Tuesday, May 9, 1995 by Dave Winer.
The dynamics of the web are completely different from email. Your email box is your private space. As I've said, it's an honor to be allowed to share that place with you two or three times a week.
The web, on the other hand, is more like television. Don't like the show? Switch channels! Because you have more freedom, so do I.
A few of weeks ago I added a new section to my website, things that aren't going out over the DaveNet email channel. [This site is no longer on the air. DW 5/97.]
One of the pages in my personal section is a list of my favorite movies. No need to list them here. I just added another one to the list today -- Being There, with Peter Sellers, released in 1979. A beautiful movie, way ahead of its time, with a wonderful title and cast, a light plot, but it makes a huge statement -- that reality is less than we think it is. That gardening and loving and simply *being* is enough. That everything else is just a big misunderstanding. Chauncey Garndener, played by Sellers, becomes a central figure in world politics, totally by accident. The ending of the movie delivers the message that to be an angel you just share your love and speak the truth. The simpler it is, the better.
Sometimes the truth is so simple it's hard to understand. It's taken me years to understand the Sting song that goes like this -- "If you love someone, set them free." A profoundly beautiful statement, because, from one view, it totally defines the concept of love. Love is about sharing creativity and freedom. Love makes you want to create, and creativity is an act of love. When you try to capture love, to hold it, isolate it, control it, to own it -- it disappears and is replaced with stagnation, boredom and eventually misery. The opposite of creativity.
I've often wondered if software development is a right-brained or left-brained activity. So analyitic; at the same time so intuitive. This question stops me, I wonder -- was VisiCalc an act of love or an act of analysis? Was Windows? How about PageMaker? Can you feel the defiance in Quark XPress? I can! I've decried the head-trips that generated products like Lotus Symphony and OS/2. And note that OS/2 didn't become a contender until IBM got angry, and allowed their personality, such as it is, to flow into the product. Yes, I believe there's more love in software than we previously dared to imagine.
On the other hand, there's a lot of perceived security in them numbers! I thought that spreadsheets were creative tools until Bob Frankston, firstname.lastname@example.org, one of the co-inventors of the spreadsheet genre, told me that many spreadsheet users do the math on a calculator, type the numbers into a spreadsheet, and print out the result. They don't know about formulas. Wow! No what-if. No play. No fun (spreadsheets *are* a lot of fun...).
Conclusion: even if Dan and Bob poured their hearts into VisiCalc, many of their users hold their love in reserve, and are unwilling to trust the software. It's merely a presentation vehicle. Bob pointed out that Borland had completely captured this idea in Quattro Pro by integrating a full-featured presentation program. Many spreadsheet users view them as CYA tools. Forget passion. Keep the boss happy, keep your job.
I discovered the same thing in 1986, when I shipped MORE. Before then, outlining was a small niche, it appealed to total right-brainers, despite the misperception in the software industry, at the time, that to use an outliner you had to be a complete detail nut, a very left-brained sort of person. Our earlier products, ThinkTank and Ready, appealed to intuitive people. The ability of the computer to assist the organization of ideas relieved the human being from being an idea organizer. It can be frustrating to have thousands of ideas if there's no way to capture them. With ThinkTank there was less frustration. You could free-flow your ideas into the computer, evaluate them, organize them endlessly. It seemed a shame to print them out, because once they were stored on paper, they became rigid and hard and un-malleable. It seems Sting would have something to say about this -- "If you love your ideas, set them free." Uhh hhuh. Yes!
Then came MORE, a product name chosen because we had no idea which new feature (there were many!) would be most appreciated. It turned out to be bullet charts. Type in an outline, flip a switch and voila! a beautful color presentation. Want to switch the order of two slides? Flip the switch back, press cmd-U, flip back. Fast, easy, high leverage, fun and (ugh) your ass is covered!
When we figured this out, that CYA was our main selling point, we switched our ad strategy. Huge picture of a boss, taken with a fish-eye lens to accentuate the size of his nose. An angry, maybe even hateful look on his face. A caricature of *your* boss. Underneath the picture we said "He has three questions. You have two answers. You need MORE." We laugh. Ain't it the truth! The message -- if you used this software you could keep this guy from firing your ass.
MORE was at least two products inside one app. Some people liked MORE's outlining but had no use for the bullet charts. Some people liked our bullet charts but didn't like the outlining. Our salespeople came back from the field with the request from customers that we unbundle the two, and give them a choice. Yes! I agree now and I agreed then. But there was just one small problem -- there was no way to do it!
Lesson learned: integrated software is a dead-end. Look for another approach.
I remembered a speech that Bill Gates gave at a Roger Von Oech conference in Palo Alto in 1983. The question -- What About Integrated Software? It was a hot issue because of Lotus 1-2-3, which showed that presentation graphics and spreadsheets were a natural fit. Bill said that technologically, Lotus was barking up the wrong tree. He said the right way to do it was to make a spreadsheet program and a graphics program and link them together with a scripting language. Yeah! [Bill often figures these kinds of things out before a lot of us do. Go hear him speak when ever you can, you'll always get new ideas you can use.]
