Sunday, May 8, 2016

the real survivor

The three legged stool of business economics are something like [1] know your product [2] know your customer [3] know where the money is coming from. But there is also a silent fourth rule. Know your exit strategy or how to scale. If you know how to scale the business then you might have the next good and if not then just cash out and let someone else scale it. Most success stories result in one or two major successes and then a multitude of small ones. l refer to Shark Tank and The Profit.
Hidden in there are a few technical details that are typically missed by founders and technical leaders.
Certainly if you are building a compression algorithm, an electronic trading system, autonomous vehicle or ABS braking system, fly by wire; then writing your project in machine code, assembler or some variation of C or Ada then that makes the most sense. Compute cycles matter as long as you can count them.

But if you are building a business application like an accounting system, reporting, CRM, CMS, HRM then the intellectual property is in the workflow and not the algorithm. There is ABSOLUTELY no advantage to writing the business logic of the application in a native language like Go, C, C++, C#, Java, perl, python, ruby or any heavily compiled or interpreted language. If there are more than 100 to 150 APIs then its the wrong tool chain. By example I look at tcl, th1 and Lua. These languages are lightweight and require simple interpreters. In fact there is likely to be a basic interpreter for your favorite language which you can expand on.

Once the core system has been implemented then it can be bound to a GUI on the front end and "macros" on the back end. Both should be highly templated so that all systems can be manipulated easily. Furthermore migrating from domain to domain is just a matter implementing the basic services instead of the entire system. (at one in the process you will stave off the competition by implementing features so that the cost of entry is higher for the new competition. However, the velocity curve will always converge and so you'll need millions of lines of code to establish and hold a leadership position but then there is also LOC explosion that takes place when using templates and macros that the competition might not see.

No comments:

Post a Comment

another bad day for open source

One of the hallmarks of a good open source project is just how complicated it is to install, configure and maintain. Happily gitlab and the ...