Last week I talked about a collaboration I had with someone. This week I would like to talk more about the idea I had. The idea will involve collecting data, scoring it. But the target audience is in-house development teams and COTS applications. The internal development team version will help educate your team. The COTS version will help you not make a costly purchasing error. As part of the RFP process. The lack of quality I see in COTS applications is the driving force on this entire design. I’ve seen vendors with horrible code, less desirable designs, or outdated patterns that I have to work with because we purchased the lowest qualified bidder. And the weak products should be removed. The pricing model will be subscription. As a consumer, I hate this model! As a developer, I love it. The old model of buying it and pay maintenance is a better balance. The subscription model will generate renewable money. That is hard to argue. For talking purposes, I am guessing at 1-2k range. The scoring model will five categories and will use a weighted average rule. We’re the weights will be configurable. But generate one overall score, like a grade or credit score. Each test will score a number between 0 and 1, yes, a ratio or percentage for the results. I am concerned about the pass/fail type that could skew my results. For example, a true result score 1 (100%) and false at 0 (0%). Too many of these could pose a problem. But the goal is to have the score be a number for math and attempt to follow school grading with 70% and up being a pass. Below 70% fail. The second thought is to use - 0-20 - critical, - 21-40 - danger, - 41-60 - warning, - 61-80 - alert, - 81-99 - information The results will be a category summary (5 groups). The details will be a table to tell you why and what. My first step was to create the storage and defines what each column would hold. When possible, I also represented the actual values I would use for them. Now I started to brainstorm the items I have already created and what category they would be. Easy work was done. I added the things I would like to add in, irreverent of effort. Parameters were next. I have been working on a standard for this, so the first few came from that. - Help - Version - Debug - Expire check - Output (grade, category, detail) - Category (1-5) - Type (COTS, RFP, Developed) - style (Universal standard, custom) - User standard with path - ignore feature list - Module - Joe check these I am sure the parameters will change. It is early in the process. I know I will need to create a table for us as a reference for a standard. The table can house a universal standard or a homegrown standard. The homegrown one will read from a configuration file to override the universal standard. The target for completion is the August / September timeframe. I hope I don’t do like some Kickstarter programs and deliver a year late. The issue is I have a day job, swim coaching, and am writing three books (one generic business book, two swim coaching books), all in the early stages of completion. I am completing the base design. Tomorrow, Sunday, I will start to create the test, one by one. I am sure the first one will be the hardest, and it will get easier after that.
