I thinks there's space for both.

For some markets, a native look and feel is a must. In my opinion, this happens when:
- the user may easily find a replacement for the software -- like a shareware text editor
- the software does a secondary task for the user -- he/she does not have time to learn
- the software is used for short periods of time -- he/she fires up it, use it and close it

For other markets, the users simply don't care about the look and feel. This happens when:
- the software is mandatory -- like a company-wide accounting/management software
- the user needs/wants to get the job done -- he/she will probably gets training on the software
- the software is used all day long -- get customer's data, prepare invoices, check payments, etc

Websites are in between: sometimes the browser use native widgets, sometimes it does not.

I agree. That is always what I hoped FPGUI would be, and even a limited LCL-fpgui would be useful here.

I currently have an acquaintance trying to customize QT/X to get it working on top of a own build embedded solution. Good luck.

But that is something else than refocussing the entire Lazarus in that direction, like Graeme was always pleading. Lazarus is IMHO primarily about delphi compatibility, multiplatform (and that is always a precarious balance) and native widgets.

When looking for multiplatform options, a few years ago, I tried FLTK. It's a C/C++ GUI toolkit with custom drawn widgets. They have the standard widgets plus a few nice controls like knobs and rolls and a very basic theming. However, the lack of printing/reporting and other higher-level features made me write it off.

2010/3/9 Marco van de Voort