Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
💡Background
As a user I found the terminology used in the module quite confusing, and difficult to understand, as some of the terminology wasn't really using the words or phrases that I'd expect.
🤔 Ideas for improvement
In this PR I've made several proposals which I think make the spam module more understandable for the regular user.
Some of this is fixing typos, but some is changing the actual wording used to describe things, to something which I think might be a bit more familiar.
I think that these all make sense in the context, but some I could not find in the UI directly so I am just working on assumptions.
🔄 Changes proposed
Consistency, e.g. always using TLDs rather than TLDs and tlds.
Typos - in a few places there were extra letters or words which were mis-spelled which I've fixed.
Adding a note that the admin will still need to manually block the user.
Using their instead of its when referring to the user.
Avoid the -> Don't allow the (it makes better grammatical sense, I think).
Instead of saying what happened (word matches with dictionary), I changed it to what this means (forbidden word found, forbidden domain extension found) which I feel as a user makes more sense, because I can directly understand the impact.
Clarity in the fact that the user gets an email about their account being locked.
An error occurs -> An error occurred (better grammatical sense).
Making more of a sentence, instead of 'when spam', use 'when spam content is detected' - explains what is actually happening (spam is found in the content submitted).
I think that instead of 'agent', which doesn't make a lot of sense to me as a user, perhaps using the word 'trigger' would be better? It informs them directly what that does, whereas agent is a bit unclear (do you mean a secret agent🕵🏻? an estate agent🏠?) whereas trigger is explaining what that thing actually does (trigger, or not, the actions to take place) - e.g. 'fire spam trigger if rules match' sounds clearer to me than 'fire spam agent if rules match'.
I think 'detection symbols' is also a bit confusing, perhaps 'triggers' is better? e.g. Check the trigger/s required to fire the rules rather than Check the detection symbols required to fire the rules.
😱Things I could not find to correct
Dictionnary is mis-spelled several times, but it's not in this file - I would probably change it to 'Block list' and change 'Dictionnary configurations' to 'Block list configuration'
'Sinalize' is not at all a common term in English, but I can't find where to change this - perhaps it's a core language string?
List forbidden words -> List of forbidden words
'Remove scan' on the button when in a dictionary setting - I'd say 'remove block list'
Hope that this might be helpful!