Will Manifold switch to optimistic UI updates?
60%
chance
Sep 1
M$284 bet
Currently while performing an action on Manifold, there is a delay where the frontend waits for a confirmation from the backend before updating the UI. Sometimes this delay is as long as a few seconds. Will Manifold instead switch to an optimistic approach, where the UI is updated immediately assuming the backend action will succeed; and only in the small chance of an error, will that error be communicated to the user. Jun 7: Closing date to Sep 1 Close date updated to 2022-09-01 11:59 pm
Sinclair
Sinclair Chen bought M$10 of YES
We should. Most of our users are on mobile. Also it's easy to do with react query.
0
MattP
Matt P is betting NO at 53%
@Sinclair oh man that would be great. The mobile experience is less than ideal right now from a technical perspective.
0
mqp
Marshall Polaris bought M$20 of NO
We have put in almost zero effort to optimize latency so far, so I think it would be inappropriate to add complicated stuff until we make a reasonable effort to get latency down and then see where we stand.
0
MattP
Matt P sold M$46 of YES
@mqp funny story - I sold my YES in response to this, and then the website promptly locked up instead of updating and showing me the new probability.
0
akhil
Akhil Wable is betting YES at 30%
@mqp makes sense if you think it’ll disincentivise reducing latency internally - though even once latency improves, updating the UI immediately might make for a better user experience, and something most sites do? A minimal indicator that the change is pending (like greying out) is is another option which covers both needs.
0
mqp
Marshall Polaris is betting NO at 34%
@akhil I just like making the simple common sense thing first, and then prioritizing from there. Common sense says that the round trip latency for placing a bet from a normal Internet connection, if we don't go nuts on geolocating servers, should be perhaps 100-200ms. I think if that were our actual status quo, we would not be having this conversation, because it just wouldn't seem important at all to introduce any complexity to improve perceived latency on crappy connections. So I think we should first make some reasonable effort to get latency in something like that ballpark and then consider this.
0
akhil
Akhil Wable is betting YES at 34%
@mqp Cool. Moving out the close date to Sep. I can then try to make the case for it's simplicity and usefulness even though you've already kicked ass on making the site super quick.
0
MattP
Matt P bought M$60 of YES
Man, I certainly hope so. Website performance on spotty internet connections feels super sluggish/laggy with the current implementation.
0