惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

WordPress大学
WordPress大学
大猫的无限游戏
大猫的无限游戏
B
Blog
阮一峰的网络日志
阮一峰的网络日志
IT之家
IT之家
Hugging Face - Blog
Hugging Face - Blog
博客园 - 【当耐特】
Jina AI
Jina AI
博客园 - 聂微东
T
The Blog of Author Tim Ferriss
宝玉的分享
宝玉的分享
L
LangChain Blog
M
MIT News - Artificial intelligence
Blog — PlanetScale
Blog — PlanetScale
腾讯CDC
酷 壳 – CoolShell
酷 壳 – CoolShell
Y
Y Combinator Blog
F
Fortinet All Blogs
H
Help Net Security
B
Blog RSS Feed
J
Java Code Geeks
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
Apple Machine Learning Research
Apple Machine Learning Research
S
SegmentFault 最新的问题

Hacker News

GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis Bonsai 1-bit WebGPU - a Hugging Face Space by webml-community Moving a large-scale metrics pipeline from StatsD to OpenTelemetry / Prometheus GitHub - Nightmare-Eclipse/RedSun: The Red Sun vulnerability repository GitHub - SethPyle376/hiraeth: Local AWS emulator focused on fast integration testing, with SQS support, SQLite-backed state, and a debug-friendly web UI. GitHub - macOS26/Agent: Any AI, replaces Claude Code, Cursor, OpenClaw. Over 18 LLM providers (Claude, OpenAI, Gemini, Ollama, Zai, HF, Qwen) wired into a native Mac app that writes code, builds Xcode projects, bumps versions, manages git, automates Safari, use AppleScript, JS or Accessibility, extend Agent! w/ MCP Servers, run tasks from your iPhone via Messages. YouTube now lets you turn off Shorts I Made a Terminal Pager Burgers | マクドナルド公式 Commands — HackerNews CLI documentation ChatGPT for Excel PiCore - Raspberry Pi Port of Tiny Core Linux Live Nation illegally monopolized ticketing market, jury finds Google Broke Its Promise to Me. Now ICE Has My Data. Founding Engineer at Adaptional | Y Combinator CRISPR takes important step toward silencing Down syndrome’s extra chromosome GitHub - saffron-health/libretto: The AI toolkit for building reliable browser automations US v. Heppner (S.D.N.Y. 2026) no attorney-client privilege for AI chats [pdf] Retrofitting JIT Compilers into C Interpreters IPv6 – Google The Accursèd Alphabetical Clock Cybersecurity Looks Like Proof of Work Now Fragments: April 14 Cal.com Goes Closed Source: Why AI Security Is Forcing Our Decision | Cal.com - Scheduling Software for Online Bookings Laravel raised money and now injects ads directly into your agent When moving fast, talking is the first thing to break Too much Discussion of the XOR swap trick – Heather Cafe Introduction to Spherical Harmonics for Graphics Programmers The Grand Line
Map Clustering is Not My Favorite
2026-06-13 · via Hacker News

Friends, this post has been in the making for maybe, 20 years.

Google Maps (and me) are (at least) that old. And mapping, of course, existed before that; those were the days.

But ~~~20 years ago, or like a year after that, something appeared on our collective mapping radars that I have not been able to shake off since - and which I haven’t kvetched about in writing, ever, for everyone to “enjoy”. Enjoy! It’s a Greg Rant! Amazing…

What happened ~~~~20 years ago, after Google Maps was hoisted on us all, was that “people wanted to see more than 1 point on a map” and sometimes that meant 100 points, or a thousand! And things were not easy or good then. I mean, 1000 points, that’s a LOT.

But then, who said 1000 points (was it you?) was too much to show? Well… browsers? Ancient, decrepit (err, just not-recent) browsers from back then didn’t necessarily struggle inasmuch as they begged/hoped you wouldn’t create a literal DOM element for every single dot you’d want to add to your map.

But that was what Google Maps did, back in 2005-6-7-8-9 (they don’t do that anymore)

Adding 1000 points back in 2005-6-7-… meant adding 1000 DOM elements to your map, and each of those DOM elements had more DOM elements in them. If you are the leetcod’ing type, you’d say something like “O of N!!!” and you’d be right - go back and study now.

And so, hence, we all, mappers, mapping-people, map-folk, accepted/self-ingrained that “you better not show too much information on that map! It might break!” and discussed/shared strategies around how to do that. It was fun. We didn’t have Instagram.

