Re: Eaten by the Tiger? Open Source WO is coming...

Am 25.08.2005 um 00:21 schrieb Alex Raftis:

>
> Anyways, I thought people might be interested in a little  
> background on the project.

Here is my history to ink out the picture a little more:

Last winter (2004) my company (http://www.codeon.de) was handed the  
task to provide within a years timeframe an EOF replacement for a few  
desktop and daemon applications (no Java). Our customer expects at  
least the same level of quality he is used to. He really needs the  
replacement at the end of 2005. The applications use EOF fairly  
feature complete, a simple one-model/one-database-context/one-editing- 
context implementation will not do.

In the beginning I looked at GDL2, Ogo GDL1. BDControl, and  
AJRDatabase. (http://www.mulle-kybernetik.com/weblog/archives/ 
000611.html). I started extending GDL2 and then AJR to get a feeling,  
for what is there and what wasn't there and the quality of the code  
base. I judged how much effort it would be to extend or rewrite the  
code base vs. rewriting it from scratch.

The decision was fairly multi-facetted and not just based on a  
comparison of estimated useful existing code lines vs. expected  
number of needed code lines. There were also cooperation, testing and  
migration issues considered. And last but not least there were also  
business issues to consider.

I decided to do at least EOAccess from scratch and I was thinking of  
possibly branching off or contributing to AJRDatabase for EOControl  
(when the conversion off the DBPrefix would have been done).

I was not alone in that judgement. Another company, that shares our  
customers problems, reached a similiar conclusion independently. And  
I knew of yet another company in need of an EOF replacement. So then  
in the spring of 2005 three parties tried to set up an EOF project as  
an equitable joint venture.

Unfortunately two would-be partners couldn't get their funding  
secured for 2005, so I had to go ahead on my own anyway because of my  
timeline pressure. That's where I am today, still in the process of  
writing EOF with good prospects of making my deadline.

As it turned out, that it was beneficial to have some portions of  
EOControl done for testing EOAccess, I code and accumulated many  
parts of it over the months. When the time came to do the rest of  
EOControl recently, there remained basically just faulting and object  
store classes. That turned our not to be a big enough task, that it  
would have been an incentive for me to reuse AJR code.

So what does it mean for people on this list ? Currently it shouldn't  
mean much. The code is developed in house and it will be used in a  
closed source project. It is an option but there are no plans to  
release this EOF open source. I will be finished first and then we  
think about the next step.

> Finally, for those interested, I've pretty much decided not to  
> clone EOInterface at this point. I just don't really see the need,  
> since Apple's added most of the EOInterface layer to AppKit and  
> Foundation. That being said, I have written a very small and  
> primitive EODisplayGroup as a subclass of NSArrayController, just  
> to see if our abstraction layer could be tied into modern AppKit,  
> and it works fine, even if it's incomplete.

I'd like to mention that an EOInterface clone already exists, http:// 
www.mulle-kybernetik.com/software/MulleEOInterface/. I am confident  
enough to claim, that it doesn't suck :)

Ciao
     Nat!
------------------------------------------------------
I like the word Irksome, it sounds like it should mean
more than it actually does. -- Chris Redgrave