[Mono-osx] Requesting tool attention for OS X
Miguel de Icaza
miguel at novell.com
Sat Nov 28 16:35:57 EST 2009
>> Another option is for someone to complete the Monodoc/Cocoa#
>> version so that
>> we can ship that by default (it has to be Cocoa# because we do ship
>> already with Mono) or it could be written against one of the other
>> for Cocoa, but it would have to be shipped as a separate tool.
> Is Cocoa# the de facto / recommended framework on Mac? There are a few
> out there and I've lost track as to which is "the one to use".
Well, we do ship Cocoa# but it has not been maintained for a long
time. I am merely saying that if we want this to be part of the core
Mono, it has to be done using Cocoa# as we do not currently plan on
bundling any of the other libraries until one of them emerges as the
clear superior framework.
That being said, I did look at the other frameworks a few months ago,
and they had various degrees of completeness, design goals and
licenses. I would not say that there is any framework (including
Cocoa#) that I really like.
For us to bundle a new Mac framework it would have to satisfy a couple
* MIT X11, Apache 2 or MS-PL licensed class libraries.
* Commitment to API backwards compatibility.
* It should probably be closer to what we did with MonoTouch's API
design than to Objective-C's native APIs.
Mono has accumulated third party libraries over the years that did not
meet some of these criteria and we ended up with the burden of
maintenance, so I would not like to repeat that. In particular we
have ended up in a situation where we either deprecate libraries, try
to merge two APIs and expose them both and do multiple builds or ship
two versions of the same library. I would not like to repeat this
process, not for the Mac, and not for anything else.
In the short term, Cocoa# is what we have, and although not perfect,
it is bundled and would be fine to build a tool like native MonoDoc.
More information about the Mono-osx