JSON是否属于二进制编码格式？作为可读文本的它如何编码为字节序列？

Great question—let’s break this down clearly, since there’s a common confusion here between text-based formats and how they get converted to bytes for storage or transmission.

First, let’s clarify the core distinction:

• JSON is a text-based serialization format: Its entire structure is built from human-readable Unicode characters (like {, ", :, letters, numbers) that follow a strict syntax for representing data structures (objects, arrays, strings, numbers, booleans, null).
• Binary encoding formats (like Protobuf, BSON, Thrift): These skip the human-readable text layer entirely—they encode data directly into raw bytes with no inherent character-based structure. You can’t open a Protobuf file in a text editor and make sense of it, but you can with JSON.

The confusion usually comes from two places:

• Every piece of data (text or binary) must be converted to a byte sequence to travel over a network or be written to a file. This step applies to JSON too, but it’s a separate process from the serialization format itself.
• In some workflows, JSON might be wrapped inside binary containers (like in certain message queues or database blobs), which can make it look like it’s a binary format at first glance. But the JSON itself is still text underneath.

The process has two distinct layers, which is easy to mix up:

Step 1: Serialize data to JSON text #

First, your in-memory data (like a Python dictionary, Java object, etc.) gets converted into a valid JSON string—this is the human-readable text part. For example, a user object might become:

{"name": "Alice", "age": 30, "is_active": true}

Step 2: Encode the JSON text to bytes #

Since computers only understand bytes, this Unicode text needs to be mapped to a byte sequence using a character encoding scheme. The JSON specification (RFC 8259) explicitly recommends UTF-8 as the preferred encoding (though UTF-16 and UTF-32 are allowed too).

Here’s how this works for a simple example:

• The character { (Unicode code point U+007B) maps to the single byte 0x7B in UTF-8.
• The string "Alice" uses ASCII characters (all within U+0000 to U+007F), so each letter maps to one byte (e.g., A is 0x41, l is 0x6C).
• If you have non-ASCII characters (like {"name": "张三"}), the Chinese characters are encoded into multi-byte sequences: "张" (U+5F20) becomes 0xE5 0xBC 0xA0, and "三" (U+4E09) becomes 0xE4 0xB8 0x89 in UTF-8.

When sending this over the network or writing to a file, you’re sending/writing that sequence of UTF-8 encoded bytes. On the receiving end, the bytes are first decoded back into Unicode text, then parsed into the original data structure.

内容的提问来源于stack exchange，提问作者user1008636