Hi,
I think a feature is just one issue, though an important one, among the many factors that characterize a product. What attracts a graphic artist to a program, in a large extent, is its user friendliness. I started working with Illustrator about a decade ago, and throughout the years, I became pretty comfortable with the program. I think Illustrator is a great product with well-designed tools. It is user-friendly, intuitive, feature-rich, and flexible. You can use it as a web-graphic tool and it can create great artworks for offset printing as well. Then years ago, I tried out Freehand and I immediately fell in love with it. It has an extremely intuitive interface and a unique personality that I found attractive right away. Ever since, I have been using it on all kinds of projects; even if it is slightly less features-complete than Illustrator, I prefer it to other products. This is just to say that features are just one factor in the overall picture.
When designing an application in a proprietary environment, developers always start with the development of the overall layout of the program. When that worked out individual features developed and added, in accordance with the general blueprint and the look and feel of the program; they understand that a tightly designed and compact layout is a big thing when it comes to winning over users; it also essential for efficiency and productivity. For graphic artists user-friendliness and intuitiveness is very high on the priority list.
A badly implemented and confusing interface puts users off. Gimp is the perfect example of this. Here we have a feature-rich app that cannot take off because of its weird look; its scattered, confusing, floating palettes turns graphic artists off completely. It is around for about a decade (as far as I know) but so far, it couldn't make noticeable inroads among professional designers. Unless it changes, it remains forever the graphic tool for amateurs and hobbyists.
The counter example is Scribus that already has started developing a personality and small businesses have already started adopting it. It has a clean look, a limited but well-implemented tool sets. In terms of features, it is lacking many essentials, like Undo or Bleed, for instance; yet more and more people find it as an attractive alternative to QuarkXpess.
This is the strategy Inkscape developers should consider adopting–first, develop a personality, and leave the bickering about individual features for a later time. For if the overall concept and design of the program is well thought out details will fall into their own place more "naturally".
Jozsef Mak