July

Latest survey results + various under the hood improvements

Yet another month has passed by in a flash. So I am here to bring you the latest update on all things Kenmei. This month have been sparse with noticeable updates, due to two reasons:

  1. Main features in development require a lot of design and experimentation to be completed, so it's uncertain how long they'll take and it's quite an intensive work, compared to features that are more straightforward to do.

  2. The backend required a long-overdue improvement in how supported sites were stored. That's delicate work, as it touches on almost all of the Kenmei codebase, but I am almost done and it will help tremendously with adding new sites as well as having better control over currently supported sources.

But let's dive into things step by step. First of all, I'd like to take a look at the latest survey that got concluded.

Thank you for participating in this survey! It's very useful for me to understand how satisfied everyone is with Kenmei and what features are most requested. I make sure to work in a data-driven way, so these surveys are a must.

First of all, when it comes to satisfaction, it mostly stayed the same, with even less negative reviews. Which is great, but I am hoping to have a lot more people rate the experience as 5, after many of the requested features get implemented. Those of you who voted 3 and 4, do let me know on Discord if it's just missing features or something else making you not rate Kenmei higher.

It's also great to see that a lot more people are using Kenmei now. With a similar number of people responding, it's an over 20% increase in usage, which is great to see just after a couple of months.

Finally, the most interesting bit is the most requested features. While I was expecting Dark Theme to still reign supreme, it turned out to be 3rd, after being able to set chapters manually and automatic tracking as you read. I have adjusted my roadmap, to make these a higher priority than other features that I planned.

I wanted to take a moment to shed light on improvements that took up most of this month. Feel free to skip this section, if you are not that interested in the more nitty-gritty details.

For starters, let me explain a bit on how it currently works. Every supported site on Kenmei API gets a designated API (similar to Tachiyomi or Paperback extensions if you are familiar with those), which encapsulates the logic of getting all the data I need. Of course, one of the pros of Kenmei is that you, as a user, don't have to know about these (unlike Tachiyomi/Paperback, where you manually download and update extensions). So I can quickly fix them up, if something goes wrong, without you having to do anything, perhaps just seeing a brief downtime. I also have a sophisticated generator for new sites, which makes me create these APIs almost completely automatically, making it quite easy to add new sites.

But, as good as this system is, it could be a lot better. This approach takes up a lot of code, which it doesn't have to. And with 14 sites supported, I knew it was time to improve this system so that we can scale to 100+ supported sites, without adding tens of thousands of lines of code.

So what I've been doing this month, is move a lot of this information into the database, instead of having it in code, re-writing Kenmei API to figure things out on its own. I'll spare you the details, but this approach brings lots of new improvements and future possibilities:

  1. If any of the sites are down, I'll be able to deactivate them instantly (eventually instantly showing it in your dashboard, for manga that has a site issue, prompting you to switch to a new one), while previously I had to deploy code, which would take time.

  2. I'll be able to scale to 100+ supported sites, without adding ungodly amount of code. This will make maintaining sites a lot easier as well as making them less error-prone

  3. Eventually, I envision an ability to add support for new sites, without writing almost any code. At this point, I expect this system to be miles ahead of anything out there (I think it is already quite a bit better than other popular manga apps on this front)

So this work was super important and will allow me to get back to adding a lot of new sources next month. And make even more awesome improvements going forward.

As mentioned, some of the work that went on didn't quite fit into the roadmap as well as a lot of time spent prototyping and experimenting, which doesn't quite translate into a simple completion percentage. With that in mind here is the updated roadmap:

  1. Ability to choose title type globally (English or Romaji) → 50% Complete

  2. Improving data consistency in the database → 100% Complete

  3. Dark theme → 10% Complete

  4. Cover grid mode → 30% Complete

  5. Revamped manga list UI → 70% Complete

  6. Improving mobile experience → 50% Complete

Based on the survey, I will be adding the ability to set chapters manually and possibly exploring the automatic tracking in August. But otherwise, most of my time is going to be put into finishing revamped manga list UI with the improved mobile experience.

Last updated