Closed barbarer closed 2 months ago
I don't think its ever been 100% accurate. We just use a single counter stored in a redis server. There is no serialization or synchronization for updating the variable. So in a busy class there are bound to be many race conditions where multiple people read the same number update it, and then write it back.
It is usually much better than it was during the instructor workshop. It would say 2 when we each had sent several (about 14 people).
Is there any way to make it accurate? Are the messages being saved to the database so you could just get that number from the database?
Dr. Barbara Ericson Associate Professor, School of Information University of Michigan
On Thu, Jul 25, 2024 at 11:27 AM Bradley Miller @.***> wrote:
I don't think its ever been 100% accurate. We just use a single counter stored in a redis server. There is no serialization or synchronization for updating the variable. So in a busy class there are bound to be many race conditions where multiple people read the same number update it, and then write it back.
— Reply to this email directly, view it on GitHub https://github.com/RunestoneInteractive/rs/issues/514#issuecomment-2250677795, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKOZ7MY4FNOJCEVB3SX5I73ZOEKMXAVCNFSM6AAAAABLOXQ6UOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENJQGY3TONZZGU . You are receiving this because you authored the thread.Message ID: @.***>
It was never my understanding that the accuracy of the count was important. You always explain in the workshops that its when the number slows down or stops going up.
Of course there are ways to make it accurate, but it doesn't seem like a high priority thing to me.
It should be a high priority. I use it to know when to start the second vote. Is the number of votes accurate? I also use that.
Dr. Barbara Ericson Associate Professor, School of Information University of Michigan
On Thu, Jul 25, 2024 at 5:55 PM Bradley Miller @.***> wrote:
It was never my understanding that the accuracy of the count was important. You always explain in the workshops that its when the number slows down or stops going up.
Of course there are ways to make it accurate, but it doesn't seem like a high priority thing to me.
— Reply to this email directly, view it on GitHub https://github.com/RunestoneInteractive/rs/issues/514#issuecomment-2251462655, or unsubscribe https://github.com/notifications/unsubscribe-auth/AKOZ7M4N2ZMUW5YOKIXA3TTZOFX5RAVCNFSM6AAAAABLOXQ6UOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDENJRGQ3DENRVGU . You are receiving this because you authored the thread.Message ID: @.***>
This hasn't changed in how it works from day one. Is there any evidence that you have been starting the 2nd vote too early or too late?
Changed this to use the database to count.
In any course the number of text messages is not updating correctly on the instructor PI interface. At the PI workshop we knew that we had sent lots of messages, but the number of messages did not update correctly.