By John Gruber
WorkOS: APIs to ship SSO, SCIM, FGA, and User Management in minutes. Check out their launch week.
Assorted follow-up points on the Dashboard-vs.-Konfabulator quote-unquote “controversy”:
If I regret anything about the original article, it’s that it might be construed as my taking Arlo Rose and Perry Clarke to task, personally, whereas my intended targets were the bystanders loudly crying foul. To be clear, it wasn’t Rose and Perry who started or unfairly fanned the flames.
The “Cupertino, Start Your Photocopiers” banner afront Konfabulator.com — that’s just good clean marketing fun. What I did use against them were Rose’s histrionic comments to CNet’s Ina Fried (e.g. “It’s insulting, is what it is,” Rose said in a telephone interview. “They could have at least offered to work with us or to buy it.”) But I realize how easy it is to have a quote taken out of context by a publication like CNet, well-known practitioners of the hype-any-conflict brand of journalism.
I have nothing against Rose or Clarke, and I’m certainly not rooting against them.
I wasn’t the only person who presented desk accessories as prior art. Arlo Rose, however, rejected this analogy in an interview at Geek Patrol:
Matt: The point has been raised that Apple had a similar concept to the Dashboard in Desk Accessories back in the original OS. Do you think this at all validates the Dashboard as their own idea?
Arlo: No, People keep bringing up Desk Accessories. Desk Accessories were just mini applications accessed via the apple menu. The idea behind K is not so much about the basic Widgets as it is about empowering people to easily create their own and this is the key part that Dashboard duplicates.
Rose is ignoring where the comparison is apt, which is in terms of scope. Konfabulator widgets and Dashboard gadgets are very much similar in scope to the roles taken by desk accessories. They are very much a modern-day equivalent. By Rose’s logic espoused here, he wouldn’t have had a problem with Dashboard if it were harder to develop for.
Here’s the connection. Desk accessories were a set of APIs that allowed early Mac developers to create tiny, focused, single-purpose apps that could run alongside full applications. They were easily accessible via the Apple menu (in fact, that was the only thing the Apple menu was used for originally). Because of the insanely tight resources — 128 total KB of RAM, 400 KB of disk space, and 8 MHz CPUs — they were inherently difficult to develop. You try writing a sliding puzzle game in 600 bytes. (Note that this very article is already 1.6 KB in length; 1.7 including this parenthetical.)
Dashboard is a set of APIs that are going to allow Mac developers to create tiny, focused, single-purpose apps that run alongside full applications. They will be easily accessible via Exposé shortcuts. Because these APIs are based on web standards implemented in WebCore, they will be inherently easy to develop.
Others have compared Dashboard to Hypercard; e.g. Dori Smith:
It’s about all of us, programmer and non-programmer, scripter and non-scripter, being able to, for the first time, create little useful applications on our own Macs, using simple code and markup (i.e., HTML, CSS, and JavaScript; i.e., not AppleScript) that we probably already know something about, and then being able to share those widgets with anyone who has Tiger.
I would argue with her premise that Dashboard will allow “non-scripters” to create anything at all — if you’re not a scripter, it’s going to be tough to write JavaScript, and you don’t write any JavaScript, your gadgets aren’t going to do much — but, still, the comparison to Hypercard in terms of ease-of-use is apt. Many times more Mac users will be able to write Dashboard gadgets than can write GUI software for their Mac today.
The difference, of course, is that Hypercard was its own world unto itself. Hypercard used its own scripting language (Hypertalk), which, no matter how easy you thought it was, was something new to learn, and something which didn’t work anywhere except in Hypercard. And, of course, Dashboard will let you use color.
Dashboard gadgets are like desk accessories in terms of scope — but they’re different because they’re easier to create. They’re like Hypercard stacks because they’re easy to create — but they’re different because they’re based on cross-platform web development standards.
Clearly, much of the same thing can be said of Konfabulator — that it’s relatively easy to develop for and is based on a standard language like JavaScript. I never argued that Dashboard isn’t doing very much the same thing as Konfabulator. But nothing in what they’re doing is new. My point boils down to this: making it easier to write client-side software using scripting languages is clearly something that makes Mac OS X a better platform. That Konfabulator offered something similar first doesn’t entitle them to anything.
Another argument against Dashboard is that its widgets “look like” Konfabulator’s, that regardless of their implementation differences it’s the fact that they share an unmistakable similarity in terms of look-and-feel — that that’s what Apple has taken without recompense.
Widget aesthetics are indisputably a major factor in Konfabulator’s appeal. And while none of the example Dashboard gadgets is specifically reminiscent of any of the default Konfabulator widgets — it’s clear that with Dashboard, Apple is aping the overall gestalt of Konfabulator’s widget aesthetic.
