When Code Sidetracked a UI Designer: My Journey Through Themes, Input Methods, and Product Experiments

01. Early UI fascination sparked programming interest

Jensen introduces himself as a UI designer whose curiosity for graphic work (“美工”) in high school led him to experiment with early smartphones such as Symbian, BlackBerry, and Windows Mobile, buying second‑hand devices to modify icons, signal bars, and battery displays.

He created a Launcher theme that gradually brought elements of Windows Vista to the phone, writing XML by trial and error after learning the syntax from forum experts.

These experiments marked the first imprint of Windows Mobile on his path, showing how the platform’s decline alongside Android and iOS motivated him to pursue UI customization further.

02. Winning a Xiaomi typing competition with a custom double‑key Wubi layout

During university, Jensen joined a Xiaomi 2 typing contest that required recording a promotional video; the prize was a Xiaomi 2 phone.

He leveraged his five‑stroke input practice to design a double‑key Wubi skin that combined the compactness of a QWERTY layout with the low ambiguity of T9, improving typing efficiency.

After two weeks of intensive practice, he achieved a peak speed that beat all opponents, winning the championship by an 8‑second margin; the runners‑up used voice input.

The double‑key layout also featured a 17‑key double‑pinyin scheme that merged QWERTY‑based initials with combined finals (e.g., iao/ao, en/eng) and displayed available finals after selecting an initial, with an optional stroke‑based filter to further narrow choices.

03. Discovering a market in Xiaomi theme development

Recognizing Xiaomi’s paid‑theme policy, Jensen began crafting a series of MIUI themes, emphasizing lock‑screen animations, rounded‑corner icons, and dynamic wallpapers.

He used the MAML DSL (a XML‑like configuration language) to build lock‑screen modules, even embedding WebSocket support so that themes could fetch remote data.

He later built a simple content‑push system on his own server, allowing themes to poll for notifications and display them on the lock screen, with clickable links to WeChat, QQ, browsers, or the theme store.

Despite positive user feedback and several themes reaching daily rankings, the Xiaomi platform later disabled external WebSocket endpoints, rendering the push feature infeasible.

04. Reflections and the “Magic Screen” concept

Jensen summarizes his sales: from August 2012 to the present, his themes generated 1 933 911 units sold and ¥43 151.78 in revenue, representing his first “bucket of gold.”

He reflects that UI design no longer adds a “halo” to his career, as he moved into software development rather than pure graphic work, yet he still values pursuing personal interests.

He concludes with a brief note that the “Magic Screen” product idea was halted when Xiaomi removed the WebSocket entry, illustrating how platform policy can abruptly end technical feasibility.