[Cocoa-sharp] The State of the Art

[Paolo Molaro]

On 02/23/07 Andrew Satori wrote:
> Let me be clear, I think the project could be great, my issues are  
> not with Geoff and the CocoaSharp guys, they are not with the Mono  
> project in general. They are very specific to a core developer who  
> delights in being an ass to people that aren't as bright or talented  
> as he is, and those who don't live and die OSS.  I write code for a  
> living, and don't have the time to play his games. 'File a Bug',  
> 'Submit a Patch', 'Quit Whining'.

Do you think that insulting the guy who actually did port mono to OSX
helps? Your point of view is not based on facts: try to be an adult
and don't let your frustration drive your behaviour.
You complained on irc that you had to patch a config file for two years
on your box and still you didn't have time to file a bug report.
How am I supposed to know about a bug if you don't tell us about it?

And let me tell you: you keeping a fix to yourself is what damages the
mono osx community, because other people may have hit the same bug and
you didn't share the fix. And by not telling us about the bug you also
prevented other osx users from enjoying the fix (the bug was actually
not osx-specific, so your complain that the bug was an anti-osx plot
from outer space is also completely off-base).

> Mono has a debugger, it doesn't work on the Mac and under the current  
> leadership, and without significant contribution by a Mac developer  
> with the know-how and financial ability to sit down and write a Mac  
> friendly debugger, no current Windows C# dev's are going to convert.

The debugger is in its early stages, how do you pretend that it gets
ported to multiple different architectures and operating systems at this
point? Especially if you're not even willing to contribute a bug report?

> Mono has an IDE, it doesn't work well on the Mac because of missing  
> features and the requirement of having most of a working Gnome  
> desktop under X11.  Xcode works fine as a code editor, and with  
> CocoaSharp it's fine (up to a point) but again, without the work of  
> some dedicated dev's and a little help / support from Apple, you'll  
> never bring Xcode up to a reasonable level. (all of the plumbing  
> exists, but it's poorly documented if it's documented at all).

Some brave people started working on an xcode plugin, other people
are porting Gtk to native OSX, which will make monodevelop much easier
to run on OSX. Have you contributed to either of those projects?
Mono, gtk, cocoa-sharp, monodevelop, the xcode plugin are open source
projects: they _require_ people to contribute even with simple bug
reports and documentation if they don't have time or inclination for
writing the code. If you want everything done and on a silver plate,
then yes, none of these projects are likely for you.

> ASP.NET is unreliable in anything but single user cases under OS X  
> because it's never been really tested.

Did you file bug reports for the issues you found? I can't fix a bug I'm
not aware of. how can you pretend we fix issue we don't know about?

> Simple things in the Mono for Mac code continue to get fixed and  
> broken because few of the core developers use a Mac as a primary  
> platform.

Regressions sadly happen all the time on all the architectures: the only
difference here is that you don't want to report a bug.

> All of that said, the problems could be fixed, but my conclusion was  
> that the Mac community is not large enough to warrant the effort from  

The main issue is that the community needs to be educated to report the
issues and learn to contribute as all the other communities around mono
do. Mono is not perfect, but it can get better only if each one of you
does his little or great contribution. I know that people have day jobs,
but there are several ways to contribute, one of which is to file bug
reports with reproducible test cases, for example: this is how open
source projects work. (And no, insulting people doing the osx port
doesn't count as a contribution).

> ODBC is *STILL* broken because the CONFIG file doesn't in fact load  
> like it's *supposed* to, and fixing it so that it works the way it  
> works, not the way it's *supposed* to work makes you an idiot in the  
> eyes of a certain core developer, who then blames the whole damned  
> thing upon you because your were patching it to work rather than  
> filing a bug, but I digress)

The story is like this: while I wasn't paying attention to irc
you complained about the bug and someone suggested you to fix it
in an incorrect way and you kindly offered to submit a patch.
When I read about it I told you to not change the config file, because
it would have been an incorrect fix: the bug was elsewhere. You got
upset over it. I'm not reponsible for other people giving you
incorrect info in the way to fix a bug. But you took it personally, I
guess because you had spent some of your valuable time writing the patch.
As mono developers we also must ensure that changes committed in svn
don't damage other users: it's not enough for us to workaround an issue,
otherwise we'll never get a working mono.

As a matter of fact, the bug has been fixed in svn less than two hours
after it was reported (and no, irc is not a good place to report bugs,
because people that are responsible for that area of code may not be
online, or reading or they may be busy with other work: we don't ask
people to submit bugs to bugzilla because we're sadists, but because
that is the best way to help us fix the bug: it gets on our bug list, we
can contact the reporter for more information or to confirm a fix, the
reporter can include a test case etc.).

lupus

-- 
-----------------------------------------------------------------
lupus at debian.org                                     debian/rules
lupus at ximian.com                             Monkeys do it better