This book is exactly what it says it is, providing a tour of distributed object technology: what it's good for, why we need it, and how it may/will happen. It can be very cute. As a teaching device it explains everything to a Martian named Zog who, like most of us, is befuddled by the explosion in acronyms, activity, and interest in distributed object technology. Yet it does a good job explaining things for the quasi-technical user, e.g. a technical manager who must know whether or how to deploy object technology but who isn't the programmer who must actually do the deploying. That is, it's a good overview, not a reference manual.
One of my favourite features is that they ``put their biases on the table'' by placing their outright editorializing in Soapbox sections, typeset with a grey background to set them apart visually. As with any hot topic, there are bound to be strong opinions. It's best to keep the inevitable preaching separated from the necessary informing. Thank you.
But the book was just not dense enough for me. I found myself waiting for the authors to get right down to the meat. I'm sure it's there. But I might have learned more faster from a good Usenet FAQ. So I'm twisted. I would return to this book later if business interests warrant it.
On the other hand, it was published in 1996, i.e. the Cretaceous era in distributed object terms. So it misses out on the Java and JavaBeans revolution. But it appears to cover CORBA, OpenDoc, and Microsoft's OLE/COM well. Oh, but OpenDoc appears to have been abandoned (largely in favour of Java and JavaBeans), and Microsoft has changed to ActiveX, etc. I suppose you should expect a new edition of this fine book pretty soon.
I'd recommend this book for technical managers, but please do supplement it with knowledge of Java and JavaBeans.
Larry Wall created Perl, and this is the definitive guide to the language, unless you trust the source code more... This is a quirky book, with much tongue-in-cheek humour. In some places it reads like a Douglas Adams book. There's even a section entitled ``Wake Up!!!'', (see page 18).
If you need the real low-down on this popular kitchen-sink ``system administrator's language'', this is the book to get.
NB: The book covers version 4.x of Perl, and Perl 5 has already been released and reportedly breaks some old scripts. So beware. I haven't played with Perl 5 myself, so I don't know how different it is. I imagine almost everything is the same. So get this book and read the release notes to Perl 5.
This slim volume describes in great detail how applications should behave: how selections ought to be made, how menus behave, where the menu bar goes, etc. It's interesting to read from a historical perspective: GUI programs that don't behave in the way described now feel unnatural. That's a testament to the influence of this whole movement toward a unified style for applications.
Interestingly, one thing that's missing from the guide is a description of those shortcut buttons just below the menu bar that have become popular: they're called "SmartIcons" by Lotus, and I can't remember what by Netscape and Microsoft.
Did I learn something? Yes. You can extend a selection to a non-contiguous object by depressing CTRL and selecting the new object. This is described in the style guide, but I haven't yet seen it implemented. It's a good idea. (Late breaking news: some programs use SHIFT in this manner. For example, Lotus WordPro96 lets you select more than one frame this way. This is useful for grouping frames together.)
If you write GUI software, then this book may be indespensible.
I'm waiting for the fourth volume in The Art of Computer Programming series by Knuth. It's on graph theory and combinatorial computing. In the meantime, his The Stanford GraphBase: A platform for combinatorial computing should keep us busy. It's an excellent example of the fruits of careful programming under the CWEB system. And Knuth says its the final preparation for volume 4.