This is one of those posts that I like writing once in a while, to analyze the status of the project, draw lines, set directions. It helps to build an overall picture of the matter, filter what is relevant, vent a bit.
Tagline: cut rough edges, eradicate weed. Or, “bulldoze the place flat“.
Codebase Complexity + Bugs = Kaboom
ScrewTurn Wiki accounts for more than 84 KLOC (53 KLOC code, 31 KLOC comments). It’s nothing particularly big, but over time the codebase has become particularly intricate and has grown without control. This is our fault obviously, but that’s not the point. Even having 1,811 unit and integration tests, there are still quite a few bugs, especially in all those parts that relate to the web front end. Fixing those bugs has become more and more complex and time-consuming, and many of you have surely noticed how we are not actually fixing any non-trivial bug lately.
The problem is that users are hit by the same bugs and rough edges over and over again, and they demand fixes.
On The Visual Editor
Let’s face it: the Visual Editor sucks. It probably did a lot to bring more users to STW, but it’s extremely buggy and moody. One day it works, the next it doesn’t. Plus, when it’s raining or if the temperature is higher than 26Â°C, it corrupts content.
Well, the idea is to remove it. Much better to have fewer happy users than a lot of disgruntled ones. It turns out we’ll keep it after all, and improve it.
Tagline: get less users, look for more quality.
Size Does Matter, But What About Quality?
We’re now at almost 10,000 downloads per month. If only 1% of those downloads converts to a real wiki instance, that’s 100 new wiki instances per month. We have about 100 new forum messages per month, which is a lot considering that all of them are help requests. The fact that we are not fixing bugs, let alone answer forum questions, fast enough is causing a lot of users to grow more and more unhappy with ScrewTurn Wiki.
Funnily enough, we’re de-facto monopolists in the free .NET wiki engine market. Or so it seems, feel free to correct me if I’m wrong.
Tagline: “open-source runs on time, not on money“.
We’re working day and night to make our other project successful. That’s hard enough for two developers. If you add STW to the equation, well…
You’ll surely know that we have commercial licenses for those companies that cannot or don’t want to use GPLv2-licensed software, but they do not generate enough revenue to hire someone to work on STW for a significant amount of time. After all, we’re not WikiMedia Foundation.
Pre-emptive comment: of course we could offer commercial support, but that’s the most boring thing to do for a software developer. I’d like to avoid that like the plague.
We now want to experiment what we can do with a couple of interns. So, if you are a university student with some .NET experience and want to work on a somewhat successful open-source project, drop us a line.
I’ll spare this for another post, as I have no idea it will take some time to figure out.