This old paper still seems relevant: http://oraclemagician.com/white_papers/book_article.doc

No flamebait intended.

Without wishing to pry, is there any useful money in writing a technical book?

From what you wrote, it all looks to be a bit of a punt, with regard to revenue streams.

By the way, have you thought of producing some sort of visual Tuning course on, say, CD-ROM/DVD/whatever?

Sometimes pictures really are worth a thousand words. A company called CBT Nuggets does that sort of thing (as do several others). I bought a (gasp!) SQL Server course from them a few years back: it’s a very useful format, and I learnt much faster, and with less uncertainty, than by using books alone.

Food for thought!

Saves me lugging a hundredweight of books around, people forgetting to return them, spilling coffee over them, pages falling out, etc.

Not exactly low cost, but having a rather brilliant selection of books on all topics that I may want to know something of (e.g. UNIX, TCP/IP, Java, C++, etc., etc.), available wherever there’s an Internet connection, it seems good value. One can also download a few titles per month, to print off, I suppose.

Businesses can buy corporate licences, too, which, I think, is a no-brainer for any business that wants its staff to do well in their work.

That should have been “2nd Volume” – which is not yet complete.

I hadn’t come across Safari before – I’ll have to have a look into it. Once the book has gone to the publishers the subsequent appearances are up to them really; there have been a couple of deals on the book that I haven’t heard about – beyond receiving items with no apparent origin on the Royalty statement.

Any chance of the 2nd edition being made available on O’ Reilly Safari?

The 1st edition is there.

Incidentally, if affordable, I recommend Safari to all. It’s a genuine boon.

It’s one of those comments where I should have left it out completely or written half a dozen pages explaining it. I make the comment in the 2nd edition volume of the book.

“A little while ago I wrote some notes on why monitoring the buffer cache hit ratio (BCHR) was a pointless exercise. ”

But in your CBO Fundamentals book (P. 25), you say:

“… plotting trend lines from regular snapshots can give you important clues that something is changing, or that a performance anomaly occurred at some point in time.”

What am I missing here?