good decision, as we got the opportunity to improve security in places, separate concerns and generally get to grips with the platform we'd be hosting on.

Furthermore, separating production from staging allowed us to put a granular permission model for operations in place, to make sure we have the right permissions for the right people.

If we learnt something from this; it is not to give up, but to continue grinding it out, until we're done.