Closed thescurry closed 6 years ago
Beng's estimate: 3 or 5
This ticket requires some investigation - we are using lazy loading for all lists, how to do 'scrollTo' if card is not loaded yet... My estimate is between 5 and 8.
Yeah, the list navigation could get a little dicey. I go with an 8 to be conservative.
I created https://github.com/hippware/rn-chat/issues/2478 since there's not a clear path to take with the biz logic.
Should we talk through this one as a group?
I don't think lazy loading is ideal for the home mode view... mostly because the list will comprise of only the pins on the screen (and possibly some tutorial cards). Once a user moves the map to a location and taps a pin, then we should quickly load the list based on what pins are viewable on the screen (1 to X number of pins).
For discover mode, maybe lazy loading works? That will be a chronologically ordered list, so maybe it's possible. Could we fetch 10 cards and when those load we "pre-fetch" 10 more? If we could always have a large enough buffer of cards available when in this mode, the swiping could seem fairly seamless. How do other apps that use a "swipeable list" solve this problem? I believe apps like Tinder, Shapr and a few others use a list like this, maybe we could borrow some ideas?
For discover mode, maybe lazy loading works?
Yes, I think we're safe to stick with our current paging/lazy loading scheme for discover mode since the order will still be date-based.
I don't think lazy loading is ideal for the home mode view
I agree. I don't think we should be using lazy-loading/pagination for home mode view.
For discover mode, maybe lazy loading works?
Yes, because there is an implicit assumption that it's date based.
I agree. I don't think we should be using lazy-loading/pagination for home mode view.
The trade-off here is that we have to pull in (potentially) a lot of information every app start. This is the reason I proposed 2 "levels" of queries here.
The trade-off here is that we have to pull in (potentially) a lot of information every app start. This is the reason I proposed 2 "levels" of queries here.
Fair enough.
This is starting to get into (premature?) optimisation territory and I don't have enough visibility to give specific insights, only generalisations. You'll have to try it out (or try different approaches) and see how it works out.
Verified on Staging Version 3.1.3
Verified on Prod Version: 3.4.2
Verified on Prod Version: 3.8.1
Zeplin link: https://zpl.io/blm4w6k