August 7th, 2019

Replying to a tweet from @davedisasterman

Could you explain what you mean? Using navigation preload shouldn’t use any more data (or time) than a regular navigation to a URL if there were no service worker present.

51° N , 0° E

Also on Twitter

Reply Retweet Favourite


Threadpool: The merc with the c: mount

Perhaps I misunderstood what you are doing, but if there’s no data involved, what is being ‘preloaded’? As I understand preloading, sites load data into memory for the most likely navigation choices the user will make, and this is how it makes the experience faster. This can 1/2

Threadpool: The merc with the c: mount

eat user’s data plans behind the scenes, and arguably uses resources unnecessarily in all cases, which ultimately costs energy. While we worry about page bloat, package sizes, 3rd parties and the like, it seems a valid concern. I made an assumption tho, so will take another look

Jeremy Keith

Yeah, that’s something completely different that also uses the word “preload”, but it’s nothing to do with navigation preloads in service workers.

Threadpool: The merc with the c: mount

Ah I see now. I was thrown by my assumption. I apologise for jumping to that conclusion. Unfortunate consequence of shared nomenclature. Nav preloading in the context I mentioned has been around for a few years. Good for perceived performance but dubious in practice.

Have you published a response to this? :