Comment by yerushalayim | Hacker News Reader

Comment by yerushalayim | Hacker News Reader

Parent Full thread

You’re right that a lot of software comes out of Anglophone or Western contexts, but that’s exactly why these issues persist. The problem isn’t that RTL is “hard” — it’s that most text engines, layout systems, and PDF toolkits were originally architected with implicit LTR assumptions baked deep into the rendering pipeline.

Once those assumptions are embedded in things like glyph ordering, bidi resolution, cursor movement, hit‑testing, line breaking, and font fallback, fixing RTL becomes a retrofit instead of a design principle. By the time a team realizes the gap, the shaping and layout stack is so tightly coupled that adding proper bidi handling feels like a massive rewrite.

You see this pattern everywhere: PDFium (used by all Chromium browsers), various UI frameworks, and even some OS‑level text components still mishandle RTL in 2026. The symptoms are always the same — disappearing text, reversed glyph order, broken cursor navigation, or failure to commit text at all.

This isn’t a niche corner case. Hebrew, Arabic, Farsi, Urdu, and other RTL scripts represent hundreds of millions of daily users. The real issue is that global language support is still treated as optional rather than foundational, and the technical debt from those early assumptions keeps compounding.

Even basic local names in Western contexts (with characters other than just [a-zA-Z0-9_\ ]) cannot be properly and consistently input or searched in most of corporate or retail software, at the basic text level, because they rely on broken assumptions that even modern ERPs still follow. I routinely sift through 3 or 4 different spellings of my name every time someone asks for my name in their systems.

I cannot imagine the nightmare it must be for non-western languages.

I don't know exactly what the real, hard incentive is to make it happen at last, as this needs a strong perspective over software as a tool to serve people, as well as some kind of artistic literacy: we need more people to care about the tools they build, and more people to pay the makers because of that. Steve Jobs, with all his downsides, had this kind of focus and impact. But this needs to be systemic, not exceptional.

You’re absolutely right that even “Western‑adjacent” names break in systems that were never designed to handle anything beyond ASCII. It’s no surprise that entire writing systems fall through the cracks.

Same goes for Steve Jobs. One of his most underrated contributions was his insistence that typography, calligraphy, and the aesthetics of written language were not decorative extras but core to the human interface. Apple invested early in system‑wide text rendering that treated all scripts as first‑class citizens. It wasn’t perfect, but it showed what happens when leadership actually cares about the universality of writing and makes it systemic.

>things like glyph ordering, bidi resolution, cursor movement, hit‑testing, line breaking, and font fallback

I think part of the problem is that things like this are often just barely working on the one set of test data that often just barely covers the fundamental requirement in its most simplified "happy path" form to begin with. LTR in this case.

And then the coders move on to get the next objective barely working, but nobody ever goes back far enough to reinforce the weak points in the fundamental structure before it's a far more challenging task. And then if done it's still too challenging to go all the way back.

I've gotten into the habit of replacing the accented "é" in my first name with a plain ASCII "e" in forms to avoid troubles. The worse part is that the form is usually accepted, then later on you encounter random issues (cannot log back in, ...).

Reply on news.ycombinator.com