It's actually difficult to choose best path to develop upfront. This is why refactor is a phase so popular in software development. You need to accept the fact that it's nearly impossible to produce perfect software from scratch - usually it means having several iterations changing approaches, fixing validation with design patterns or updating naming to be most accurate and so on.
Major plan for this week was to start implementing value goals as next feature of the app. However, just before diving deep into the topic I decided to go through my list of goals to simply check what's my structure of goals. And then the lightning just struck me...
This was a week that was designed to become a start of second app milestone - value goals.
When it comes to visual part, I'm always concerned. On the one hand I feel that I can distinct `not-a-bad` from `not-good` designs. On the other hand, my design skills are comparable to my singing. Trust me, you would not like to hear me singing 😱
I'm currently on holidays so this ain't gonna be typical post describing the progress towards building great dev company of one. Holidays gave me no space for working progress, but also gave some space for thinking and doing some cleanups. One of the things that needed cleaning up was ~50 cards in my mobile Safari browser kept `for later`. This list of course grew for months.
I'm actually surprised. What is more, my level of surprise rises every week. What I'm surprised with is the fact that I'm still in the game. I'm still developing the app I planned to develop. I started developing, I had several difficult weeks but I'm still progressing. This week is the week when I can celebrate finishing a first milestone - `daily goal functionality`.
The last 10-15-20% of something is most difficult to reach. It's one of the variants of Pareto principle - the one principle to rule them all. Right now I am observing it with finishing first milestone of the app - daily goal. For last couple of weeks it seems like next week will be the last one before moving on to next type of goal. But it always ends with realizing that something still is missing and requires adding - editing goal, displaying calendar, fixing calendar, fixing calendar again, and again, and again, deleting goal, updating progress bars of goal, updating date switching and so on and so on.
Finally detail view became understandable for my beloved Tester which is actually a great sign. You know, there are two levels of understanding when developing something. First level of understanding is the one in which developer understands what he created. Second level of understanding is the one in which someone being bit outside understands what you wanted to show. Finally, after couple of iterations we reached a level in which detail screen becomes understandable the outside world which is a great thing!
Pause - Stuck - Pause. This three-week series did not sound too optimistic for the future, but finally I was able to move on this week. A task that made me freeze for some weeks was implementation of calendar view presenting completed days in a monthly view. Initially I decided not to implement it from scratch but to integrate a library that looks like matching my needs
Can you feel a bit unsatisfied even if you worked as long as you planed and actually did some progress on detail screen and also fixed some bugs across the app? Well, sure you can, if you overexpect.
Only one hour of week time actually spent on project development is not a perfect way to push the project down the line. But, like it pretty often happens, it's all in the details. Life is not only about one project - it's actually an asynchronous race of multiple threads fighting for resources and the app project is just one of the threads - important one for sure, but not the only one. So it is what it is - detail screen is still under development, but a little progress was still made.
So, this was a very educational week. I dived bit deeper into async/await way of threading in iOS and it actually not only took my breath away but also took significant part of time left for coding. Still, detail screen is under construction, it's already part of app flow. What it's lacking is presentation of data - levels, progress and completed days / strikes. This will be the subject of next week.
It actually took me six days before I figured out how to approach the logic related with levels completion popups. Luckily a week consists of seven days, that is why the last one was the productive one.
Swift Charts for me was the shining star of The WWDC 22 week. Surprisingly, for this week I also planned adding little chart to the app presenting in the list progress in completing goal levels. As the app is for now developed based on SwiftUI in version for iOS15, I did not have the opportunity to test Swift Charts, but I believe I'll use it slightly later. Thanks to modularity of the code it shall be super easy to replace current chart rendering with SwiftUI code.
Creating goals without defining levels you wanna achieve is like participating in race without a finish line. Goal levels are checkpoints clearly showing you if you actually progress. There is no simpler and better way to monitor your progress.
So, the initial plan for this week was kinda different - I wanted to focus on adding goal levels to track progress. However, after some days of testing local notifications I found a bug that could cause duplicated notifications or missing ones depending on moment user launched the app. This is why I started this week trying to fix the bug and also add some unit tests covering logic of selecting time for this notification.
When thinking about development, sooner or later you start thinking about automation of some tasks, processes, tests etc to simply save time when doing the same set of tasks from time to time. And, when you think about automation in world of iOS development, you need to think about Fastlane. This is the one tool to handle app automation in the way you would expect from tool like this in 2k22 - dependencies management, certification, building, testing, distribution - just to mention key areas you can automate using Fastlane.
Once we started gathering data the next step was to introduce series calculator checking best & current series of successes for each of daily goals. Actually it took me some time to find an approach that makes sense but it looks like approach I finally took looks like working 💪 I finally decided to recalculate goal progresses once per a change and store them for quick future accesses to limit unnecessary recalculations. By storing data I mean storing in device memory and also in local memory cache for quick access during runtime.
When building list of goals I decided to use Combine to produce stream of data updates and to reload UI only when it's actually needed.
Inversion of control desing pattern is a programming principle increasing modularity of the software. Introducing dependency injections gives us a chance to quickly replace implementations, if needed, without a hassle and without changing anything in main program code.
Real work on the project has started this week, finally.