0xCE0

I love Fossil, I love SQLite, and I also like Althttpd.

https://sqlite.org/althttpd/doc/trunk/althttpd.md

Just like Fossil vs Git, SQLite vs $SomeRealSQLServer, I wish someday Althttpd would become a no-bullshit self-contained replacement for Nginx/Apache/whatever bloated HTTP servers. It has already proved its working by serving Fossil/SQLite, but configuration/features for serving actual web site is not yet "real production quality", at least that is how I feel.

Overall, what an amazing legacy this set of software has been to the world.

Real life is the new forum of 2026, just like Roman Forum 2k years ago. You can see, feel and communicate with the reality as it is. Go find people and talk to them, and maybe you will meet someone with whom you can have a discourse of your life. Observe how the emotions of yourself and the other person(s) change during discourse, because they reveal hot points: limits, pain points, dreams, un(knowingness) etc. Learn/listen and teach/speak.

I don't like at all. My experience for iPhones is iPhone 6+ (which was amazing), then I've been at Android world (Sxx series), and now I have iPhone 17 Pro Max, and truly wonder what is the value proposition here. Feels really buggy, UI tries to be Jolla-like gesture-based, but fails to do that, default style reminds that it is made by/for a kindergardener. And there is no "home" button anymore, that resets my virtual position to something safe/familiar, now I'm trying to swipe up/down/left/right, and there is always some new screen layout.

On the macOS side, it gets more locked and buggy and inconsistent with every release.

This is just a general comment for Apple OSs today, not specifically version 26.2.

These kinds of cases just triggers all the rage for me, be they true or not and whatever is the actual case.

I have fresh experience of setting up Azure/M365 and AppleDev for my startup. Those things are scary as f*uck, in many perspectives:

(1) Dark patterns everywhere (click this checkbox and we'll buy you a license, oops +xxxxx €/$ per year just came; get one-month trial for O365 to get bizaccount, select 1 license, see that there is 25 licenses (~ 4k €/$) to be renewed if I don't cancel).

(2) Microtransactions everywhere (e.g. Azure VM SSD I/O: every read/write operations costs), DDOS and 10/100 k€ bill coming. Everything "scales", especially bills. And no billing caps, of course.

(3) Codesign with Microsoft: I have option to wait weeks for freight ship to ship USB cert token (if it ever survives past toll/postal service after that), or use AzureKeyVault, but that is officially only for companies that has taxes/accounting for 3 years of operation. So no startup can use that by this requirement to codesign?!

(4) AppleDev (and kind of Azure/MS too) requires DUNS number, which takes 6 weeks to get in normal case. Apple's 5 bizday route doesn't exist anymore (at least not for non-US-based companies). Or just use D&B magic link from Grok and get it immediately in 5 mins.

(5) If you base your business on Azure/M365 and AppleDev and be obidient and compliant (as I am doing/being, because I'm building real legit and long-term company, not some hussle project), it still doesn't matter, because they can just can decide by human/ML to shut your business operations and means of living. And getting answers like in the title's article's screenshots with those emojis are just the most non-human interaction that there can be done for affecting so devastatingly to someone's life/business.

These are the most disgusting things that I know of.

Make sure you eat your own dog food by doing your own company's accounting 100 % with your own product. Because if you don't dare to trust fully your accounting on it, why would anybody else.

Excel is a real CEO/CFO-grade tool, because the mental/simulated model stays local, stabile and personal, Excel rarely corrupts data (never happened at least for me, even with massive file sizes) and you can just generally trust it to get the needed job done with it whatever it is.

Yeah, I can't quit Excel. Every day there are spreadsheets open in my taskbar, even right now :)

This.

Product-Market Fit, i.e. finding the point where your product/service resonates with the market.

I know what you mean by charting/mapping the problems and their solution. But I think/feel this method doesn't work anymore, it feels like "2012-2022 world". How can one sell something that doesn not exist even by the seller? You are selling just the idea, not the implementation. Of course that could lead deal also, but it's more of people-to-people trust/business, than about real existing products.

