Earlier this week, Curl put out a press release that talked about their pending relaunch into the North American market at Web 2.0 Expo. Curl was one of the very early Rich Internet Application players, but they were ahead of their time so didn't see massive adoption here in the US. In Japan, however, Curl did quite well and has been providing applications to some big clients including Toyota and the bank of Tokyo. The "relaunch" means that Curl will once again bring marketing and sales resources to the North American market. It also coincides with the 5.0 release of the product.
Curl was founded in 1998, made a bit of a splash here in North America, but really took off in Japan, gained the attention of Sumisho Computer Systems Corp, which purchased the company. Since then, they've focused more and more on the Japanese market while the North American market matured. They think now is the time, and with a lot of buzz and excitement around Rich Internet Applications, they may be right.
There are three parts to the Curl environment. The first is the Curl language, which looks a lot like any other scripting languages, and according t the press release, offers " full-featured, object-oriented programming language that integrates rich text formatting, GUI layout and presentation scripting". The second part of the ecosystem is the Curl Run Time Environment. This is the browser plugin that is required for Curl to render. Currently, it works for Windows and has a beta version for Macs (on Vista it took forever for me to load, but your mileage may vary). The download is about 7 megs, so it's large by runtime standards, but with broadband it isn't outrageous. Finally, to build Curl applications, you use the Curl IDE, which is pretty standard, and includes a debugger, performance monitor and an HTTP traffic monitor. Building Curl applications is a lot like building HTML. You made changes to the code, then browse to the .curl page, and the browser uses the plugin to render the application. You can also view source in your browser and the code displays just like you were viewing the source to a web page.
Some of the examples are cool and give a good idea of what Curl is capable of. It's almost like half flash, especially some of the graphic manipulation, but it doesn't quite have the whole picture. The applications themselves don't resemble Flash at all, and look a bit outdated to me, though the Curl team assured me that they could be customized and polished by developers.
Curl is positioning itself for the enterprise market, and the pricing is representative of that. Curl is free for noncommercial use, but jumps up to $12,000 to start for commercial deployment. Is it worth it? Well Curl seems technologically superior to some other technologies. It can handle large datasets and offers offline capability out of the box. I'm still not sure it can compete with solutions like Nexaweb or Adobe's Live Cycle Data Services product, but it has a solid list of customers that implies it's proven itself as an enterprise product. The fact that it requires a brand new language and runtime may hinder adoption, but if the technical benefits are really what they claim, enterprise customers would be interested. This is still a growing part of the Rich Internet Application world, so Curl may be able to gain some solid traction.
For more thoughts on Curl, see People over Process.