I still remember debating (and disliking) paging - you’d see 10 things on the map, then click “next page!” (a page of map?) and then see 10 other, different points. Today, I would say: not good UX.

Or you’d have.. heatmaps? Or… you’d bake the points into images and serve those images? You could look up what WMS is, but you don’t have to.


Chapter II.

There once is a beautiful - truly, without irony - web site called Atlas Obscura - which imparts on us where interesting things exist in this world. Oh, look - a beautiful and unique place! It’s here! And here’s some history about it too. Atlas Obscura is great.

But/however, I will be using as an example of what I’m talking about here - what I have a problem with - is their/this map:

This map.

It is a map of 30 thousand+ places - all on one map - where the map shows every one of those places - but clustered.

Clusters being these circles with numbers in them. Each cluster … hides&groups more items within in. To discover those items, you must zoom.

Hence, frustration number one - clusters, existing alongside markers, don’t tell you anything. Clicking on them zooms in the map… because… that’s a thing you must do… to reveal the cluster’s content, maybe. It’s like a kindersurprise, it’s a chocolate and maybe there’s a fun toy inside, but you can’t see the toy (the chocolate is not transparent). You are at the outside-the-chocolate stage in your life. You must get closer to the chocolate shell to get inside and get to the toy, which right now, you can’t see, because of the chocolate. You follow?

And hence, you click. But the clusters, also, sometimes, will not/never reveal what is inside of them. Because… if a cluster clusters multiple points that are either very very very close together (or even, it’s really multiple points at the same location - it happens), the cluster will remain a closed box forever. ((caveat, this is still a problem with non-clustered maps))

It’s like being gifted a matryoshka, but you’re not allowed to open it - which would defeat the point&pleasure. It would not be a nesting doll at all at that point and just be a sad present.

And ok, point two - the goofiest - is that sometimes/a-lot-of-the-times, zooming onto a cluster, at the last moment, reveals points that are not at all where the cluster is. You zoom in and boom, it splits like the atom and you see a point on the left and a point on the right. It makes sense! They were clustered and “averaged” together. But it’s terrible ui, frustrating, surprising in the worst of ways.

But you know, these above points are not the one(s) that matter to me/you/the world.

The truest reason - the most dandy, reason to hate clustering is that reading a bunch of numbers on top of a map - is not a thing that the eye/brain does well at all. Staring at a map covered in digits is a bizarre passe-temps and is not productive, efficient, etc. - it is, simply put, a strain.

Map clustering solves a problem (too many points, browsers not liking that) - ok - maybe - but also - that problem is not a problem anymore.

Because… mapping tech in 2023-4-5-6- is incredible, fast, and based on video game technology. Modern maps (on your phone too, oh so much) are gems - technological GPU+shader codes meant to render infinitely fast, in parallel, smartly, economically. They zoom and pan infinitely extremely fast - yes, 60fps, yes 120. More. It doesn’t matter. We are post-fps. And definitely post-DOM. Things are fast now.

You don’t have to cluster anymore. You can just be.


Chapter III.

I downloaded the Atlas Obscura data and used Protomaps to make a map to convey to you - in more ways than one - the “points” of this article.

One of the (but not the default) mode(s) just shows you all the freaking points. The map remains fast and breezy, because MapLibre GL JS is a gem, and everyone involved did great amazing work. I love everything about it (not the map I made necessarily, I just mean- the total human effort that lead us to this day where this is so easy, so fast, so beautiful).

But also - the map plays with these notions of clustered/too-muchness///well-we-can’t-just-show-everything-on-a-map-can-we’ness, and other concerns. I think that a slightly not-opaque dot overlayed on top of other dots (this mode - the default) is maybe the neatest. It conveys density, lets you see points. The hover is satisfying (on desktop only, yes). I don’t advocate heatmaps for everything/everywhere.. but maybe a hint? I don’t know. Do I even know what I like?


Chapter IV.

Atlas Obscura did nothing wrong - they are great! - it’s just that the day has come for us to put clustering behind. It’s not evil, it’s not garbage - it’s just done. It’s over. We have no need for clustering - for all its faults, it was useful to us 20 years ago and then maybe 15 years ago too and maybe 10. But we are post-cluster now. We can be unclustered, in public. We can be ourselves! Burn your clusters. Show us your points.