Errata

Though the description is correct, the rail diagram for "name" need slight modification. The representation (in the rail diagram) for all name should start with letter is missing. Meaning I can make a name as following as per the current diagram shown, however this is error as per the language grammar in strict (or) non-strict mode.

"use strict";

var 73_i = 10;
var 53i = 20;

Hello:
The grammar diagram present for integer at page number 8 shall handle for negative number. If not representing for positive sign is okay, but its required to note for negative sign.

Some thing similar to denoted for exponent digit in exponent grammar.

Thanks,
Karmegam

In the Strings section on Chapter 2 (Grammer), There is a railroad diagram marked "escaped character". One of the options is a forward slash ("/") which in fact is NOT escaped, and can be included in strings as-is.

The tracks exiting "escaped character" should go back between the quote literal and the upper track, not after the upper track, otherwise one cannot specify a single escaped character.

The diagram is saying the following situation would be impractical

var a = "1";
switch (a) {
case "1":
alert(a);
break;
},

since the diagram is not allowing to use only one case clause and finish it with a disruptive statement. Also it is acceptable use of "default" statement without a consecutive disruptive, as follows:

var a = "1";
switch (a) {
case "1":
alert(a);
default:
alert(a);
}

I really don't know whether it is a matter of a ECMAScript version or a tiny mistake. However, worth a look. Thanks!

The lowest path of the "expression statement" railroad diagram shows that "delete {expression} {refinement}" is a legal expression statement; however, the "expression" railroad diagram shows that "delete {expression} {refinement}" is a legal expression. That suggests that "delete delete {expression} {refinement} {refinement}" would be a logical JavaScript statement.

Also (combined with the "for statement" railroad diagram on page 13), it suggests that "delete {expression} {refinement}" would be a legal increment clause for a for loop declaration.

In the comfirmed errata at page 21, says:
It should be stooge["first-name"] // "Jerome"

In my opinion, this is not the better correction.
I perfer to change code at page 20:
"first-name": "Joe"

Why? because of at page 22, in the 2nd line:
... the property value is replaced:
stooge['first-name'] = 'Jerome';

which means: "Joe" is REPLACED by 'Jerome', natually.

Otherwise, "Jerome" is replaced by 'Jerome', which is almost meanless.

That regular express removes white space from both beginning and end of line. Not only end of line.

Better example in closures section
modes[i] should be nodes[i].

myMammal object doesn't have a "saying" function so the the says function would not work unless augmented later as is done by the mycat object.

This only makes sense if my suggestion to introduce Object.create() as a builtin on page 22, is taken.

Replace "the Object.create method from Chapter 3." with "the Object.create() method."

This only makes sense if my suggestion to introduce Object.create() as a builtin on page 22, is taken.

(It also only makes sense if the modern Object.create() method still does not work on arrays!)

Replace "the Object.create method from Chapter 3" with "the Object.create() method".

The group capturing the hostname should include an escaped \., because . alone will match any character

The re "parse_number" on p.69 does not accept .78 or -.62
both of which are legal in many (all?) programming languages
including JavaScript. The repair is trivial and would not burden
the exposition. Please see the StrDecimalLiteral production in
97-028.pdf at the 1997 directory of the archives at ecma-international.org.

Suspected errata in Javascript the Good Parts by Douglas Crockford
submitted Apr 2021.
re: (c) 2008 printing history May 2008: First Edition
re: ISBN 978-0-596-51774-8

The make_a_matcher() example does not return 10 for y.lastIndex; instead, it returns 0. Maybe RegExp objects made by regular expression literals do not share a single instance?
See http://stackoverflow.com/questions/1326374/are-regex-literals-always-regexp-objects

The text reads:

"An unescaped ^ matches the beginning of the text when the lastIndex property is zero. It can also match line-ending characters when the m flag is specified.

An unescaped $ matches the end of the text. It can also match line-ending characters when the m flag is specified."

It is identically stated of ^ and $ that they "can also match line-ending characters when the m flag is specified."

It would be more appropriate with ^ to specify that: "It can also match immediately after line-ending characters when the m flag is specified."

The RegExp:
var doubled_words = /([A-Za-z\u00C0-\u1FFF\u2800-\uFFFD]+)\s+\1/gi;
will match any pair of identical strings of one or more letters separated by a space.

That means it will return many false positives where a substring at the end of one word matches a substring at the start of the following word. Matching false positives may prevent it from finding actual duplicated words.

For example:
"the end end".match(doubled_words);
returns ["e e"], a match against the last letter in "the" and initial letter in "end".
Consequently, it fails to match against the duplicated "end"s.

Erratum for my erratum about Object.create().

It should say "more an important update" not "more and".

Btw, the last paragraph there is part of the text I think should be inserted (not just commentary).

The reason I suggest adding an entry for Object.create is that if (as I suggest) you start treating Object.create as part of the language, a reader might then think page 86 is the logical place to look for info about it.

The function Expersson (second form):
var foo = function foo() {}; // A form

above experssion itself is fine, however, it is better to change as:
var foo = function () {}; // B form

The author claimed: "Throughout this book, I have been using the second form because it makes it clear ...". But throughout the book the author always used as B form and never as A form. B form is simple and clear, but A form is unclear and seems no explanation inside the book.

My edition says [2011-04-08].

If suggestion is taken to add Object.create entry on page 86, then add 86 to this entry in the index.

I see that if any number of lines were added to page 86, other entries on later pages would move to new pages! Hoping your index is automatic.

By the way, "If you lived in Github, you'd be accepting a pull-request by now."