From d1b6f192afe4069b2569e4ed4407acf4bf1d248f Mon Sep 17 00:00:00 2001 From: Donovan Date: Thu, 26 Jun 2025 13:29:10 -0500 Subject: [PATCH] fix typos --- docs/nkode_over_unencrypted_channel.md | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/docs/nkode_over_unencrypted_channel.md b/docs/nkode_over_unencrypted_channel.md index 4db88bb..9a536ba 100644 --- a/docs/nkode_over_unencrypted_channel.md +++ b/docs/nkode_over_unencrypted_channel.md @@ -3,8 +3,8 @@ ## Low-Bandwidth Architecture The standard nKode architecture will not work in low-bandwidth environments. -Keypad icons are too large to send from the sever to the client. -To over come this issue, we can move the nKode icons from the serve to the users mobile device. +Keypad icons are too large to send from the server to the client. +To over come this issue, we can move the nKode icons from the server to the users mobile device. The server only sends the indices in which the icons need to be arranged. ```mermaid @@ -15,8 +15,9 @@ sequenceDiagram Note over User,Server: Enrollment User ->> Server: Initiate Enrollment Server ->> Server: Generate Keypad Icons + Note right of Server: Ideally the icons are generated on the users device.
Since current ML models are too compute intense, a GPU enabled server must run the models during enrollment. Server -->> Mobile Client: Store Icons On Device - Note right of Server: Server does not store the icons and does not know what they are + Note right of Server: The Server does not store the icons Server ->> Mobile Client: Keypad Index Array Mobile Client ->> User: Render Keypad User ->> Server: Set nKode