No time to say more than the above to this outstanding response from Dave Winer. Great stuff coming in from the rest of ya'll. Thanks, and keep it coming!

BTW Dep't: The "no Moore's Law for software" question comes from Ellen Ullman, not me. I just thought it was provocative, which turns out to be true.

And soon. The finished deadline is December 15.

Like I said earlier, I'm writing this piece called Why operating systems suck (and what we're doing about it) for a bigtime magazine. Whether they know it or not, here are my sources, in no particular order:

• Phil Hughes
• Craig Burton
• Dave Winer
• Don Norman
• Jakob Nielsen
• Stewart Brand
• Esther Dyson
• Kevin Werbach
• Jerry Michalski
• Jim Allchin
• George Lakoff
• Andy Hertzfeld
• Steve Jobs
• Bill Joy
• Ellen Ullman
• Richard Gabriel
• Charles Roth
• Richard Saul Wurman
• Neal Stephenson
• Linus Torvalds
• Eric S. Raymond
• Tim O'Reilly
• Larry Wall
• Don Marti
• Steve McConnell
• Michael Polanyi
• John Seely Brown

Anybody I'm leaving out? I've already talked with some of these folks, but I need to hit a lot more — and I know I'm forgetting more than a few that matter. Only one of them is dead, so that keeps the list of calls to make on the long side. If it needs to grow by your name or some other obvious omission, I'm sorry I forgot to put you and the others on the list. Get in touch.

One of my questions is "Why is there no Moore's Law for software?" The first response, concerning documentation, comes from Charles Roth. The question itself comes from Ellen Ullman, who spoke at the PopTech! conference in Maine last month. In a very provocative talk, she suggested that a big problem with software is that the people writing it are often far less terrific at understanding the problems they are solving than they are at writing software.

Well, there's the phone. More later.

discuss

The amount of resistance I get when I try to upgrade someone from WordPerfect 8.0 to 9.0 is amazing sometimes. Users don't handle change well at all, so a software developer who is attempting to do something truly revolutionary is going to have a hard time getting their users to follow along. Just renaming and moving a few of the commands on the menus is enough to frustrate most of my users, and they'll start griping about how they always knew where it was in their old faithful version of their software. I prefer Microsoft Word to WordPerfect, I know them both, used WordPerfect since version 4.0 and Word since version 2.0, but try to get someone to use Word if they're a WordPerfect user, and it's like pulling teeth. I'm sure this argument extends to many other software areas, this is just the most common one I encounter.

The other users that are frustrating are those who want to upgrade, but they don't know why. The kind who say "I need Windows 98 put on my computer, that's what I have at home, and I want the files to be compatible". No matter how many times I explain that the files are compatible between them, they don't seem to catch on. It's a case where they want everything to be the same, they don't want to have to think about what they're doing. They learn how to use a computer through rote memorization of basic tasks, not through an actual understanding of the processes involved.

Technical users, "geeks", support personnel, programmers, etc. understand how the computer works, they can adapt to minor and even major changes in less time than end-users. Take HTML editors, I'm sure I could pick up Adobe GoLive, DreamWeaver, BBEdit, FrontPage, etc. and make basic web pages within ten minutes in any of them. I haven't used all of those packages, but they do the same thing, I only have to adjust to a difference in terminology or layout or icons.

For the typical end user, it's a whole different ballgame, they just know that when they click this shiny button, it does what they want. They don't think about what the term means, what other terms mean similar things, etc. Any change to their software will alienate them. It's this alienation that prevents new and better tools from replacing existing tools that are adequate for the job. Only when a tool becomes inadequate will the user accept a replacement.

Another example from my daily life is the e-mailer PINE. PINE still runs on the VMS systems here which double as the e-mail servers. Trying to explain to someone how to read an HTML formatted e-mail or easily retrieve an attached file can be a nightmare, but they don't want to switch from PINE to Eudora, Outlook, etc. because they're completely lost in those applications. As technical people, we find that hard to imagine, it's got big shiny buttons with words like "Reply" and "Forward", how hard can it be? The users don't think in that level, they don't analyze what they're seeing, it's rote memorization.

I'm not criticizing users, if it wasn't for them, I wouldn't have a job. I'm just saying that the same rules don't always apply to them as apply to you and I, and that's what holds back revolutionary changes to our software. Mac OS X is going to be a good example of this, many of the long-time Mac users are going to be angry, lost, and confused by the new OS, even though most of us are excited about the technical improvements like pre-emptive multitasking and protected memory.

Just my two cents. :)

-Seth

discuss

The terms mapping and packing are introduced in Thinking about Thinking.

Packers are those users you mention who "learn how to use a computer through rote memorization of basic tasks, not through an actual understanding of the processes involved." Typically packers become lost when they don't have a "script" for the current situation.

Mappers are the people who "can adapt to minor and even major changes", by maintaining a mental map of how things work, referring to it as necessary, and happily discarding and rebuilding chunks of the map which no longer reflect reality.

As I say, it's a generalisation, but it makes a useful lens nevertheless.

Whether mappers are better or smarter than packers is an open question, but I suspect that packers make better accountants than programmers, and vice versa. Different horses for different courses?

discuss

People who prefer "Sensing" are equivalent to "Packers" and people who prefer "Intuiting" are equivalent to "Mappers"

I'm a Mapper from way back and it always brings me up short to deal with Packers who don't want to know anything below the surface.

Rely on a Packer though, to notice when the environment changes. I need to be told that the room is too hot or cold, for example, or even if something has been moved in a room.

discuss