Hello, World!

Hello, world!

I haven't been writing blog posts for a long time. The reason why I had given up blogging is that currently there are no blog systems fulfill all my needs. I tried to get used to Medium. I love its reply system -- which guides people to write a new post instead of a simple comment -- actually this was what the world of blog used to be, but the render effects of code blocks sucks. Moreover, I'm a multi-linguist, I want to write both in Chinese and English at the same time, and sometimes Japanese and German. But the quality of Chinese character's render effects on Medium doesn't match those of Latin script or even Japanese Kana. After 1 year of bearing such a torture of playing with mismatched product quality on Medium, finally I decided to implement my own.

Jekyll and Abandoned

Firstly, I chose Jekyll. But soon, I abandoned. Frankly, I don't like Ruby, which is the language implements Jekyll. I'm a Chinese grew up in Mainland China. For one of the core concepts of Simplified Chinese and Ruby is both "simplify for simplification", and such a kind of simplification rule produces many irregular cases which increases brain burden after all, thus I don't like both. Ruby's grammar is highly irregular which is just like the strokes of characters is more irregular in Simplified Chinese than what they were in Traditional Chinese. You write less on the surface, but remember more and take more time to understand what you written under the surface.

There are more historical details I can tell in between China and Japan about how simplification "took over" two countries, which one is the "motherland" of Simplified Chinese and the other Ruby's.

Japan got started to simplify the pronunciation system of Chinese characters (Kanji) in Meiji Restoration. Several decades later, 9 days after the foundation of the People's Republic of China, Japan announced the use of its new form of Chinese characters -- Shinjitai (literally means "new character form"), which is just another approach to simplify Traditional Chinese.

While Japan simplified the Chinese character's pronunciation system, it is impossible for China. For there are already tons of homophonic words in Chinese, whenever you are speaking Chinese or Japanese, the number of homophonic words would dramatically increase after you did whatever simplification to the pronunciation system. When speaking Japanese, you can get rid of those homophonic Chinese words by "falling back" to native Japanese words or phonetic interpreted words came from languages other than Chinese. But when speaking Chinese, there are no "fallback words". Thus we don't simplify the pronunciation system of our language.

Thus the only simplification we did to our language is just the simplification to the form of characters. About ten years after the foundation of the People's Republic of China, the central government announced to gradually replace the use of Traditional Chinese with Simplified Chinese.

Well, but the reason why I had abandoned Jekyll is not as complicated as the historical details I mentioned above. Just like other blog systems that I have used before, Jekyll doesn't fulfill all my needs. The user defined excerpt markup is awesome, but I also want shortcodes in WordPress. I had to write additional codes to generate tags and category if I wanted to host it on GitHub Pages.

Zola and Abandoned

Because I've been learning Rust for weeks when I decided to implement my own blog system, I also searched some blog frameworks written in Rust. Then I found Zola.

Truth be told, Zola is a good book generator. Even you can use it to generate documentations and guides. But as a blog system, it fails.

It was designed to be a single binary such that the extension of the system is extremely difficult. Because of its simplicity, it is the best choice if it hits all your needs and the only thing you have to do is to design a theme. But soon fails totally if one of your needs is out of its spec list.

Gatsby.js and Determined

Occasionally, I found Gatsby.js. Or say, the meet with Gatsby.js is my destiny. After trails of Jekyll and Zola, it no longer scares me that to take some time on assessing a blog system by writing some trail codes with it.

Just after downloading the tutorial project and scratching with lines of code, I decided to implement my own blog system with Gatsby.js.

Gatsby.js is built upon React.js and node, which I have never got touched with before. But picking up these two skills is quite simple. The expression is straight and easy to understand. The concept behind Gatsby.js is more like data-oriented programming. But it doesn't matter that if you don't know such a buzz word. You can imagine that Gatsby.js works like a pipeline, and the only thing you have to do is just to "hook-up" on correct time and add your custom contents.

When MDX Met with Gatsby.js

After several days crawling on Gatsby.js community, I quickly found an awesome upgrade to traditional Markdown: MDX.

MDX is JSX enhanced Markdown, which means you can insert JSX tags in Markdown documents and the framework would replace it with real React.js component on behalf of you when rendering the document.

This means I can implement a replacement of WordPress shortcodes with JSX tags and React.js component now. Some features like left/right float figures, charts and components out of current imagination can also be easily extended into the system now.

Struggle with Gatsby.js

The perspective is beautiful but the road is frustrating.

Since the framework interprets MDX document in JavaScript, and the JavaScript escapes backslash (/) escaped tokens (which means interprets valid escaped tokens likes "\n" into a real newline and invalid escaped tokens likes "\L" into "L"), the first thing I have to do is to get the framework stopped doing that.

For the framework converts the MDX document into Markdown AST firstly, and then into MDX AST, this mis-interpretation can be corrected after Markdown AST generation by inserting a guard backslash before each backslash in the Markdown AST. The principal behind this is simple: Since the JavaScript blindly escapes everything after a backslash, so just get that backslash escaped.

You may wonder that there are also escaped characters in Markdown grammar, what happens to them when you blindly inserting a guard backslash before each existed backslash?

Well, the answer is simple: The escaped characters of the Markdown language in the document have been escaped in the progress of converting MDX document into Markdown AST, which means they are no longer escaped characters in the AST now, you can safely do it on generated Markdown AST.

After fixed the backslash issue, there were still many issues waiting for me. But building a system with Gatsby.js is joyful.

Stage of the Project

As I mentioned before, there are no blog systems fulfill all my needs. Does Gatsby.js fulfill all my needs?

Yes. Though there is not an out-of-box Gatsby.js project fulfills all my needs, but building what I need is quite simple with Gatsby.js.

Currently this blog implements

  • A responsive design

  • Basic Markdown syntax support

  • Inline and block KaTex syntax support

  • Tags and category

  • Table-of-contents for sectioned posts

  • Creative Common 4.0 license claimer

  • RSS feed

But this is not the end. In the next stage the project would add support to:

  • Webmention

  • "Shortcode" for major video sites

And the site would finally implement a beautiful render effect for Chinese character and Latin script mixed documents which enables me to post in languages I intentionally want to.

Since I decided to build this site in a couple of weeks, I didn't ship a very modern design. The appearance of the site is quit old school. But I have already got a new design. Since refreshing the appearance of the site needs a lot of time, it might be taken after half or one year.



Freiheit ist Wille, Handeln ist Fähigkeit.
Built with Gatsby.js.