I wrote this short prose in Classical Chinese to express my best wishes to the Chinese graduate students among us who cannot have a family reunion due to their curriculum and research in a foreign land. I hope everybody enjoy their work despite not being able to visit their families, and in the end the hard work would pay off!

歲祝犁大淵獻孟春，獺祭雁來，蟄蟲始振。九州之民咸歸鄉里，以躬事親，其樂也融融。維斯學諸君，負笈夷地，雖闕地萬尺而神州弗及。奈何？然格物致知，切磋琢磨，亦別有其趣。夫志厲青雲者不拘小節，義存高遠者不畏䠑踽。於此新歲，吾其願子窮理究業，畢其端緒，以饗方家，方不負其心。勉之！

Most people who have a passion for cooking can be divided into two camps: those meticulous souls who would revel in freshly baked pastries following recipes to analytical-balance precision, and of course, those who even when equipped with a detailed recipe, having complete disregard of units and measures, would add or remove certain ingredients to their liking and adjust the amount of seasoning to their taste. In reality this is less of a dichotomy but more of a spectrum. I’d like to think myself as more towards the second type. In everyday cooking, the tolerance is pretty high, so who cares if you added three pinches of salt instead of two in your beef stew?

It goes without saying that any Chinese pre-modern texts ever existed did not have a digital version to start with. They were either handwritten, printed with woodblocks, or inscribed on a variety of objects (鐘鼎文, 石鼓文, 甲骨文, etc.). However, for the convenience of both research and education, there are significant governmental and academic efforts to convert those texts into a format that is more accessible. Building and keeping an electronic archive of Chinese Texts serves many purposes:

1. Easier search, annotation, and cross-reference.
2. Natural language processing (NLP) of Classical Chinese texts. The accuracy of models used in NLP (if taken a statistical approach) will directly benefit from a large corpus of well-structured texts.
3. Batch processing and digital publication of Chinese texts in different platforms with minimal effort.
4. Certain types of books, for example dictionaries (字書, 韻書), could be processed with metadata attached to each entry, which would facilitate complex queries. This would allow for character lookup across many dictionaries with one click (or shall I say one SQL query?), possibly even with example usage pulled directly from the full corpus.
5. Related with the previous point, easier query of dictionaries enables students to study Chinese etymology, phonology and dialectology, as well as researchers to compare phonology of dialects and conduct field studies.

This list can go on for much longer, but I’ll stop here for brevity.

I came across this video on YouTube a couple days back. As a novice machinist I was initially very confused - aren’t lathes supposed to be used to turn parts into axial-symmetric shapes? How is it possible to make, say a hexagon, with an engine lathe?

Ok, I get that you’ll need two spindles (part spindle, and live tooling spindle) that somehow are synchronized together, but how exactly can you produce a flat surface, given that both the part and the cutter are spinning, tracing out curves instead of straight lines?

As a graduate student having to read academic papers on a regular basis, frequently have I encountered references to journal articles in the form of DOI (digital object identifiers). Quite often they don’t come in hyperlink form, so it is annoying to actually access the paper by first copying the doi, then going to the doi website, pasting the doi into their tiny query box and clicking submit, being redirected to the journal page on whatever publisher’s site, and finally being able to download the paper or hitting a paywall (which is when Sci-Hub comes to rescue…).

I have long wanted to change that.