WEBVTT

1
00:00:00.180 --> 00:00:01.170
So right now,

2
00:00:01.171 --> 00:00:05.820
I think we have this huge opportunity to broaden our reach to many of the people

3
00:00:05.821 --> 00:00:08.100
who are now accessing the web for the first time,

4
00:00:08.310 --> 00:00:10.440
but it takes a lot of care to do that well.

5
00:00:10.920 --> 00:00:15.150
And it's up to us to use that opportunity wisely. And to be clear,

6
00:00:15.151 --> 00:00:19.230
it's not just Javascript, we're sending more code overall year over year,

7
00:00:19.890 --> 00:00:24.210
and there are real indifferent costs to that increase in how much we're

8
00:00:24.211 --> 00:00:29.040
delivering some types of code that we deliver are more costly to speed,

9
00:00:29.310 --> 00:00:32.040
and some are more costly just to data consumption alone,

10
00:00:32.190 --> 00:00:33.570
but both of those are problems.

11
00:00:33.750 --> 00:00:36.540
I've always included these numbers in my performance talks,

12
00:00:36.570 --> 00:00:39.570
and these are from the HTTParchive.org again.

13
00:00:39.571 --> 00:00:41.490
And it says that the average web homepage,

14
00:00:41.520 --> 00:00:45.450
these days weighs anywhere from a megabyte and a half to two megabytes.

15
00:00:45.660 --> 00:00:49.980
And if I were to grab an updated stat on this, even today,

16
00:00:50.070 --> 00:00:53.010
it's probably even more than two megabytes. This is always growing.

17
00:00:53.340 --> 00:00:57.630
And I should add that these are median averages and medians are nice.

18
00:00:57.631 --> 00:01:02.250
But the thing about median numbers is that they tend to quiet the outliers at

19
00:01:02.251 --> 00:01:05.250
the end of, of, at either extreme on a dataset.

20
00:01:06.210 --> 00:01:10.230
And many of the most popular sites on the web are heavier than that median.

21
00:01:10.350 --> 00:01:13.320
So weight is a real problem with data usage.