github.com/nf/docker@v1.8.1/docs/terms/image.md (about)

     1  <!--[metadata]>
     2  +++
     3  draft = true 
     4  title = "Image"
     5  description = "Definition of an image"
     6  keywords = ["containers, lxc, concepts, explanation, image,  container"]
     7  [menu.main]
     8  parent = "mn_reference"
     9  +++
    10  <![end-metadata]-->
    11  
    12  # Image
    13  
    14  ## Introduction
    15  
    16  ![](/terms/images/docker-filesystems-debian.png)
    17  
    18  In Docker terminology, a read-only [*Layer*](/terms/layer/#layer) is
    19  called an **image**. An image never changes.
    20  
    21  Since Docker uses a [*Union File System*](/terms/layer/#union-file-system), the
    22  processes think the whole file system is mounted read-write. But all the
    23  changes go to the top-most writeable layer, and underneath, the original
    24  file in the read-only image is unchanged. Since images don't change,
    25  images do not have state.
    26  
    27  ![](/terms/images/docker-filesystems-debianrw.png)
    28  
    29  ## Parent image
    30  
    31  ![](/terms/images/docker-filesystems-multilayer.png)
    32  
    33  Each image may depend on one more image which forms the layer beneath
    34  it. We sometimes say that the lower image is the **parent** of the upper
    35  image.
    36  
    37  ## Base image
    38  
    39  An image that has no parent is a **base image**.
    40  
    41  ## Image IDs
    42  
    43  All images are identified by a 64 hexadecimal digit string (internally a
    44  256bit value). To simplify their use, a short ID of the first 12
    45  characters can be used on the command line. There is a small possibility
    46  of short id collisions, so the docker server will always return the long
    47  ID.