Weblog Without Honor or Humanity

Learn C++ in 10 minutes. While browsing the shelves at Borders I saw this amazing book that promises to teach you C++ in 10 minutes. That’s an improvement over 21 days, a previously held record for teaching C++ by the book, and a very sad comment about my intellectual abilities (took me years). Peter Norvig’s excellent essay Teach Yourself Programming in Ten Years is the best possible commentary.

A quote from Accounting for Dummies, page 67:

Be aware that a business may be tempted to deliberately overstate or understate its profit. When a business overstates its profit in its income statement, some amount of its sales revenue has been recorded too soon and/or some amount of its expenses has not yet been recorded (but will be later). Overstating profit is a dangerous game to play because it deceives investors and other interested parties into thinking that the business is doing better than it really is.

Looks like corporate America didn’t pay attention to “Accounting for Dummies” recently.

Programming Jabber. Just finished reading DJ Adam’s Programming Jabber. Jabber is an interesting but not revolutionary (IMO) technology. It’s an Instant Messaging protocol like AIM, ICQ or Microsoft Messenger. What makes it different is that unlike those proprietary protocols you can run your own Jabber server (there’s both open-source server and commercial ones). Jabber also puts emphasis on being a p2p platform instead of just limited IM service. The book is a good introduction to Jabber, its protocol and kinds of applications that can be built on top of it. Good if only to stimulate your thinking.

Stuid coding mistake of the day. Spent a day figuring out why debug build worked differently than release build. What’s wrong with this code:

bool FMyFunc(void) {…}
…
assert( FMyyFunc() );

In non-debug build FMyFunc() will never be called. Aaargh…

coder blog: You want to fight back against the system? Learn to code. Write software that propagates across the worldwide digital medium affecting the lives of millions using a board beneath your fingers. Robust, distributed, massively popular open source digital communication software is the most powerful tool at your disposal to reach and influence society.

Amen. And don’t forget to vote out all those bought politicians that continue to take away our freedoms with DMCA and other legislation hostile to free speech, fair use rights and human beings.

Philip Greenspun’s book How To Be a Web Whore Just Like Me shows how good a book can be if it’s written by an intelligent person who writes about the stuff he knows a lot about. The book talks about why to do web publishing and how to do it. While the “how” part is bound to be less and less relevant, the “why” part is timeless. The book has been upgraded and renamed to Philip and Alex’s Guide to Web Publishing. It’s even better than the first version, despite author’s self-effacing statement that it’s a 1997 classic that is sort of embarrassing now

“Television is a medium because it is neither rare nor well done.”

– Fred Friendly

One of the principles of good design is discoverability: it should be easy for users to discover what the program can do just by fooling around. Menus and toolbars add discoverability: they don’t interfere if users don’t care about them but they’re always there, visible, for users to try out new features of the program.

Emacs, in general, has poor discoverability. Most of the work is being done with arbitrary keystrokes and obscure lisp functions and the only way to find out what a given keystroke does is by reading copious manuals or using one of the obscure lisp functions to list current key bindings. This is efficient for experienced users but presents a very high barrier to entry for new users. There are two problems:

• it’s hard to find out what a given keystroke does
• it’s hard to find out what is a keystroke that does what I want

It would help discoverability if users could easily find those 2 things out. Here enters my idea: we need to add a new window that will list all keystrokes that are reachable from a given point. Normally it would be empty. When users presses “Alt” key it would list all keystrokes that start with “Alt” key. Entries should be sorted in the order of usefulness. That way users could see what a given keystroke does and with time learn other useful keystroke by looking at the list of available options.

This window could be small, permanent window (with the ability to hide/show it) or something that only pops-up when needed (similar to how the auto-completion usually works). This idea is not limited to Emacs - many programs implement key bindings as a short-cut for frequently used functions and could benefit from a system like this.

Look Ma, what an ugly coding style. Hope you’ll never code like this. I feel compelled to say something smart about coding style and here it is: coding style doesn’t matter. Put your braces wherever you want, child, and don’t worry about “rational” arguments that justify personal preference. Life’s too short to worry about coding style. Life’s too short to worry about anything.

The end of the free (LWN). An excellent site Linux Weekly News has just announced that, barring immediate miracles, they’re closing shop on August 1. Those are bad news - LWN has been an excellent site providing weekly pro-Linux, pro-open source and anti-Microsoft news with excellent, thoughtful commentary. This event, as sad as it is, illustrates a very simple economical principle: you cannot make money on something that can be obtained for free. The cost is not directly related to quality. Despite being truly best-of-the-web, popular and running on a tight budget, the people involved cannot find a way to pay the bills. The magazine has been running for some time on the dot-com wave when everybody thought that Internet was a goose laying golden eggs but when the reality kicked in the numbers says that the site cannot turn a profit. Donations have been tried but it seems that even popular site with loyal and sympathetic readers cannot make enough money that way to stay in business. And no other option seems to be in the sight. This is also a main problem of open-source when it comes to making money: by definition of open-source you effectively have to give away the software for free. If people can get the software for free, they won’t pay. Tom Lord, hard-working and excellent programmer, is finding this out the hard way. I wish him luck, I wish LWN luck, but don’t expect that they’ll have it.