Separate the Machine You Touch Hardware With
※本記事にはアフィリエイトリンクを含む場合があります。内容は広告の有無に影響されません。
結論
Hardware work can destroy a machine through a USB port after a wiring mistake, so the machine holding data you cannot lose and the machine used for experiments are safer kept physically apart.
The division of roles
- MacBook Air = the mothership … prototyping, documents, meetings
- ThinkPad X1 = the field machine … driving and flashing microcontrollers and sensors
In one line: the Mac defends, the ThinkPad attacks. There are two reasons for the split.
Reason ① drivers and ports
Many cheap microcontrollers and sensors carry a serial chip such as CH340 or CP2102.
That is awkward on macOS: installing the driver takes lowering security settings or handling kernel extensions in the terminal. On Windows, plugging it in is usually enough to be recognised and flashed.
Ports matter too. The ThinkPad X1 has two USB-A + HDMI as standard (plus Thunderbolt 4), so microcontrollers, sensors and programmers plug in with no adapters. The MacBook Air, by contrast, has two USB-C and a headphone jack. If you are plugging several cables in on site, the former is more straightforward.
Reason ② work that can break things, and things you cannot lose
Hardware development sits next to the risk of physically destroying the PC. Short something out through a wiring mistake and, at worst, the motherboard can burn out through the USB port.
So the split is physical: the ThinkPad is “the field machine that may as well break”, the MacBook is “the side that protects the data”.
Putting business data and development assets on a machine that might short out during an experiment is frightening. It is risk management: separate “things you cannot lose” from “work that might break something”.
What the mothership is for
The MacBook Air is the thinking, making and communicating side. Prototyping, documents and meetings all concentrate here.
Fanless, quiet and long on battery, so it works away from a desk without worrying about power. Gathering data here and protecting it is its role.
The criterion for splitting is not specs
Carrying two is heavy and costs money. The reason to split anyway is that it is separating the loss, not splitting the specs.
To split by performance, one higher-end machine would do. But “work that might break something” is not solved by performance. Do both on one machine and an accident on one side takes the other with it.
Whether to pay the cost of redundancy is decided by what you cannot afford to lose.
よくある質問
Q1Why is the hardware side Windows?
Because many cheap microcontrollers and sensors carry serial chips such as CH340 or CP2102, and installing the driver on macOS can require changing security settings or handling kernel extensions. On Windows, plugging it in is usually enough to be recognised.
Q2How much do the ports matter?
Having USB-A and HDMI as standard means microcontrollers, sensors and programmers plug in with no adapters. Where you are plugging several cables in on site, that difference is working speed directly.
Q3Can one machine not do both?
It can, and then your business data sits on the machine that could short out during an experiment. The reason to separate is not specs but separating the risk.
この記事の根拠
- ドキュメントファイル
本文の主張は、上の記録に書かれていることだけです。運用しているリポジトリは非公開のため リンクは張れませんが、どのファイルの何行目を、どのコミット時点で見て書いたかは 記事ごとに残しています。推測で書いた箇所はありません。