Pet Theory

Flex 4 a.k.a. Gumbo will treat the major pain point in Flex development–the workflow between designers and developers–by allowing Flexers to easily make complete, detailed component skins as well as non-standard interactions and effects.

And what about Flex design? Will it be impacted? It’ll be a year or more before we know, but this post makes two predictions:

1. Flex designer will fulfill the old unfulfilled dream of Flash designers that UI will become more game-like. Flex will also supplant a lot of Flash development while still avoiding Flashiness.

2. More important, Flex designers after Gumbo will focus on incorporating the best of hypertext, the dominant and transformative mode of the web that Flash and Flex designers tend to discount.

Here’s a slideshow to help explain these predictions. It provides general context for thinking about GUI. Afterwards, I’ll extend the argument with Flex-specific details.

Flex v. Flash.

Flash is great for games and banners. For complete sites, it mostly has nothing to do except animate transitions. Everyone’s visited the archetypal aggravation: a movie site with slow, pretentious transitions that lead you to….wait a few seconds…some text and pics. Flash designers have duly moderated their broadcast impulses for the sake of usability, but in their hands, the impasse between WATCHING and LINKING has proved–after nearly a decade–equally sterile in corporate sites and student portfolios.

Design-wise, Flex is a step forward because it gives Flash some content–data–to work with, and shifts it into a domain where its powers can be applied productively. Transitions actually make sense when users are in consist DOING/STAYING mode and managing space becomes an issue. Selectively drawing on the physical world to construct an application starts to make sense, too. Working on stuff, users might want to push it aside, gather it together, stretch it out, or stick it to the metaphorical wall.

Gumbo (Flex 4) v. Flash

Gumbo cleanly separates out the underlying data and the visuals in Flex. With the visuals now a thing apart, they can be skinned, customized and propelled by their own logic. This is a many-splendored turn for the better, which I explore here from a programmer’s perspective.

The emphasis in promoting Gumbo has naturally fallen on how the separation of data and visuals allows for greater visual expressiveness This is only half the equation, however, and ultimately less than half. More important, now that the component data has been freed from any particular component visuals, it becomes easier to conceptualize various kinds of data for a Flex project: ranges as timelines, trees as branching inquiries or dynamic pages, lists as narratives, etc.

Visual expressiveness in this context doesn’t mean demo-wow-itude as much as flexibility for visualizing an enormous range of content.

Flexibility could be wrung from pre-Gumbo Flex at the cost of baked-from-scratch components. But it was not easy, and easiness has a way of freeing the imagination.

With Gumbo, more projects will become imaginable. Many of those projects would have been slated for Flash development previously. (I’ll be surprised if there are more than a rump of pure, non-game AS3 developers in two years.) So given that Flex designers will soon have Flashier capabilities, and will be working on Flashier projects, will their designs become Flashier?

Excepting the inevitable spinning-globe demos, I doubt it. Compared to Flash, Flex is well grounded in data and user control. Instead I expect some really interesting, really helpful ways do and explore stuff.

Flex v. Ajax

Right now Flex and Ajax occupy slightly different niches in the web-application area. Flex for the moment looks more like desktop software, while Ajax applications look more like web pages. Both center on DOING, while Ajax is pulled strongly towards the LINKING and Flex is pulled less strongly towards WATCHING.

Providing differentiated front-end tools is a impeccably rational strategy for Flex and Adobe, and it makes sense since they have an advantage in the BROADCAST and the APPLICATION areas. Ultra-reasonable Flex evangelists regularly go as far as recommending Ajax for “content”-oriented sites.

But here’s the thing. Hypertext is good. It seems mundane only because it has assimilated into users’ DNA so quickly and completely. But the page+link+page structure is addictive to users, because it offers associating power with quick, undoable and clear interactions. If an Ajax developer finesses a page refresh now and them, that hardly detracts from hypertext.

And Ajax design is consistently good. When Ajax first went mainstream I remember taking solace in the fact that the design was often awkward, “interactive design is more difficult than it looks, no?” But its skip-intro period lasted a chilling micro-second. Incredibly well-designed Ajax apps demonstrating a well-grounded balance of STAYING and GOING now seem ubiquitous.

And lastly, the traditional web-development community has the numbers, the talent and the vision thing. How long will it be before Flash’s particular media capabilities are picked off, one by one?

I don’t know how supportable the current differentiation is over the long term. Flash and Flex have improved by leaps and bounds, but so have HTML, Javascript, and CSS. In fact, they are encroaching steadily on Flex and Flex territory…so why not push back into HTML territory?

My guess is that LINKING will exert increasing gravitational pull on both DOING and WATCHING in the coming years. Hypertext is the up-and-comer and its course is still running. Users will want to jump around apps with inline links, and undo or redo actions with a history button. Media will become a kind of hypermedia: bookmark-able, link-able and comment-able. Journalists and educators will offer less linear and more jumpable content.

Flex is well positioned to produce this new, more expansive kind of content, if Flex designers are willing to learn from hypertext. Flex is already a decent tool for creating relatively CLOSED and deliberately constructed experiences that integrate data, media and interaction. Gumbo will make it better than decent, and designers dedicated to a more relatively more OPEN user experiences could really make Flex sing.

And now? Well, currently, if you create linked text, the text has to be selectable, so when the user clicks the link, they often select the text instead!

Flash 10 (which Gumbo relies on) promises to fix text: razor-sharp fonts, multi-column text fields, inline images, open glyph APIs. Finally, text will be a full-fledged part of the platform. This is great for multimedia capability–text is naturally the dominant, mediating media in multimedia–but even greater as a solid step towards integrating hypertext-like functionality.

These new text capabilities might in the long run turn out to be more decisive than the new skinning model. Next steps might include selection-dependant background loading and enhanced history, both of which would allow for more direct user control.

I live in Portland, Oregon, and I’m looking for work. More about me.