But once again, this isn’t something new to Konfabulator. The gestalt of widget aesthetics is very much the same as the gestalt of skinning/theming — a craze that started many years before Konfabulator existed. E.g., most of the “faces” in Panic’s Audion gallery could easily pass as Konfabulator widgets or Dashboard gadgets — the irregular window shapes, the transparency, the alpha channel blending. Same for Winamp, whose “skins” probably deserve to be recognized as the trailblazing example.
What sets Konfabulator’s widgets apart is that they’re extraordinarily well done. Arlo Rose is a hyper-talented visual designer; his widgets possess a certain je ne sais quoi. But they’re still just skins, albeit extraordinarily well-done ones.
I don’t think it’s any surprise that Apple’s gadgets are well-designed too.
Any remaining argument that Apple has “ripped off” Konfabulator hinges on the idea that Konfabulator is greater than the sum of its parts. The argument going something like this: No, they didn’t invent JavaScript, nor the idea of a scriptable runtime environment for little single-purpose applets, nor the idea of arbitrarily-skinnable UIs — but Konfabulator was the first to put these ideas all together.
I see Konfabulator as a well done implementation of existing ideas. If you see it as something deeply innovative, then we’ve reached a point where we’ll have to respectfully disagree.
As for the nomenclatural confusion regarding “gadgets” vs. “widgets”, here’s a blurb from an email I recieved from a WWDC attendee:
According to what I heard, the initial name was “gadgets” and Steve vetoed it when he heard it and so they used “Widgets”.
Obviously, this is merely scuttlebutt, but it does ring true. Whatever the reason for Apple’s apparent switch to and appropriation of “widget”, it’s unfortunate. To the casual observer, it certainly contributes to the perception that Dashboard is Konfabulator knock-off, and, worse, it’s unnecessarily antagonistic.
What concerns me is the message the Dashboard thing sends. It goes like this, “If you come up with a good idea and develop a successful product, we might copy it and bundle it for free with the OS.”
In other words, you could be penalized — heavily — for doing a good job, for doing exactly what every developer works very hard to do.
It would take so little for Apple to have made the Konfabulator folks happy. Some money, some recognition. (“Little” is relative: little to Apple, big to the Konfabulator folks.) And it would let other developers know that Apple cares about OS X developers, that it wants people to develop for OS X, that it’s safe to come up with great ideas and great products.
Simmons has an interesting point here, that making Konfabulator’s developers “happy” in the face of Dashboard would indirectly be good for Apple too, by easing the concerns of other Mac developers.
My problem with this is that it’s indistinguishable from charity. If Apple saw Konfabulator as a good idea but not a good implementation — what would they have been paying for? Good will? Karma?
Where does the line get drawn? Should Apple have bought the rights to iCab before shipping Safari?
The truth is, Apple does buy products from third-party developers. A former Apple product manager told me via email, “The way [Apple works] is to buy technology that is better or more promising than what they may have in-house, and in the absence of such technology, to invent their own.” Some recent examples: Final Cut (purchased from Macromedia), SoundJam (now iTunes), LiveType (purchased from Prismo, now part of Final Cut Pro), Spruce Maestro (now part of DVD Studio Pro), and DVDirector (now part of both iDVD and DVD Studio Pro).
What Apple doesn’t pay for are ideas. I suppose Simmons’s argument is that it might be a good idea if they did — but I’m unaware of any company that does.
In the comments of his weblog, Jerry Kindall relays an appropriate anecdote as to the value of ideas versus implementations of ideas:
Science fiction writer Larry Niven once said, perceptively, that if a fan gives a writer an idea, and the writer goes on to write a story from it, the writer owes the fan a beer, but no more. Because coming up with ideas is the easy part. What matters is what you do with the idea, and I think Apple has done more. Arlo and Perry do deserve a beer for their trouble, but I don’t see what else Apple could owe them.
Using thousands fewer words than I’ve blown on this topic, Michael Tsai simply nails it:
Developing for Dashboard is nothing like developing for Konfabulator. There was almost zero chance that I would develop a widget for Konfabulator. I’d have to learn a completely new platform, and anyone who wanted to use my widget would have to buy a license for that platform’s runtime. Dashboard widgets, on the other hand, are Web pages. I already know HTML and CSS, and I have a collection of tools for working with them. The runtime is free. If I want to use Cocoa or Java, I can do that, too. […]
The Konfabulator developers are in an unfortunate situation, partially of their own making, but what’s getting lost in that coverage is that the Mac is now a platform for writing client-side Web applications.
Previous: | The Last-Minute Push |
Next: | Status |