---
summary: User id comparisons should be case insensitive
---
created: 2014-10-27 08:38:49.0
creator: matthew
description: |-
  I'm seeing more and more disasters with people failing to identify users correctly due to case sensitive IDs.  For instance Amandine's business cards say @amandine:matrix.org whereas she's registered as @Amandine:matrix.org.

  Splitting this out from SYN-78
id: '10488'
key: SPEC-289
number: '289'
priority: '1'
project: '10001'
reporter: matthew
status: '1'
type: '1'
updated: 2016-10-28 16:27:57.0
votes: '0'
watches: '3'
workflowId: '10592'
---
actions:
- author: markjh
  body: |-
    If we want any level of flexibility for this lookup we need the HS to implement mapping from printed user-id to actual matrix ID.

    Unless we restrict the space of IDs to fit into ASCII then rigorously defining a case insensitive match is slightly painful.
  created: 2014-11-04 11:13:09.0
  id: '10723'
  issue: '10488'
  type: comment
  updateauthor: markjh
  updated: 2014-11-04 11:13:09.0
- author: markjh
  body: Blocked by specifying a lookup API.
  created: 2015-01-15 13:16:25.0
  id: '11128'
  issue: '10488'
  type: comment
  updateauthor: markjh
  updated: 2015-01-15 13:16:25.0
- author: matthew
  body: |-
    The current plan of action on this (as discussed IRL yesterday evening is):

     * Store IDs case-sensitively in the DB
     * Compare them case-sensitively in the code & protocol
     * Canonicalise them either at the origin client or origin HS by querying a lookup API on the destination HS (as per SPEC-62).  This adds a roundtrip, but it may be suppressable by combining with some other roundtrip.

    Open questions are:
     * Do you do the canonicalisation lookup from the origin client (yet another random thing to expect clients to remember to do, which they will never remember to)
     * Do you do the canonicalisation lookup from the origin HS?
     * Do you use a user alias->id (assuming user aliases were separated from underlying IDs) lookup to also perform the canonicalisation?
  created: 2015-04-14 18:41:37.0
  id: '11490'
  issue: '10488'
  type: comment
  updateauthor: matthew
  updated: 2015-04-14 18:41:37.0
- author: matthew
  body: this looks to have been mistakenly closed; the problem is still glaringly open.
  created: 2015-12-04 11:37:42.0
  id: '12427'
  issue: '10488'
  type: comment
  updateauthor: matthew
  updated: 2015-12-04 11:37:42.0
- author: richvdh
  body: A solution to this (at least for invites) is proposed at https://docs.google.com/document/d/1B7q_3ruJzeQTg-uJHe1UScxbVLzgm451c25OjpYcojI.
  created: 2016-07-14 09:55:47.0
  id: '13057'
  issue: '10488'
  type: comment
  updateauthor: richvdh
  updated: 2016-07-14 09:55:47.0
- author: richvdh
  body: 'Migrated to github: https://github.com/matrix-org/matrix-doc/issues/586'
  created: 2016-10-28 16:27:57.0
  id: '13394'
  issue: '10488'
  type: comment
  updateauthor: richvdh
  updated: 2016-10-28 16:27:57.0