Outside of a dog, a man's best friend is a book

[pozorvlak]

Inside a dog, it's too dark to read. (Groucho Marx)

One of my favourite entries in my list of email signatures is the following quotation:

A really good, readable textbook/manual is worth its weight in very, very, very nice coffee.
   -- Tim Sweetman

A while ago, my sigmonster included this at the end of an email to my flatmate, who responded

There is not such a thing as very, very, very nice coffee.
There isn't probably such a thing as a good, readable textbook/manual, either.

Now, the first bit is just because he's a Philistine and doesn't drink coffee :-). But how about the second bit? This made me think about which books, exactly, deserve that description.

In computing, the book against which I measure all others is Wall, Christiansen and Orwant's Programming Perl. It's witty, erudite, clear and thought-provoking - it explains the hairy details of the Perl interpreter's inner workings by making puns in Anglo-Saxon, for goodness' sake, and somehow this is the perfect explanation. But my favourite bit has to be the explanation of the regex engine in terms of The Little Engine that Could :-). Never afraid to fly in the face of conventional wisdom, it's had a real effect on the way I think about programming. And it's incredibly comprehensive.

Coming in a close second has to be Knuth's The TeXBook, for all the same reasons. As for other books... well, K&R, obviously, and possibly Stroustrup's The C++ Programming Language. And I'd like to mention Scott Meyers' Effective C++ series, although it had the opposite effect on me from that intended: it made me decide that any language which forced you to think about this arcane rubbish was clearly not ready for prime-time. Paul Graham's On Lisp also looks pretty good, but I haven't read the whole thing. Then again, I've only read about half of the TeXBook, so what the hell. The BBC Basic manual was pretty good. The Gang of Four book is mind-expanding, as well as being clear and readable, and Martin Fowler's Refactoring, though lighter-weight, deserves a mention. And I'd like to give an honourable mention to Andrei Alexandrescu's Modern C++ Design: Applied Generic and Design Patterns, which (though the depth precludes its being clear and readable enough to make the cut) is simply mindblowing.

In maths, there don't seem to be so many good candidates. Halmos' Naive Set Theory. Birkhoff and MacLane's Survey of Modern Algebra. Ian Anderson's Introduction to Combinatorial Mathematics (though it's a bit too short). Maybe Apostol's book on analysis. Derek Holton's Problem Solving series, if they count as textbooks. Maybe Borceux's Handbook of Categorical Algebra. Mac Lane's Categories for the Working Mathematician I exclude for the same reason as Modern C++ Design: it's simply too dense to be easily readable. I had an excellent book on group theory written by someone with a German-sounding surname, possibly beginning with W, but I gave it away and can't find it on Amazon. Not that I'm biased, but Wilson Sutherland's Introduction to Topological and Metric Spaces is great. Tom Leinster's book Higher Operads, Higher Categories (which probably shouldn't count as a textbook) is if anything too readable: I often find I've read several pages but haven't really engaged with the meaning.

I suppose it's been eight years since I used textbooks for any other subjects, but I'm surprised I can't remember good textbooks from before that. I remember my history textbook from when I was about 12, which covered a period on a double page - the left-hand page was facts, and the right-hand page was a description of the time by a fictional typical character of that period. But that's it.

Of course, it would be remiss of me not to mention Charlie Dancey's truly excellent Encyclopaedia of Ball Juggling :-) EDIT: And Lofty Wiseman's SAS Survival Handbook, and Geoffrey Budworth's The Knot Book. And How to Hold a Crocodile :-). Divers readers (thanks, everyone!) have left recommendations for textbooks on topics ranging from criminal law to pharmacology, and from electronics to adult movie-making - see the comments.
So, I turn it over to you. What are your favourites? What textbooks do you think are worth their weight in very, very, very nice coffee? :-)

Comments

[neoanjou.livejournal.com]

The only book I think is especially brilliant in a classic 'textbook' type manner is: Mary L. Boas - Mathematical Methods in the Physical Sciences. Legendary book.

Then of course there are the classics:
* The Fenymann lectures on Physics (I don't think I've ever used for actual reference, but they look damn good on your shelf)
* 'Numerical Methods in C' - again I haven't directly used it, but every other mathematical technique I'm using seems to have a reference to this book.

BTW - Calibrating according to the firebox.com Civet Coffee sales price £15.95 for 57g we get a good text book being worth 279.82 UK£/kg (241.39 US$/pound). Suprisingly enough I'd guess this is more than most textbooks cost per kilo.

