Though I have been skeptical about current approaches for rich net apps, I’ve been looking at Macromedia’s Flex and Laszlo for some time to see if they come up to scratch. Of late, both have been in the news — Laszlo because they open-sourced their app server, and Flex because Macromedia announced its giving away the software to qualified non-commercial users for $9. This is all good, however, neither of these two quite leave me satisfied.

Both currently compile to Flash. Flex (I believe) is wedded to it. Laszlo, while theoretically target-independent, currently supports only Flash. Flash is known to be fast (even though non-Windows implementations have typically been a tad slow) and is used widely for ads, animations and short games; however looking through simple demos like the Amazon store shows that the the UI is far more sluggish than a standard HTML interface.

Also, both approaches currently are lousy, accessibility-wise. Even considering the work done by Macromedia to improve accessibility in Flash 7, I cannot imagine Amazon converting every page on its store to this format (and if it didn’t, there wouldn’t be any point implementing the order/checkout process in Flash, since that would mean subjecting shoppers to two different interfaces).

Then there are the developmental hurdles of declarative programming (not necessarily a bad thing; but it is unfamiliar to many developers and needs a good WYSIWYG IDE to be productive) and a costly application server sitting between your data and your users. And what does all this buy you? Why, Drag and drop! Data binding! Platform Independent Fonts! Sigh — supposedly obsolete IE has handled all of this since version 5, and has 85% of the browser market. The Mozilla crew is catching up, and the WHAT-WG process will ensure Safari and Opera do, too.

I can understand there may be a market for these products among those who need rich net apps right now, but those looking at this from a strategic point of view would do well to either target IE only, or wait for XAML (which is going to be baked into IE in Longhorn), or work with other browser vendors and the WHAT-WG to ensure that Web Forms 2.0 ships ASAP (and withstand a diversity of UIs as browser vendors work to iron bugs out of their Web Forms implementation, as they did with CSS, for a while).

Bottom line: upgrading the browser results in a far superior user experience than hacking together kludges on the server that execute on the client via plugins. And because of this, Flex and Laszlo, while attractive, look like products whose windows of opportunity are closing — fast.

Update Oct 12: Dr Dreff comments; he says Rich Internet Apps (RIAs) have a future without MS — he specifically mentions Apple — and MS never ships on time anyway. True, but Safari developers have been active on Web Forms 2.0, and his points do not invalidate my proposition that delivering your entire UI through a “presentation server” and requiring a plug-in to view it has no future when HTML itself can be extended to support modern UI niceties like drag/drop, fonts and autocomplete. In fact, if Adam Bosworth’s Caching Framework ever sees the light of day, it could, in conjunction with a modern widget set, revolutionize the way web apps are done — web users can finally get the same experience current Lotus Notes users do when they work with their apps offline.