Last Call Review of draft-ietf-ecrit-data-only-ea-18
review-ietf-ecrit-data-only-ea-18-opsdir-lc-schoenwaelder-2019-08-21-00

I assume all my comments are easy to resolve. Some are purley editorial,
others may require minor content changes, hence I ticked 'has issues'. The
comments are in document order and not ordered by 'seriousness'.

- The acronym 'LoST' pops up in section 3 without any explanation what
  it expands to or what it means. This appears to be to only
  occurrence of 'LoST' in the document, which makes me feel lost.

- It seems that in 4.2 there are three cases distinguished under the
  'sender:' list item but the indentation somewhat confuses this. Perhaps
  the typesetting can be improved, e.g.

  sender: ...

  - Originator is a SIP entity, Author indication irrelevant: ...

  - Originator is a non-SIP entity, Author indication irrelevant: ...

  - Author indication relevant: ...

  Right now the last ends up on the level of 'sender:' and the text
  indentation of the first two is different.

- What is a 'PIDF-LO' structure? References missing.

- Section 4.3 says that a data-only emergency call is sent using a SIP
  MESSAGE transaction. Earlier text says that "this document only
  addresses sending a CAP message in a SIP INVITE that initiates an
  emergency call, or in a SIP MESSAGE transaction for a one-shot,
  data-only emergency call." It seems there is no further text
  detailing the SIP INVITE case nor is it clear how one decides
  between the two options. The examples also focus on the MESSAGE case
  and the security considerations also discuss MESSAGE and CAP (and
  not INVITE).

- Do the authors use 'human understandable' and 'human readable' as
  synonyms? If so, it seems 'human readable' is the more common phrase
  with which we describe texts in protocol messages.

- Section 7 warns about sending 'large quantities of data'. Is it
  possible to be a bit more precise what 'large' means? Is 1k large?
  Is 10k large? Is 1m large? There is also the phrase 'very large', is
  that the same as 'large' here? I am not looking for a precise number
  but rather an indication up to which magnitude sending data inline
  may be safe.