[Mono-devel-list] Mano - Mono on PalmOS

Wolf Logan wolf at circle-cross.org
Thu May 1 17:52:08 EDT 2003

The problem is that Mono is, like .NET itself, an environment -- 
essentially, it's a set of libraries that are invoked by a machine process
to handle bytecode interpretation and resolve method calls to the runtime
classes. this assumes a certain OS-level programming model, namely that of a
kernel with running processes. the PalmOS doesn't work quite that way, and I
believe that attempting to simply port Mono to the Palm would be extremely
frustrating. WinCE (and embedded linux) *do* in fact work that way, and I
can easily imagine how a port to one of those systems would fit into the
mainstream of Mono development.

what I was hoping to do was write a PalmOS5 app *using* Mono (as ported to
the ARM) as a basis. I will still probably do so. but I'm at a loss as to
how any of the code I'd end up writing will be of any use to the mainstream
Mono project, since it will mostly be concerned with Palm-isms. For example,
ARM code does not run "bare" on the Palm -- the OS exists in a compatibility
environment that requires ARM routines to be wrapped in stubs for access
from a pseudo-68K world. calls to the OS go through a similar wrapping
process -- although the OS itself is ARM code, and it's being called from
ARM code, the call goes through a 68K compatibility layer.

but never let it be said that I'm stubborn. if "the Mono community" at large
thinks that it's a bad idea to have a project to port Mono to the Palm, so
be it. I'll be happy to close down the project at sourceforge and retract my

----- Original Message ----- 
From: "Sergey Chaban" <serge at wildwestsoftware.com>
Sent: Thursday, May 01, 2003 1:33 PM

> So if I understand correctly you're not proposing to do a fork but more
> writting a standalone lib that would implement missing APIs on top of
> For example, implementing mono/io-layer (essentially a Win32 subset impl.)
> PalmOS interfaces? So with such a lib in place it would be possible to
> Mono for PalmOS5 without changing a line in the Mono itself.
> Sounds good in theory, but may be extremely difficult in practice.
> Besides it's simply not enough to do a port.
> I believe many people on the list will be interested.
> This would benefit development for other small systems like WinCE and
> Linux (and in the future maybe QNX and Symbian).

More information about the Mono-devel-list mailing list