Errata

the word in is repeated showing "in in our current....."

Binary to decimal description is WRONG.

= 9 + 1 = 10

Should be

= 8 + 2 = 10

I haven't done "real" math in a long time, but isn't there an extra minus sign in:

log2(x) = -(log(x) / log(2))

It should instead be:

log2(x) = (log(x) / log(2))

right?

e.g. https://en.wikipedia.org/wiki/Binary_logarithm#Conversion_from_other_bases

iPod was introduced in 2001, not 1998. I remember because the original iMac came out in ’98, and the iPod / CD-RW were after that, when Apple was pitching the Mac as the “digital hub” of our homes.

Hi, your book is really intearsting, the error I foun in page 30 is a minor error,

2^3 = 8
2^1=2

And you put

2^3=9
2^1=1

The following sentence is incorrect:

Perhaps the simplest way to encode some text information would be to number all of
the English characters—A to Z—with numeric values 0–26.

It should be "0-25" at the end of the sentence.

The following sentences are incorrect:

Perhaps the simplest way to encode some text information would be to number all of
the English characters—A to Z—with numeric values 0–26. You could then use the
number of pulses, along with pairing, to determine what digit you were transmitting.
For example, you could translate “THE HAT” into 20-8-5 8-1-20.

Either the numeric values need to be "1-26" not "0-26", or each numeric coded value at the end of the second sentence needs 1 subtracted from it; eg "19-7-4-7-0-19".

The original sequence has 5 Ds, not 4 as described later in the example.

The values for n=1 and n=2 don't seem to be right.

The code for n=2 should probably be "100", because you encode the value of 2 by doing (2^1 + 0). The code "101" would match n=3.

In the table on the prior page, it's explained that the number 0 is not representable in unary, which suggests that n=1 cannot use (2^0), which means that the encoding of "0" must be some kind of special case that isn't listed in the algorithm.

The algorithm specified after "To decode Elias delta" seems like it would fail to correctly decode when the code is "0" as there wouldn't be any way to tell if the code word is done.

Following the encoding algorithm before the table, I encode n=1 as "10".

When compression formats for video are mentioned, webM is also mentioned. But webM is a container format, not a compression codec.

On this first line
.. write n=2^(N + L) (where L =n-2^N)...

Should this be n = (2^N) + L like the example below on the same page?

The adaptive VLC steps miss the one that should output the "B" character for the first time.

"GZIP, BZIP, and now" should be "GZIP, and now"

For most of the history of HTTP, DEFLATE and its wrapper GZIP were the only supported means of compression. BZIP2 is not a part of the "Standard HTTP stack" -- the only browser I'm aware of that ever supported BZIP2 was an early version of Chrome, and that code was ripped out long ago.

Today, everybody does GZIP/DEFLATE, Chrome+Firefox+Opera do Brotli, Chrome+Opera do SDCH, and Opera does lzma.