Monday, February 7, 2022

Linux and github

So, this is probably the real reason why Linus Torvalds is so dismissive about using github as a dev tool for Linux...

http://lkml.iu.edu/hypermail/linux/kernel/1603.2/01906.html 

http://lkml.iu.edu/hypermail/linux/kernel/1603.2/01926.html 


I guess rightly so.

Thursday, February 3, 2022

Embedded marketing in Google Maps directions is a thing

 https://news.ycombinator.com/item?id=30171800&p=2#30172766

Apparently embedded marketing in Google Maps directions is a thing.

Yuck.

I generally don't mind ads as a concept since people don't pay for $things much, but somehow this feels so horrible (I can't articulate why but probably for multiple reasons)

Saturday, January 29, 2022

First world problem

First world problems..... 特登走去睇返 egress 價位,確認返我冇搞錯 order of magnitude (1GB $0.01 同 $0.1 同 $1 係好大分別架) 另外好在 target audience 得千幾萬人,每人每日最多都玩兩三次,個 upper bound 好清晰。

睇嚟呢個有機會係我人生最 hit 嘅(個人製作)產品。

 




Tuesday, January 25, 2022

"10x programmer"

睇緊死鬼舊老細啲訪問,都唔係第一次睇,每次都領悟到啲新嘢。

佢話世界上好多嘢,最好水平嘅同一般嘅水平,都係相差幾成至一倍。但好嘅軟件工程師同一般嘅,差距有五十倍。呢個係九十年代嘅訪問嚟。

以今日嘅軟件開發工具嚟講 (*cough* copilot *cough* stack overflow),可能可以將普通軟件工程師嘅程度提升,所以 50x 可能係高估咗。但你問我嘅話,「10x programmer」係絕對存在嘅,呢樣嘢嘅存在簡值係無容置疑。唔知係咪受美國啲奇怪政治正確風氣影響,見好多人唔接受「10x programmer」呢樣嘢,真係唔知係意識形態驅使,定係真係見識太少。

當然,所謂 10x 都係好粗疏嘅量化,雖然呢樣嘢絕對存在,但如果要標籤某某係或唔係 "10x programmer " 實在唔係好恰當。喺一個健全嘅團隊入面,唔應該有任何人工作效率比其他人高出 10 倍,呢樣嘢對被標籤嘅人都唔公平。何況有時 10x 係低估咗,好多時勁人就係能人所不能,你搵十萬大軍都做唔到嘅嘢,佢一個人就搞得掂。呢啲人才係「無價」嘅,不過最後佢份糧通常最多都係你(十)幾倍。

好多概念雖然喺現實存在,但唔代表應該隨便用。人概價值好難用單一維度去量度。

Thursday, December 16, 2021

Sunday, November 21, 2021

Konqueror


It was more than 20 years ago when I first saw it.

Those were the days when Internet Explorer was at its peak glory. IE5, being (unsarcastically) the best web browser in the world at the time, killed off Netscape in a high stakes battle to become the world's de-facto interface to the Internet.

At that time, I recently switched to using Linux as my main OS, somewhat as an ideological statement against the inevitable Microsoft dominance over the tech industry. I also loved the developer-friendly environment having discovered the power of having the *nix OS at my fingertips. The only problem with Linux at that time, was that it did not have a functional web browser.

Yes, in 2000, Linux did not have a functional web browser.

Some older Linux distributions gave you Netscape 4, but Netscape 4 was well known to be buggy and by 2000 already very out of date. (Netscape 6 was utter garbage) *Everyone* was using IE5, and many web sites just flat out did not work on Netscape. Mozilla was at version "0.6". Previously it was called "Milestone 18", which did not inspire much confidence. Whatever they called this thing, the UI was slow and laggy. Memory intensive too. They invented something called XUL that allows developers to write UI for applications using XML and Javascript -- if you thought Electron is slow now with hyper-optimized JS engines, imagine that happening in 2000.

There was also w3m of course, but I never got the hang of it, even with its graphics extensions on xterm.

Anyway, shortly after KDE2 was released, I downloaded it and ran it for the first time. Imagine my shock when I saw a fully functional, mostly standards compliant, *fast* browser that incidentally showed up as one of the apps. I literally didn't believe it, I loaded a couple web sites I had in development, tried to run Javascript on it, and it just worked. I still remember how mind blown I was when "<script>alert()</script>" popped up a new window. Javascript engines are a dime a dozen these days, but in 2000 given how many man-hours were poured into developing IE and Netscape/Mozilla, I think I can be pardoned for naïvely presuming that it requires a multimillion dollar project to make one.

https://en.wikipedia.org/wiki/K_Desktop_Environment_2#/media/File:KDE_2.2.2.png

I mostly did web development at the time, which precluded me from using Konqueror as my main browser since I couldn't use it to check HTML rendering fidelity (as in, rendering properly on IE and Mozilla), but it did imprint a lasting impression on me -- this was a small sub-project in KDE, yet it was more than competitive with the big browsers, arguably much better than Mozilla was at the time. It was such a pity that only KDE users would have the chance to use it.

After using KDE for a while, I eventually settled on GNOME2 as my main desktop environment (I hated KDE's IE/ActiveDesktop-like asthetic, and the important apps in GNOME, Evolution and Nautilus, were better), and by that time Mozilla/Firefox was improving to become a somewhat reliable web browser, so my time with Konqueror was actually quite brief. (btw, KDE and QT were both great software that didn't get as much credit as they should have.)

Fast forward 20 years. Some of Steve Job's underlings must have been similarly impressed by the project and adopted the code in Apple's new Safari web browser as part of WebKit. Google forked WebKit and created Chrome. IE died of neglect and Microsoft's new browser is also based on Chrome. Mozilla's Firefox is now on life support. Microsoft never achieved its then-seemingly inevitable world domination. Who would have thought that 90+% of web traffic these days originate from a Konqueror descendant.

Anyone who was there at the time would have thought IE's domination was unassailable and seemed capable of lasting forever. It took Konquer's descendants more than 10 years for them to do it, but looking back, it feels pretty amazing how a tiny little browser nobody had heard of eventually beat IE at its own game. 

Monday, October 25, 2021

成語 (好似係

- 削髮如泥

- 飽暖知榮辱

- 三文兩語

- 春雞一燒值千金

- 迅雷不及掩耳盜鈴

- 酒mate菜僕

- 此亂即棄

- 一年四季在於春

- 兩Bi俱雙