|
Size: 2206
Comment:
|
Size: 3078
Comment: Revert apparent spam.
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 12: | Line 12: |
| idempotent *Deprecated* This term has in the past been used to describe the property we now call "invertable". Idempotent really means you can apply a function to an input, and if you apply the function again to the output from the first application, you get back the same result. That's not what the email package does. |
**idempotent** A property of certain operations in computer science and mathematics. An operation is idempotent if multiple applications of the operation do not change the result. Formally, given an operation 'g', 'g' is idempontent if and only if: |
| Line 19: | Line 18: |
| invertable The email package attempts to maintain *invertability*. By this we |
g(g(x)) = g(x) For example, the 'lowercase' operation is idempotent. There are operations provided by the email package where it makes sense to require either strict idempotence, or idempotence when possible. **invertible** The email package attempts to maintain *invertibility*. By this we |
| Line 23: | Line 28: |
| put in. For well-formed input, this is an absolute guarantee, and | put in. For well-formed input, this is an absolute requirement, and |
| Line 25: | Line 30: |
| to break invertability. | to break invertibility. Note that invertibility is a stronger requirement on an operation than idempotence, but it applies only when an inverse operation exists. |
| Line 27: | Line 34: |
| raw data *Deprecated* because its usage has been ambiguous. In some cases it is another term for wire-format, used especially when the data is expected to not be RFC conformant. But it has also been used to refer to transfer-decoded bytes, on the theory that the decoded bytes are the 'raw data' that went into the transfer-encoding pipeline at the originating MTA. |
**conformant** Conforming to a particular specification or standard. The Internet RFCs use this term to refer to implementations and data that conform to the requirements of the RFC. |
| Line 35: | Line 39: |
| string | **raw data** Data in the form it enters the email module parser, or exits the email module generator (primarily the former). [*]_ A related term is 'wire-format', the difference being that wire-format data is understood to be RFC conformant. Raw data may or may not be RFC conformant, and may or may not be bytes (if, for example, it comes from a doctest or other text input source). .. [*] Note that in the past this term has been used ambiguously to also refer to the original source data that was transfer-encoded into the form that is the actual raw data that the email module deals with. **string** |
| Line 38: | Line 55: |
| text | **text** |
| Line 41: | Line 58: |
| transfer-decoded | **transfer-decoded** |
| Line 44: | Line 61: |
| transfer-encoded | **transfer-encoded** |
| Line 46: | Line 63: |
| "over the wire", ie: to wire-format. | "over the wire", ie: to wire-format. Mostly used in the verb form (ex: "after the data has been transfer-encoded") in discussing operations involving the RFC defined transfer encodings Quoted Printable and BASE64. |
| Line 48: | Line 68: |
| wire-format | **wire-format** |
| Line 51: | Line 71: |
| containing the data of the message nominally transfer-encoded. Wire-format data may or may not be well formed according the RFCs; the term refers to the data actually found in the wild. |
containing the data of the message transfer-encoded according to the RFCs. |
Glossary of Terms-of-Art for the Email Package
This page is an attempt to standardize on the language we use to describe concepts relevant to the email module. It also mentions some terms that are deprecated as incorrect or ambiguous, and why.
NOTE: this is a proposed draft, not a final document!
- idempotent
A property of certain operations in computer science and mathematics. An operation is idempotent if multiple applications of the operation do not change the result. Formally, given an operation 'g', 'g' is idempontent if and only if:
g(g(x)) = g(x)
For example, the 'lowercase' operation is idempotent. There are operations provided by the email package where it makes sense to require either strict idempotence, or idempotence when possible.
- invertible
- The email package attempts to maintain invertibility. By this we mean that if you feed an input into the package, and later ask for that data to be serialized back out, you should get out the data you put in. For well-formed input, this is an absolute requirement, and any deviation is a bug. For other input, we may find it necessary to break invertibility. Note that invertibility is a stronger requirement on an operation than idempotence, but it applies only when an inverse operation exists.
- conformant
- Conforming to a particular specification or standard. The Internet RFCs use this term to refer to implementations and data that conform to the requirements of the RFC.
- raw data
Data in the form it enters the email module parser, or exits the email module generator (primarily the former). [*] A related term is 'wire-format', the difference being that wire-format data is understood to be RFC conformant. Raw data may or may not be RFC conformant, and may or may not be bytes (if, for example, it comes from a doctest or other text input source).
[*] Note that in the past this term has been used ambiguously to also refer to the original source data that was transfer-encoded into the form that is the actual raw data that the email module deals with. - string
- python3 unicode string
- text
- unicode text (stored in a python3 string)
- transfer-decoded
- Data that has been decoded from wire-format into 8 bit bytes.
- transfer-encoded
- Bytes that have been validly encoded per the RFCs for transmission "over the wire", ie: to wire-format. Mostly used in the verb form (ex: "after the data has been transfer-encoded") in discussing operations involving the RFC defined transfer encodings Quoted Printable and BASE64.
- wire-format
- The format that data is in when transmitted "over the wire"; which is to say in a binary format rather than unicode, said binary format containing the data of the message transfer-encoded according to the RFCs.