Dropping iOS 15 & other old browsers in July 2025

Announcements
84

I’ve been following along here, and am interested in continuing to better understand what the impact of this change will be on different communities.

The way I’m thinking of it, for a given community, one could break that down as follows:

  • T total number of users in the community
    • x number of those users impacted by this change
      • a number of those users who don’t have an acceptable workaround
      • b complement of a
    • y complement of x
  • t time we wait to make this change.

For a given community, I assume there is a high probability that x > 0, and that even a > 0.

We could think of the impact of this change on a given community as being modeled as a function f(t) that returns the values T, x, a.

If we accept that x and a will be hard to get to 0, what should we be aiming for?

Perhaps we bucket impact by a/T and set some threshold on what we find acceptable.

We could think of this impact of this change on communities as a whole as being a similar function F(t) that returns a population of communities []{T, x, a}

We could use the same threshold above and measure how many communities are impacted beyond what we’ve determined as acceptable.

If we were to wait another year for this, I assume that for many communities, both x and a would decrease, but still be greater than 0.

So there’s no perfect decision here.

Then, what should factor into our decision?

What is an acceptable value for a for a given community? How many communities are we willing to have cross that threshold? should we make this change?

We haven’t done anything as rigorous as this, but we have looked as some of the data we have to inform our decision, and felt like May is a reasonable answer for t.

6 Likes