Why is Konfabulator "Pretty?"

Bill Marriott wjm at wjm.org
Sun Dec 4 20:24:16 EST 2005


Richard,

> I maintain links to the Human Interface Guidelines for most popular 
> operating systems on the top-right of this page:
> <http://www.fourthworld.com/resources/>

This is a great resource and what an incredible service to put it together! 
Having said that, there are nearly 30 links there. :) This provides evidence 
for a few points:

1) HIG is what you make it.
2) No one agrees what constitutes correct HIG.
3) HIGs change over time.
4) Appropriate HIGs vary depending on the context.

> When I wrote about making "HIG-compliant apps", oddly enough I was 
> referring to apps.  You know -- menu bar, documents, About box, etc. That 
> stuff.

Is Revolution more appropriate than K. for constructing a full-blown app? 
*Definitely.* Is K. "better" than Revolution? No, I'm not saying that at 
all. Does K. make it easier to create these little widgets that look so 
sleek and sexy? Probably. Do those little apps themselves follow a usable 
HIG convention? Yes, I think they do. Am I trying to start a K vs Rev war? 
Absolutely not. Can Rev create an stack that looks and feels like the K. 
clock widget <200K? I don't think it is possible.

> Fortunately that wasn't the core question, which was much simpler:
>
> To what degree would RR making it easier to import images from Photoshop 
> into Rev help its market adoption?
>
> Considering that relatively few application categories benefit from the 
> novelty single-window translucent whizbangness that distinguishes 
> widgets/gadgets from the rest of the world's software categories, and 
> coupled with the sober recognition that importing graphics isn't the hard 
> part of producing such UIs, I'm not sure it would make much difference in 
> the big picture.

No that wasn't the core question. The core *statement* was a belief that 
these widgets have a modern look-and-feel distinct from stacks created with 
Rev, which is the main reason why they get attention from companies like 
Yahoo.

> 1) About five to seven years behind the curve in UI. The reason why 
> programs like "Konfabulator" get the attention is that they create 
> widgets/programs that look modern/exciting, almost "by default." The same 
> look-and-feel can be created with RunRev, but it's a lot more difficult.

Did the "aqua" look help the Mac deliver better business productivity 
software? Is a scrolling, dockable expanding whatchamacallit at the bottom 
of the screen deliver superior *functionality* than the Windows Start menu? 
Not really. But it's part of the cachet of using a Macintosh. It makes users 
feel like they are using cutting edge technology.

An development environment that could not create apps to match this look 
would not be very appealing, as the products would look "long in the tooth" 
and users would not adopt them. As proof of this, consider when Rev finally 
was able to present standard buttons that looked like true Windows XP 
buttons. This was a big deal and made a lot of people happy.

It isn't a widgets/app question anyway. Interfaces are growing more 
sophisticated and complex. Distinctions are blurring in the user's mind 
between what is online or offline, what is a widget vs. an app, what goes in 
the system tray vs. what sits on your desktop, etc. Modern apps are not 
monoliths anymore. They have little bits and pieces accessible from all over 
the place.

You're entitled to your view but my take is such things matter a heckuva 
lot. Consumers want programs they can almost "touch" with reflections, 
depth, animation, transparency. Stuff Rev needs to catch up with in my 
opinion. And developers want environments that help them accomplish this.

> If you feel strongly that it might you can put that to the test in under 
> an hour:
>
> Write the import script and post it somewhere.  If Rev takes off like 
> wildfire from that we'll know it was a great idea, and Rev can pay you 
> handsomely for that hour's work.  :)

I'm afraid my skills would not allow me to complete this task in an hour or 
even a week. Honestly, I think you're minimizing the amount of work that 
went into that for K. as well. It's just not easy. The results are amazing. 
Furthermore, even were I to write such a utility, it would not be as useful 
in Rev as it is in K. I know this because I imported the folder of PNGs from 
the K. clock and used the XML file generated by K. for the clock to apply 
all the names and properties of the objects in Rev. (Yes, via a script.) 
This shortened the work as much as possible. But I could simply not 
manipulate the graphics in Rev as fluidly as the JavaScript libraries in K. 
did. They achieved amazing results.

> Photoshop output can be used in any app (as can output from Fireworks, 
> which is arguably superior for making UIs since it was designed for that 
> task from the ground up rather than morphed into it through accidents of 
> history), regardless whether the language that drives the app is C++, 
> Java, HTML, or Transcript.

That's not the point, either :( It's tedious and time consuming to export 
individual objects, import them into your development environment, name 
them, position them, layer them. The script from K. distills what could be 
an hours-long process into a single menu click. It's quite useful. Moreover 
it's an innovate concept, that one could design "outside the box" using an 
"artists" program for user interface. Microsoft has something similar going 
with Visio where you can rough things out in Visio -- whether it be user 
interface, web sites, or database schema -- and then generate the template 
for the "real thing" from those schemas.

> Sexy interfaces isn't something that comes from the runtime tool; it comes 
> from talent applied to a graphics tool.

If that were true, we wouldn't need a palette with tabbed folders, radio 
buttons, checkboxes, etc. Just draw them in Fireworks and bring them in. 
That toolbox makes it easier to design apps that fit in with the look of the 
operating system. The bigger the toolbox, the more likely it is that a 
programmer can create something "sexy." Look at how the UI for Microsoft 
Office has evolved. It's arguably "sexier" than prior versions. Revolution 
does not have the built-in tools to create an Office 2003 interface, though 
it can create an Office XP-like interface.

Cheers,

Bill 






More information about the use-livecode mailing list