Not my list. This list is derived from the SWEBOK project, an IEEE Computer Society list. It is a very process oriented list. Seven books on UML. I don't know anyone who uses UML. I've never worked at a company that actually used it. At least as many books on Agile Programming; how many do you need? A book on CORBA. Really! Knuth is there; I guess he is mandatory. No other good books on algorithms, but lots on patterns.

The author said that he explicitly left out books on technology. Sure there is a lot of technology cruft out there, but surely there a few language specific books that a good software engineer should read. Or maybe I'm just a hacker.

Since there isn't anything else positive on here right now..

It's a pretty good list, and I think he did a pretty decent job. Sure, everybody's list would be different, but there are some really good books on there that would serve people well to read them. It's interesting that Psychology of Computer Programming didn't make the list, as it frequently does.

It'd be nice to see another list without the exclusions though. K&R (C) and the Llama (Perl) book especially were very helpful to me. SICP made the list, and I kinda consider that a Scheme book.. I certainly learned plenty of programming/CS from other language-specific book. The 4.4 BSD book helped a lot of people too. Had I come up later, Accelerated C++ is a really nice book too.

> No other good books on algorithms, but lots on patterns.

CLRS is on the list; 28.

And I, likewise, was unhappy with his exclusion of technology-specific texts. It seems as though he is asserting that generalizable ideas can only come from books that purport to be general.

why the UML hate? Sequence and class diagrams and such can come in pretty handy as a design or reverse engineering aid.

I've been tempted to read one book on UML. But seven? Surely at least four or five of them are redundant?

Imagine what you'd get if you used the same technique to find the "best novels ever."

Modern Library held a "top books of the century" thing, using both critics and popular vote to make two lists.

http://www.randomhouse.com/modernlibrary/100bestnovels.html

The critics one is decent, if a bit off (Ulysses 1: yes, Great Gatsby 2: hardly), but the Ayn-Rand-meets-L-Ron-Hubbard on the other side always makes me laugh, even though I'm a fan of Rand.

Their popular list suggests they didn't have any technology for discounting the effects of voting blocs. I wonder if to this day they believe they had a representative sample of popular opinion.

If they did, they'd have to be pretty seriously deluded. Atlas Shrugged I can see as #1 popular - there was the poll that said it was second only to the Bible for average Americans - but Hubbard's stuff is incredibly derided. I don't know anybody who's read a single book of his.

I'm sure they released a result more to gain popularity through controversy than they did to get a truly accurate reading. And it worked, since theirs is always the one I've heard named when people discuss "great books," and it always gets people angry.

Using this article as a guide, then apparently Grisham.

I think you're being a little harsh. There are, like, 10 books worth reading on that list.

Seems to be rather American-centric, and extremely topheavy with recent fads, which isn't surprising, given the ridiculous method used to compile it. No Dijkstra, no Wirth, no Hoare, no Brinch Hansen. No compilers, no Lisp (although SICP might qualify). Totally absurd.

There are books on here that I think are worthwhile, that could be called classics, given the attention they've received. But not the whole list.

Generally, they are the ones that are specifically tied to the act of software development, of making things and getting good at doing that well, rather than the process and other roles surrounding it.

But then it seems that, given the definition from SWEBOK and others like http://c2.com/cgi-bin/wiki?SoftwareEngineer, software engineering is more about the process than the product. So this list may be correct, and the reason I don't hear about some of these titles so much is because I'm not as much a part of the process crowd.

For people who are looking for a shorter list: 1. http://www.joelonsoftware.com/navLinks/fog0000000262.html 2. http://www.hanselman.com/blog/SixEssentialLanguageAgnosticPr... 3. http://www.codinghorror.com/blog/archives/000020.html

Not inherently bad, but... I can think of a least a 100 things I'd rather do than read 100 books on the same topic.

Very interesting for sure. Just wonder how much the 'school factor' influences rating. How many reviews and Googlings are due to students forced to read an 'arbitrary' title? Which titles are in fact professors favorites? I think Applying UML and Patterns could be one.

Holy god, "Head First Design Patterns" is #2? That's written at a 9th grade level. The illustrations are written at a 4th grade level.

Knuth all the way down at 16? Tisk.

You don't read Knuth. You put him on your nicest bookshelf and check out a page or two for emotional support when working on Java Enterprise Application #431404.