Laszlo demo: The write once/run anywhere (Flash,DHTML,Java,Silverlight) RIA dev tool?

If you're like a lot of Web developers who want to turn their static Web sites in to something more rich and interactive (these days, referred to as Rich Internet Applications or RIAs), one of the challenges is in choosing a platform to target. The main choices are Adobe's Flash, the more standard DHTML/Javascript, Sun's Java (actually JavaFX), and Microsoft's Silverlight platform. But, given how much of a strategic investment is required (and how painful it might be to go back to square one in the event your first choice doesn't work out), which of the four should you pick? Or, will you have to write to multiple platforms in an effort to make sure your application is available on the broadest range of end-user devices. For example, if you pick Adobe's Flash as your target and suddenly realize that you want your app to run on the iPhone which requires DHTML, the long term costs of maintaining two (or even three codebases) could overwhelm you or your development team.

David Temkin, CTO and co-founder of Laszlo Systems, thinks he's got the solution to the problem -- basically a middleware-for-middleware product that allows Web developers to write their applications in an XML/Javascript-based RIA description language known as OpenLaszlo that in turn can be compiled for any or all of the four primary RIA platforms in such a way that the end-products (the RIAs) are virtually indistinguishable from each other.

To prove his point, you can see Temkin demonstrating an RIA in the reviewcast (above) that was developed once in OpenLaszlo and deployed twice: once to Flash and a second time to DHTML. The application is a rather simple app called LZPics and the user interface looks identical in both cases. In claiming that OpenLaszlo can make the four target platforms do just about anything (except play video), Temkin says that downsampling (finding some lowest common denominator of cross-platform functionality) wouldn't be required in order to deploy more complicated applications to two or more of the target platforms.

Another reason Temkin thinks OpenLaszlo merits a look by RIA developers: not only does OpenLaszlo take the risk and cost out of writing specifically to one platform or another, he says the language is also far better suited to the task of writing RIAs than the native languages of the target platforms in the first place. For example, he says, DHMTL never had today's sorts of RIAs in mind. Nor did Flash or Java.

Initially, OpenLaszlo is supporting DHTML and Flash with Java and Silverlight support coming next. The entire platform is open source which is good. But Laszlo Systems has selected what is probably the most insulting open source license of the ones that the Open Source Initiative (the OSI) has blessed: the IBM-authored Common Public License (which even IBM doesn't select for its projects any more). The problem with the CPL is that it gives the licensor (in this case, Laszlo) a license to steal intellectual property from licensees. Basically, under the CPL, if the licensor infringes on a licensee's patent that's completely unrelated to the IP being licensed under the CPL, the licensor gets to revoke the licensee's license to that IP. For example, if I give you a CPL license to my search algorithm and then you sue me for infringement on your hard drive patent (completely unrelated to the search license), I can revoke your license to my search algorithm.

At first you say, what's the big deal. But, if you have an entire product line that depends on your license to my search algorithm and I know that, the CPL paves the way for me to infringe on your other IP without fear of reprisals (since there's no way you can kill your whole product line). Most open source licenses have such terms of mutually assured destruction (MAD) -- but it only relates to the IP being licensed and not unrelated intellectual property. It's basically a license to steal and I personally would never accept anything under the CPL and I shared this opinion with Temkin.

I'm not for a minute implying that Laszlo would do such a thing and I personally think the technology is very cool. But with all the other OSI-approved open source licenses that are ripe for the picking, there is literally no reason to pick the CPL over the others other than its "license to steal" character. There are other open source licenses that accomplish the same thing as the CPL without the out-of-bounds provisos found in the CPL (for example, the Eclipse Public License which is what the Eclipse Foundation moved to, from the CPL, knowing full well that the CPL was a bad license).

Also in the video, Temkin shows us the Laszlo Web Top -- a self contained desktop RIA that runs on Flash and into which any Laszlo developed RIA can easily be plugged. The Laszlo Webtop reminds me very much of what Sapotek has done with Desktoptwo by offering what is essentially the functionality of a desktop PC inside of a browser. In other words, there's a notion of a desktop (in the browser or Flash runtime) with applications that run on it just the same way you have a desktop with applications in one of your more common desktop operating systems (Windows, Mac, Linux). Today, the only applications that ship with the Laszlo Webtop are Mail and Contacts and, much the same way Sapotek is hoping its open source community starts to build applications for Desktoptwo, Laszlo is hoping that an open source community will rally behind the Laszlo Webtop with a library of useful applications.

One big difference between the approaches being taken with Laszlo and Sapotek is that the Laszlo Webtop is a Web-based desktop that organizations would deploy to their users on their own (as opposed to the hosted solution that Sapotek runs). The example he shows in the video is of the Webtop that Verizon is deploying.