# Poisoning Survey Initial Results

###### tags: `Blog Drafts` `Kitchen Sync`

## Content

- Include some graphs of results
- Link to raw results
- Some insights (don't draw specific conclusions yet)

## Total respondents
786

## Do you use the standard library’s lock types (Mutex or RwLock) in any of your projects?
Yes: 88.8%
No: 11.2%

## Do you use locks from outside the standard library (such as from antidote or parking_lot) in any of your projects?
Yes: 55%
No: 45%

## Which of these is closest to how you acquire locks? (Select all that apply)
.lock.unwrap(): 84.3%
.lock?: 28.6%
.lock().unwrap_or_else(PoisonGuard::into_inner): 6.1%

## Have you ever thought carefully about lock poisoning in your projects?
Yes: 37.5%
No: 62.5%

## If so, what do you want to happen when a lock is poisoned? (Select all that apply)
I want my program to terminate: 62.8%
I never want my program to terminate: 18.2%
I want to access the data again: 22.8%
I never want to access the data again: 24.5%
I want to check and fix the data: 26%

## Do you think you’ve benefited from the standard library’s lock types providing poisoning by default?
Yes: 32.4%
No: 67.6%

## How much friction do you think would be involved in migrating any of your projects from the standard library’s poisoning lock types to a non-poisoning lock crate?
None: 41.4%
A manageable amount: 56.7%
Too much: 1.9%

## Would you use a poisoning implementation from the standard library if it was independent of Mutex or RwLock?
Yes: 52.5%
No: 47.5%