TL;DR
As researchers continue to migrate from our older Pipe version of EFP to the RecordRTC version (Jan 2024), we should consider options that would allow us to cancel this 3rd party subscription.
Current state
We pay a base rate for a Pipe Pro subscription, plus additional usage charges if we go over the amount provided with our plan. The fees are per-recording (not storage!) and based on type/quality and duration.
Following the rollout of the RecordRTC version, our monthly Pipe usage has dropped steadily. The last time we went over our subscription usage was about 1 year ago, in November 2024. Since then, our usage has averaged around 27% of what’s included with our subscription. This means we are paying for usage that we are not using. However, the range of usage over this time has been variable (5% - 50% of our plan) and not decreasing predictably month-to-month.
Number of studies currently using a Pipe version of EFP
48
(This number is hard to figure out because we don't store the necessary info in the database. I think I would either need a list of all commits prior to RecordRTC (or the inverse of that), or I would need to use the Github API to dynamically check the dates associated with commits).
Researcher impact
- Studies that were built with a Pipe version would need to update their experiment runners. There could be some compatibility issues, depending on how old the study's EFP version is.
- Pipe studies that are still collecting data would not be able to claim to have used the exact same experiment version for the full duration of the study.
- Researchers who have been getting their videos in mp4 format would now be getting webm files.
Implementation and estimates
-
Limit Pipe EFP versions in the validation step of the study design form This is very straightforward. This would only prevent the use of Pipe versions going forward, when a study is built/re-built. This could be done without the risk of breaking existing studies. Estimate: 2 days
-
For existing studies that have already been built with a Pipe version of EFP, these studies will simply stop working when we cancel our Pipe subscription. Ideally we would not allow a study to just stop working. Options:
- Give researchers a 'please update' warning with a deadline, communicated via email and maybe a banner on study details page
- After the deadline, deactivate the study if it is active, and delete the pre-existing build files or overwrite them by rebuilding with the latest version
- Tech support for version upgrading/compatibility issues
- Others?
TL;DR
As researchers continue to migrate from our older Pipe version of EFP to the RecordRTC version (Jan 2024), we should consider options that would allow us to cancel this 3rd party subscription.
Current state
We pay a base rate for a Pipe Pro subscription, plus additional usage charges if we go over the amount provided with our plan. The fees are per-recording (not storage!) and based on type/quality and duration.
Following the rollout of the RecordRTC version, our monthly Pipe usage has dropped steadily. The last time we went over our subscription usage was about 1 year ago, in November 2024. Since then, our usage has averaged around 27% of what’s included with our subscription. This means we are paying for usage that we are not using. However, the range of usage over this time has been variable (5% - 50% of our plan) and not decreasing predictably month-to-month.
Number of studies currently using a Pipe version of EFP
48
(This number is hard to figure out because we don't store the necessary info in the database. I think I would either need a list of all commits prior to RecordRTC (or the inverse of that), or I would need to use the Github API to dynamically check the dates associated with commits).
Researcher impact
Implementation and estimates
Limit Pipe EFP versions in the validation step of the study design form This is very straightforward. This would only prevent the use of Pipe versions going forward, when a study is built/re-built. This could be done without the risk of breaking existing studies. Estimate: 2 days
For existing studies that have already been built with a Pipe version of EFP, these studies will simply stop working when we cancel our Pipe subscription. Ideally we would not allow a study to just stop working. Options: