On Mar 6, 2006, at 4:28 PM, Eric Jonas wrote:
First, let me say that I use inkscape for at least two hours every day. Often more. I just finished my first hour-long presentation using slides made in Inkscape. I think inkscape is, beyond the linux kernel and python, the most awesome free software project I've ever had the pleasure of using.
But every time I try and get into the codebase, I'm always put off by a few things:
- is it a C++ app or a C app? Is seems like the gtkmmification is
always what people "should be" working on, but it never really moves forward. I try and understand the code, and there are function pointers and all sorts of other stuff hidden under there. Lots of the codebase still has functions with names left over from the SP fork.
So every time I'm like "I think I'll try fixing blah" I get frustrated because 1. I can't understand the code base 2. I worry that it'll all be for nothing, because once we switch to gtkmm/cairo/tiny-robot-api it'll all be undone.
[SNIP]
From my past experience I've seen this slow-but-steady approach pay better in the long run. One thing we have to do is get things defined, tested, switched, and still working.
Most of the newer work is C++. Plus as we go, parts get changed over bit-by-bit. A lot has already been fixed under the scenes, and more is actually starting to pick up momentum.
There's been a lot of work done in regards to hashing out design, and the Jabber room is actually a fairly good way for developers to get this done. However, one of the biggest payoffs we might get is in isolating and dropping code. So instead of changing everything, we can shift interfaces and module bindings to better C++ and then kill the C stuff module-by-module, with much of it ending up getting deleted instead of rewritten. Thus if we rewrite too much up front, we'll just end up wasting effort.
Now, others have different feelings about all sorts of things, but my personal take on things is that spending the money on travel won't pay the largest benefits.