WEBVTT

00:00:00.800 --> 00:00:03.530
PROFESSOR: Hi, have you found it
particularly difficult to find

00:00:03.530 --> 00:00:05.484
a specific item in your house?

00:00:05.484 --> 00:00:07.400
Let's say you're looking
for a pair of gloves,

00:00:07.400 --> 00:00:11.160
but you just can't find it, and
you spend your entire afternoon

00:00:11.160 --> 00:00:12.660
looking for it.

00:00:12.660 --> 00:00:15.060
Well, you've just
encountered the same problem

00:00:15.060 --> 00:00:17.840
that companies like
Google or Microsoft

00:00:17.840 --> 00:00:20.650
encounter every single day.

00:00:20.650 --> 00:00:22.910
And that's the
problem of search.

00:00:22.910 --> 00:00:25.750
Just like a house which stores
thousands of different items,

00:00:25.750 --> 00:00:30.390
Google stores 45 billion
index pages of information.

00:00:30.390 --> 00:00:32.640
If every page was
a sheet of paper

00:00:32.640 --> 00:00:37.760
and we stack them up real high,
we'll create a tower 600 times

00:00:37.760 --> 00:00:39.560
taller than Mount Everest.

00:00:39.560 --> 00:00:41.820
Well, how can Google
find my results

00:00:41.820 --> 00:00:44.090
so quickly when I
find it so difficult

00:00:44.090 --> 00:00:46.880
to find a pair of gloves?

00:00:46.880 --> 00:00:48.460
Well, searching
on Google is kind

00:00:48.460 --> 00:00:52.870
of like looking for
person in a big school.

00:00:52.870 --> 00:00:55.770
Let's say you're looking for
James in a row of classrooms.

00:00:55.770 --> 00:00:57.380
One of the easiest
method would be

00:00:57.380 --> 00:01:02.760
to go to every classroom nearest
to you until you find James.

00:01:02.760 --> 00:01:06.940
There's a better method
called binary search.

00:01:06.940 --> 00:01:10.030
Now, let's say the students
were arranged from a to z,

00:01:10.030 --> 00:01:12.200
one in each classroom.

00:01:12.200 --> 00:01:13.720
We would then go
to the middle room

00:01:13.720 --> 00:01:16.680
first and see if James is there.

00:01:16.680 --> 00:01:18.340
And if James isn't
there we will look

00:01:18.340 --> 00:01:21.340
at the first letter in the
name, and if the letter is

00:01:21.340 --> 00:01:26.670
before j, we head to the right,
if not, we head to the left.

00:01:26.670 --> 00:01:28.470
We would then approach
the middle room

00:01:28.470 --> 00:01:30.790
in the newly sectioned area.

00:01:30.790 --> 00:01:34.460
Eventually we will repeat the
process over and over again

00:01:34.460 --> 00:01:36.980
until we find James.

00:01:36.980 --> 00:01:40.940
Now, this is just like the first
method, but it's at a much,

00:01:40.940 --> 00:01:43.350
much faster rate.

00:01:43.350 --> 00:01:45.320
How much faster would that be?

00:01:45.320 --> 00:01:48.540
Well, that depends on the number
of students in the school.

00:01:48.540 --> 00:01:51.670
Let's say there are 500 students
and we're looking for one.

00:01:51.670 --> 00:01:54.400
It will take about 80
minutes in the first method,

00:01:54.400 --> 00:01:57.650
But one and a half minutes
with binary search.

00:01:57.650 --> 00:02:00.350
But let's say there are
1,000 students in the school.

00:02:00.350 --> 00:02:04.080
It would take 160 minutes
with the first method,

00:02:04.080 --> 00:02:07.965
but 1.6 minutes with
the second method.

00:02:07.965 --> 00:02:10.310
Now that's a whole
lot of difference.

00:02:10.310 --> 00:02:12.080
So a name is just a word.

00:02:12.080 --> 00:02:15.640
But Google searches a
combination of words,

00:02:15.640 --> 00:02:18.690
making it a little
bit more complicated.

00:02:18.690 --> 00:02:21.510
So just like how we
identified the first letter

00:02:21.510 --> 00:02:23.660
of each alphabet
of the name, Google

00:02:23.660 --> 00:02:27.900
identifies 200 unique factors
making your search terms

00:02:27.900 --> 00:02:29.610
faster.

00:02:29.610 --> 00:02:32.390
Well, if you recall, the
effectiveness of binary search

00:02:32.390 --> 00:02:35.810
depends on the
prearrangement of data.

00:02:35.810 --> 00:02:38.110
And that's why computer
scientists are actively

00:02:38.110 --> 00:02:42.210
looking for ways to sort,
manage, and eventually retrieve

00:02:42.210 --> 00:02:45.220
data faster and better.

00:02:45.220 --> 00:02:48.990
In the same way, the TV
remote goes near the TV,

00:02:48.990 --> 00:02:53.200
the shoes go to the shoe rack,
the coats go into the cupboard,

00:02:53.200 --> 00:02:56.230
and the winter gloves go
into the winter jacket.

00:02:56.230 --> 00:03:00.000
Aha, so that's
where my gloves are.