> > In general i think that we really need task forces in plural.
> 
> Agreed, but some we already have some (like kernel hackers, X.org guys,
> people who build desktop environments, ...).

I know that a few project have very organized task forces. However neither of those you mention shows much of that.

> > There are a lot of overlapping projects. One example is that KDE and Gnome should merge, at least when it comes to key technologies.
> 
> That would certainly help to reduce the desperate shortage of manpower.


> > 1. We need a standardized method for embedding objects. Like Java Beans and OLE/ActiveX in Windows. I mean how do you view a office document in a VS application, or create a PDF report? You find an ActiveX that can do it. KDE has KParts and Gnome has something similar, but we really need a standard for this. Or in other words we need a standard for Custom Controls. In OLE/ActiveX an object can be inproc (dynamically or statically linked) or out of proc (external server process that starts on demand). Also the app is totally oblivious to w 
> > 
> > 2. We need a visual editor that can create new Custom Controls, or design containers like forms/windows.
> > 
> > The Lazarus FreePascal IDE is an interesting project.
> 
> So instead of learning how to properly use existing libraries and then
> writing nice and clean code yourself you prefer to have some kind of
> gizmo spit out some crap for you and you actually call that programming?
> Thank god that most people writing free software don't take the same
> approach, because I'm convinced that's just a way of making things
> rapidly deteriorate. Don't get me wrong - I know these are very powerful
> tools in the hands of a true professional, but they're also helping to
> make things suck so much more when some lazy and incompetent fuck is
> just abusing them to roll his piece of crap ASAP so that he could start
> making easy money out of it and doesn't even have a clue of how to use
> them properly. Unfortunately I've seen way too much crap that was
> "programmed" exactly this way to prove my point.

What? Can you be more specific in how you come to this ramblings out of what i wrote? How would having a standard rather then (at least) two incompatible solutions create such a mess?

> > Another thing that we need is a good native database library. We have
> JDBC and ADO.NET. Perhaps one of them could be ported to C/C++? We also
> have PHP Data Objects that is quite nice and probably implemented in C,
> perhaps the C code could be used to make a "C Data Objects"?
> 
> There are quite a few - libmysql, libpq, libsqlite... many of them with
> bindings to whatever language you like. Just pick the best fit for your
> desired application.
So how do i use libmysql to connect to my postres database again? None of those are database libraries. They are database-specific client libraries. 
The only C database lib i know that is actively developed i libzdb (http://www.tildeslash.com/libzdb/documentation.html).

> I think the bottom line is that the programming model is completely
> different and those who just want to make their job easy and cobble
> something up without actually knowing anything about coding are gonna
> have a very hard time getting used to it.
In what way is it different? Most things i suggest are already in the programming model. The OLE/ActiveX-component type of component model for OO programming does exist. It exists in the form of beans, kpart and bonobo. The only difference with creating a standard is that kde objects would be usable in gnome and gnome objects in kde. 
 		 	   		  
_________________________________________________________________
Hotmail i mobilen på 5 sekunder!
http://new.windowslivemobile.msn.com/SE-SE/windows-live-hotmail/default.aspx