And now in the "post-AI" reality, trusting anything people sell is hard.

Booting up startup (even the startup needs to be started, it doesn't start by itself) is a constant struggle. Almost every simple and basic sounding action/operation/step causes heavy reaction/resistance by the reality. The reality just doesn't want you to succeed.

I am lucky to run my startup with my own funding, and it causes me to be very responsible with spending.

Finding PMF is the most mysterious thing I have ever encountered. How is it possible to sell anything in this world? People who can sell are the most valuable people in this world.

I guess I have had "burnout" so long, that I cannot anymore feel it. It has now become the feeling of accepting and observing the reality as realistically as it is. "It's just life.", says my friend when reality slaps/hits on face.

Technical problems for technical startups and for everyone else, normal life.

Customer acquistion, individual case of PMF? Mysterious.

Co-founder issues. Hopefully you can resolve things out, because it is the only good way out for the co-founders and therefore for the company. You need to have a real talk.

Runway stress? Time is money (because of expenses), so do things now that leads to making deals (= your product/service + marketing/selling). Do them now. Now. Do it.

Yeah, the struggle and pain is hard. So hard. Let us wish power to for the survival of each other. Good luck, and don't fuck it up (message directed also to myself).

Seems like Will's clock drawing test in Hannibal :)

I think what makes Mathematica so great product is that it is not open source (development is strongly guided by one man) and that it costs (can afford to hire many hundreds of people to build the professional design).

I think the most important thing in life is to feel how precious and potential it is to be "alive", whatever it means - having self-awareness, consciousness, agency (mind-body-environment interaction).

After this comes the question of how you use/direct this agency - the "meaning of life", so that you feel fulfillment, you know who you are, what your "story" is (and this story can be seen clearly only afterwards).

The state of constant happiness is bad. One needs times of sadness in order to understand/see/remember the state of happiness. One needs rainy days to appreciate sunny ones.

IBM IMS - fast until abend :)

Is this vibe-coded?

Variable naming cases differ between snake/camel cases even they are at same row? And variable naming is almost too consistent and "clear" to be written as human.

Uses GCC instrinsics that are not in MSVC, so not cross-platform or "C99 syntax".

No comments in code to explain/clarify anything at all.

Besides, at least in my real world first hand experience, in Windows, you cannot load all OpenGL functions just by using wglGetProcAddress(), because it doesn't return function pointers to those basic OpenGL functions that exported by opengl32.dll. And if you want to check valid pointer addresses, in Windows you should also consider all return addresses below 4 as invalid/"NULL". I wrote this note in my OpenGL function pointer loader as comment to remind myself what I learned as first hand experience when actually writing the code myself and struggling through the messy reality :)

If one is "experienced software developer", I'd say one then can guide oneself to right direction for learning by starting asking the most basic/obvious/childish questions, e.g. what libraries/languages/datasets/data formats/concepts is needed to understand and map/chart the territory (yes, that is what OP is basically doing now, asking but not just barreling into messy action) with a goal to create a minimal prototype.

True experienced developers have learned how to solve any programming problem. This related reading was actually yesterday mentioned at HN: https://beej.us/guide/bglcs/

"CSCI 2100: Unlearning Object-Oriented Programming" immediately caused me to disagree this one.

When I code in C, in the end, I usually miss the syntax for defining "objects/classes" (structs with functions and access controls), the syntax/notation that encapsulates/binds/groups the related state and its functions/API to define some specific concept/model == custom data type.

Of course OOP can be taken to extreme complexity and then lose its usefulness.

If someone wants to see SKs in their dreams, read following :)

https://writings.stephenwolfram.com/2020/12/combinators-a-ce...

Computer executes only one programming language: its own machine code. Everything else is just people's own mess of how to convert something that usually resembles English to this machine code.

The topic reminds me at least these two documents, which shows in more detail how simple but well-thought data model can create great results.

https://www.notion.com/blog/data-model-behind-notion

https://git-scm.com/book/en/v2/Git-Internals-Git-Objects

Having worked with real business databases, I truly believe and feel that foundational information modeling (database schema, data types) makes or breaks the utility and capabilities of that data.