So, in 1986, when my customers were asking for two separate programs, the answer was there, but the technology was missing. After selling my company to Symantec in 1987, in 1988 I left to start working on the missing technology, which eventually shipped in 1992 for the Macintosh. We called it Frontier.
An aside, it's interesting that one of the seeds for this product came from Gates, and I unwittingly transferred that seed to Apple. The conventional wisdom, that all good ideas flow out of Apple only to be commercialized by Microsoft, is wrong. Here's an example of the reverse flow.
Frontier is many things, but the central idea is that if you have a competent scripting language that has its hooks into all applications, that the temptation, or need, for integration is less. If Frontier existed in the mid-80s, we could have unbundled MORE, using scripts to connect our outliner to our presenter, and presumably opened the market for competition. A more rational world for our customers, and probably a more rational world for MORE too. MORE would have been useful as an email organizer, as a DTP text composition tool, or as an HTML editor, and probably a few other things if it had been opened up. It's still possible, and it still would be a good idea.
At the center of Frontier is a C-like language called UserTalk. I had originally wanted to call the language Juicy, because it is C-like, and because both Doug Baron, email@example.com, and I are Jewish. But decorum prevailed and we went with the more predictable (and boring!) UserTalk name.
UserTalk is different from C because it's totally dynamic. All binding takes place at runtime. The symbol table is fast, but it's also hierarchic, persistent and disk-based, with a huge 31-bit address space. And it's browsable, and editable (even while scripts are running). It's inherently a multi-threaded environment, you can launch as many scripts as you like, and they can communicate with each other: there's a vast shared name space. And each script can have its own private name space too. Scripts can call other scripts allowing you to build layers of functionality. Scripts can run from the desktop and from the menu bars of apps that support our menu sharing protocol, which has become a standard in the Mac developer world. Scripts can run in response to messages from other apps, and across the net, a feature I'm using heavily in the connection between Frontier and Netscape.
Frontier has a bunch of first-time features. The one I am most proud of is its outline-based script editor. It was my third try at a structured program editor, and this one really works. You can collapse the body of a loop, for example, and when you move the loop, you move all the statements nested underneath it. It's right. It's beautiful. I love it! This is how program editors should always have worked. Someday all program editors will work this way, IMHO.
Financially, I could have written checks forever to keep Frontier afloat as a commercial product. But here's the problem -- for whatever reason, our sales had dwindled to an insignificant trickle. We had gathered a loyal following, but they were a small minority. To really have a head of steam, you have to appeal to at least a large minority. It's even better if you have a consensus behind you, that your product is perceived to be a standard. There's safety in standards. I've learned that people will sacrifice performance, features, maybe even reliability, in order to feel safe in their choice.
I've been trying to resolve this contradiction for at least two years. At some level I always understood that in order to be all that it can be, Frontier had to be free. Like all thruths, this one is a pun, it's equally valid no matter how you look at it. To be free means free of cost, and free to be itself, independent of me, its primary owner. If you love someone, set them free. Correct.
So -- for the last couple of weeks I've been working on a new release of Frontier, codenamed Aretha, streamlined and updated for 1995, incorporating much of what I've learned about script writing and leading a community of script writers. At first it was a difficult process, because visiting the old parts of Frontier was like visiting an old wounded friend. I snipped and trimmed, cut out the stuff that didn't prove useful and focused on new frontiers, like the connection between Eudora and Netscape. I saw that AutoWeb was going to be truly useful now that everyone would have an opportunity to customize its scripts. I could use important features that weren't present in our Runtime. I could use Frontier's built-in outliner to store lists of email addresses for DaveNet-style essaynets. It's cool! Most important, I could feel Frontier becoming free, and in a real sense, in doing so, I also became free.
I think this gift will re-enforce the community of co-related Macintosh interests that have already built on Frontier, and bring it to new audiences, especially people who run web servers on Macintoshes. Maybe it leads to a Windows and Unix version of Frontier? I hope so too. Can we do what Sun's HotJava proposes to do? Yes! How about General Magic style agents? You bet! Do you have to use Netscape's bare-bones email window? Not any more. Eudora is scriptable, and so is Netscape. Connection made. Someone *is* home! I think it's going to work. More important, it *feels* like it's going to work. There's the right brain kicking in.
All that is great, but the greatest benefit of freeing Frontier is that it allows me to share the primary result of my life as a software developer with anyone who wants to experience it.
Finally, please don't worry how I'm going to make money. That's my problem, not yours.
Let's have fun!
PS: CYA stands for Cover Your Ass.
PPS: MORE is not an acronym.
PPPS: The left brain is responsible for analytic thinking, the right brain specializes in intuition. I always have trouble remembering which is which.
PPPPS: David Siegel, firstname.lastname@example.org, is producing the website for the fourth U.N. World Conference on Women, to be held in Beijing Sept 4-15. They are expecting 300,000 visits per day when the website is on the air. He needs help. Send email or check out his website at http://www.best.com/~dsiegel/.
PPPPPS: Unfortunately, largely due to my current focus on the Aretha project, I will not be at Apple's WorldWide Developer's Conference this week. I've gotten a bunch of email from people wanting to meet with me there. I am in town, so it's possible that later in the week I'll have more time. Send email if you're in San Jose for WWDC and would like to talk face-to-face.