In my last post, I just mentioned some thoughts about where I would be going with the post-V1. I not making major development changes at this point, so as not to delay the product any longer. Even before adding WPF and developing with Orcas, though, there are quite a few items such as VB.NET and partial or complete C# 3.0 support as well as other postponed items such as specifications that are relatively easy to add.
This week I am going through the user interface and getting the final appearance down and fixing all the little UI and keyboard issues. I moved up the Scan window from a docked pane to a ribbon tab per Scott Hanselman’s suggestion. It’s almost July… I don’t want to still be worrying about UI after this month; I am growing tired and want the program out.
I won’t be posting much this week. I’ll post more NStatic examples next week, revealing a few more aspects of the tool for the first time. These posts will make it easier for me to pull together documentation before I ship.
After I ship, there are some proposals I will be entertaining from a few companies. So, I may be selling NStatic independently for a few months before any deal is finalized, after which the price will likely go up dramatically; bargain hunters might want to buy early. For now, I am continuing on my original plans.
Part of my original viral strategy was to get people used to NStatic by making a minimal version of the software, which could also act as a Ribbon-based notepad clone with enhanced editing capabilities—maybe pull people into purchasing the full version. This goes with other forms of traditional marketing—PR, advertising, etc. I am not currently sure now how I will have people evaluate my software.
I am hoping that maybe, just maybe, that, with a unique sophisticated static analysis product that’s a must buy, I might be able to capture some percentage of the .NET developer marketplace. Other categories such as unit-testing, component libraries, source control, text editors, obfuscation, refactoring tools, bug-tracking are heavily fragmented with many competitors or face freely available alternatives. I am effectively all alone with an advanced technology and no overhead, too, so I can underprice any large competitor. This ties in with the central message I got from my business school’s strategy course, which is to “avoid competition.” (For example, Walmart eventually overtook Kmart because most Walmart stores were local monopolies based in rural areas.)
How big is this market, though? I used to have data on market sizes of different software application categories, but since I left business school, I lost access to all sorts of privileged market data. I have to rely on serendipitous information, such as slips by Microsoft executives about there being over 1 million Visual Studio 2005 shipments and 15 million VS Express downloads.
Market information is hard to extract from the web, but I did find a bevy of information from O’Reilly’s Radar weblog on the State of the Computer Book Market (Q1 07). The numbers come from a database containing information about Bookscan’s weekly top 10,000 titles sold. Approximately, 200K books on C# and an additional 120K books on .NET are sold each year. Visual Basic books sell at half the rate of C# books, so I could get a 50% boost in sales by including VB support.
The data also suggests that porting to Java and C++ would each effectively double my C# sales, which are not currently in my plans. I don’t want to damage the business model of companies with top-notch researchers who dedicated their lives to source code analysis. I don’t have a tight attachment to source code analysis; I am just applying my symbolic AI techniques to different application areas.