Ambrose

Add friend

http://c.gniw.ca/works

Ambrose’s Bookshelves

to-read‎ (10)
currently-reading‎ (5)

read‎ (17)
did-not-finish‎ (0)

half-read‎ (0)

Stats | More…

Ambrose is Currently Reading

Ambrose is currently reading

A Designer's Research Manual: Succeed in Design By Knowing Your Clients and Understanding What They Really Need

by Jenn Visocky O'Grady

bookshelves: currently-reading

Rate this book

Clear rating

1 of 5 stars 2 of 5 stars 3 of 5 stars 4 of 5 stars 5 of 5 stars

Ambrose said: " This book looks very useful, but I cannot wholly recommend it because of obvious factual errors.

On page 78, when explaining web analytics, the authors claim “users’ IP addresses” are “unique to every computer on the web”. Even if technically true at This book looks very useful, but I cannot wholly recommend it because of obvious factual errors.

On page 78, when explaining web analytics, the authors claim “users’ IP addresses” are “unique to every computer on the web”. Even if technically true at any given point in time and when network paths are considered, from a web analytics point of view this is false and has been false for at least two decades.

Even in 2006, when the first edition came out, this statement had already been factually incorrect for a few years: All computers behind a proxy share the same public IP address (i.e., you can undercount visitors, possibly by several orders of magnitude in the case of university campuses, for example), and most residential computers were already given random “dynamic” IP addresses (i.e., you can overcount and you can mix up the identities of your visitors).

(Version 1.0 of the Squid proxy ceased development in 1996, so firewalled HTTP access must have existed more than two decades ago. It’s very hard to believe anyone would claim IP addresses are “unique” even in 2006.)

I obviously can’t claim to have seen any “proprietary” analytics software. Maybe the author is talking about the military or the police, but even if that were the case they aren’t doing this by IP addresses alone (they would need to use their special powers to do things normal people aren’t allowed to do). As for the publicly available ones, I’ve never seen any take these facts into consideration, or when they do, it’s done wrong so the results are wrong.

By 2017, when the second edition came out, the world had already officially run out of IPv4 addresses for six years. Unless the authors have been misled to believe IPv6 adoption was already at 100%, it’s incredible that the authors continue to claim, in the second edition, that IP addresses are “unique” to “every computer on the web”. (Even if IPv6 adoption were at 100%, I would not claim IPv6 is “unique” even if every IPv6 address is tied to MAC addresses, since MAC addresses can be changed on some devices.)

Modern analytics partly depend on cookies and fingerprinting (even in 2006 the better analytics software depended on cookies) precisely because IP addresses, as seen by analytics software, are not unique; but these are not mentioned in the book. Also, even these are imprecise (e.g., cookies can only identify unique browsers, not unique individuals) and can violate privacy laws.

I can’t talk about the rest of the book because I’m not an expert in research, but misstating the uniqueness of IP addresses and omitting the pivotal role of cookies (and, in 2017, failing to mention fingerprinting) make the rest of the book much less credible—it makes you wonder what other errors are in the book. ...more "

progress:

(page 105 of 208)
— Sep 26, 2018 08:49AM

Sep 26, 2018 10:40AM · see review