Also patents and other per-country legal stuff.

And also the current application ecosystem: all the apps for iPhone/Android must be runnable. Not mentioned payment systems etc.

The most obvious answer would be to use the OS's native UI widgets, e.g. in Windows maybe even Win32 widgets. With under 200 lines (!) of C/C++ code you get windows, buttons, text fields, lists, tables, whatever you need. The end result is an executable of size less than 8 kilobytes, and has been working last 20-30 years in the past, and most probably will work decades in the future.

Besides, there isn't really "cross-platform GUI", because not every platform has same kind of widgets and behaviors.

I smell the "AI creativity tools" coming, meaning it cannot be fixed priced anymore.

The only way to win copycats and competitors is to just execute your own vision. The best example I know is Wolfram Mathematica. ~35-40 years of incremental development guided by the same "product owner/visionary", and there are no real competitors left and it is nearly impossible to build competing product now.

Low-tech is needed only when the high-tech is crippled/toxic/uncontrolled/broken/overengineered/expensive.

Couple of days to get basic commands to muscle memory and to get intuition how git works internally.

Save yourself from confusion and waste of time, and read/skim through the Pro Git book. It has very good diagrams explaining the workings of git. As of 2025, the book is a bit dated for the very latest (unnecessary?) commands of git, but the basics are still the same.

That said, git is the de facto version control tool in the industry, but I prefer Fossil for my personal projects. Nothing beats the timeline view and other simple tools and easy customizability in Fossil.

General Magic (movie)

Halt and Catch Fire (series)

Silicon Valley (series)

This question is so multifaceted and psychological (personally, and in the end, collectively), that it cannot be answered satisfyingly.

First thing I'd say is to figure out if the kid has real interest and willingness to learn programming/computers at all. If you want to find motivation, figure out what topic/hobby the kid is interested in, and code together something useful/fun/inspiring in that domain. After all, all real world programs are used for some specific domain.

Second thing is the young age of 11 (of course depends on invividual). "Real" programming requires understanding very abstract concepts (like memory, objects, functions, variables, loops, program state/execution model etc.). I remember when I was young (maybe ~13 yo), and I struggled alone with C++ and Delphi. I could copy and paste some code, but I couldn't really understand what make them work (or didn't). Of course it was different time then (no youtube tutorials, no AI, no good quality open source stuff etc.).

Third thing, if the kid is interested to learn, is good environment/language to start with. I'd start with plain vanilla HTML5 (HTML + CSS + JS). It's relatively easy to get feedback and interactivity with it, and it runs on every computer and mobile device. And the whole program - content (HTML), layout (CSS) and interactivity (JS) - can be in a single file and "recompiled" just by clicking refresh button. I wouldn't use any JS frameworks, because then it becomes learning something else than real basics. With couple of lines JS, you can draw graphics, play sounds based on keyboard/mouse/touch interactivity etc. But of course, even if its just handful of HTML5 APIs, to teach someone why/how they work is a big task.

Fourth thing, now with AI, for many, learning seems to have become a process of collecting random factoids here and there, and then trying to make some sense of the whole mess. I'd say a great book that presents an overview of a topic from start to end is overly underrated at this point of time. When book(s) stops giving lessons, use LLM(s)/professional(s). When LLM(s)/professional(s) stops giving you lessons, you are at the frontier of knowledge, and you have to figure it out yourself by teaching yourself :)

Here is my strategy for bookmarking:

I have my own custom frontpage.html as browser's homepage, where are links for websites I visit daily (e.g. HN). In that frontpage there is a tab, which shows links.txt content in iframe, where I throw links that I feel I should sometime revisit. I also have random.txt for random copypaste storage that shows in its own iframe tab.

If I encounter something I feel contains very important reading, I just take a full page screenshot of the webpage with Firefox's screenshot tool as png, so it won't linkrot/vanish. If the page is too long for png, I print it as PDF.