[totherme.livejournal.com]

I've just shocked myself with the following thought: I'm not sure I've ever really got on with the sort of book that you're talking about :)

When I was really quite small, I used to read factual stuff voraciously, cover-to-cover. I had a shelf full of books that were essentially lists of facts with titles like "The DIY Genius Kit" (which google now tells me was written by Gyles Brandreth), all of which I'd read cover to cover. I had various books about animals (and animal factbooks) and a subscription to zoobooks (which seem to have had a couple of make-overs since then - they used to look a lot like the spanish editions mentioned on the back issues page). I remember books with titles like "How things work" and "The big book of everything" too - but the vagueness of my memories defeats my google-fu here. I remember at least one of them being huge (so probably A4) and red.

So yeah - those were all things I reckon to have been worth their weight in belgian chocolate. But I don't remember ever having read a textbook cover-to-cover. These days the only books I seem to read cover-to-cover are fiction. I get most of my non-fiction from articles of one sort or another, or from dipping in to diverse (often anonymous) textbook like things - using them as reference material. Stuff that's online is better for me in this sense, because search tools make dipping in to the right things at the right time a much easier prospect. I particularly like hyperlinked things - wiki style. If I know the word/concept in question, or if I don't really care about it, then I can read on past it. If I don't understand it, and need or want to know more about it, I can dive deeper in another tab.

Perhaps those sorts of tools have taken the place of the parents I used to pester with that sort of question? (what's this word? What's that actually mean? Why not just...)

So - do those half-memories of childhood books count? Does google or wikipedia? (I suspect not)

Having said all that - I'm increasingly finding myself interested in history and economics. A part of that is almost certainly Neal Stephenson's fault, but I flatter myself that I was thinking about using systems and process analysis tools to try to understand the humanities even before I read his stuff. Perhaps I should look for a slightly more grown-up version of the big book of everything - Freakonomics perhaps...

On a tangent - I'm interested to note the spread of the comp textbooks you particularly value... A fair spread of imperative styles and a touch of functional (controversial statement I know - but I don't think lisp is all the way there) - but no static typing. I wonder if a really good book (by your definition - since I'm working with the possibility that they may not quite exist for me) explaining powerful static types exists? It could be a part of a book explaining something else of course - a methodology, paradigm or language. As a first (double barrelled) shot, I'll suggest the haskell wikibook and YAHT.

[half-of-monty.livejournal.com]

Ueno's introductions to algebraic geometry: I,II and III.

These are the only algebraic geometry books I've read cover to cover, but I'm sure they qualify, as they explain (a small part of) the core material properly, giving geometric intuitions on the way, and also, they are very light. :-)

Also `the not so short guide to latex 2e', being as it is virtual and thus weighing nothing! Despite really being quite useful!

Sorry, I fear I may be cheating. But then, however really really nice your coffee may be, it still won't agree with me.

[elvum.livejournal.com]

Statistics: A Guide to the Use of Statistical Methods in the Physical Sciences, by R J Barlow. Also the only textbook within reach as I write this. Most recently used to justify some recommendations for project staffing at work, to great success: nobody understood it, but everyone was very comforted by the fact it had stats in*...

* apart from the guy whose job it is to assign people to projects

[r-e-mercia.livejournal.com]

Gotta be McKendrick's excellent book on contract law. Pithy and inspiring, much like the man himself. Craig's textbook on EC law is also very good. As a door stop. Or a paperweight. Or for standing on if you can't reach something.
But enough about law textbooks. My favourite textbooks are without doubt the "worst case scenario" series. They're utterly joyous, and might come in very handy one day. I'm also an enthusiastic convertee to the handbag book of girlie emergencies, which has taught me everything i need to know about being girlie (cos let's face it, I'm not exactly intuitively girlie now am I?)
Incidentally, there was a thing on BBC R2 about how much the average person's weight in gold would be. Can't recall how much it was, though. Gosh that was a pointless witter.

[michaelp-j (from livejournal.com)]

I'd offer up Horowitz and Hill's immortal, though increasingly out-of-date masterpiece, The Art of Electronics, which has the distinction of being the only electronics textbook that is readable and interesting. Most other books treat the subject in a very mathematical way - they take a circuit, state some axioms and then derive their way down the page. H&H actually tells you what you want to know about a particular circuit, and then explains why it works.

[totherme.livejournal.com]

Oh - almost forgot!

The first book my undergrad tutor recommended to me, despite not being on any of the primary reading lists (that I can recall) was what he called The MIT Bible. There's lots of good stuff in there, and I lugged it around faithfully for quite a long time. Well worth having around to dip into, if you're interested in computers as science.