Technically Arwinss may not be the best possible architecture, but IMHO right now is the only viable one in order to reach beta in reasonable time.

Sure, we will always see how a more “native” implementation could be more efficient at the end, but the reality is that given the current human resources it is not a realistic approach, it simply won’t happen in many years with the actual resources.

Arwinss allows us to use most of a working win32k subsystem (wine’s) with minimal effort, thus saving huge amount of work. So we can focus in implementing other very needed areas to have a complete os.

Why to invest an huge resources that we don’t even have to implement something what is already done, better or worse? After we have the needed partitions, filesystems, complete kernel compatibility, etc, if we have more resources, we may consider t keep the the native win32k ss development, with the advantage of having a complete and working system to compare and test against, and most probably with more resources after we deliver an usable system.

ReactOS goals are to achieve maximum windows compatibility at both application and driver/kernel components. We don’t have any part finished. Arwinss solves with minimal effort the application APIs side that would require he largest effort otherwise, so we can dedicate our very limited resources to finish the other parts.

It is just my opinion, but I see it so clear… I hope you all understand this: the best architecture can be the worst one if there is no a realistic plan to develop it.

Jose Catena

DIGIWAVES S.L.