RPGResearch/bcirpggithubmirror
RPGResearch/bcirpggithubmirror
7
0
Fork 0
mirror of https://github.com/RPG-Research/bcirpg.git synced 2024-04-16 14:23:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

Updated UML Guidelines (markdown)

RPG Research 2023-01-15 11:36:42 -08:00
parent 189dec9d29
commit f5dd9afee3
1 changed files with 7 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
UML-Guidelines.md

@ -2,13 +2,17 @@ UML Guidelines
General Flowcharts and Unified Modeling Language (UML) is used as part of the design phase of all our projects.
While it has become (arguably erroneously) popular since about 2010-2015+ to regress back to code-first, design-later, this invariably leads to unscalable and unsustainable projects. The code-first approach works okay for small, non-distributed projects, the "quick and dirty" code. But for more complex projects, the code-first approach invariably leads to chaos the moment you try to distribute the work across other developers. The code-first approach gives the illusion of more progress sooner, and does in the short term, but in the long run begins to bog down horribly, leadin to (at best) only being able to release code for 10% of the time toward new features and improvements code (forward progression of the code base features and scaling), 85% debugging/fixing (trying to maintain/fix the existing code base), <5% documentation. This all leads to an overall garbage codebase that can't be scaled, or maintained, especially when constantly changing developer eyeballs/brains.
While it has become (arguably erroneously) popular since about 2010-2015+ to regress back to code-first, design-later, this invariably leads to unscalable and unsustainable projects. The code-first approach works okay for small, non-distributed projects, the "quick and dirty" code. But for more complex projects, the code-first approach invariably leads to chaos the moment you try to distribute the work across other developers. The code-first approach gives the illusion of more progress sooner, and does in the short term, but in the long run begins to bog down horribly, leadin to (at best) only being able to release code for 10% of the time toward new features and improvements code (forward progression of the code base features and scaling), 85% debugging/fixing (trying to maintain/fix the existing code base), <5% documentation. This all leads to an overall garbage codebase that can't be scaled, or maintained, especially when constantly changing developer eyeballs/brains. Typically unable to move beyond 3.x iterations without major overhauls.
The design-first approach, if correctly implemented with Object Oriented principles, typically leads to much more sustainable and distributable codebase over time. This usually means it works out to time spent around: 10-20% on documentation, 20-50% toward new features coding iteration (forward progression of the code base), 20-40% debugging/iteration (trying to maintain/fix the existing code base), and is SO much easier to collaborate, maintain, and support over many years and many different eyeballs and brains.
We follow the principles of Object Oriented Design (OOD), even when not using OO languages, and design-first approach means it takes longer before we actually write working code (in the early phases), but leads to a much more sustainable, scalable, distributable, flexible, and manageable product overall.
We follow the principles of Object Oriented Design (OOD), even when not using OO languages. The design-first approach means does mean that it takes a little longer (at first) before we are actually write working code (in the early phases), but overall it leads to an exponentially more sustainable, scalable, distributable, flexible, and manageable product overall, able to iteration past 10.x revisions and beyond, without complete overhaul, especially if Design Patterns and Refactoring becomes part of the regular process.
We also use UML as a visual aide to help with the design documentation.
We also use UML as a visual aide to help with the design documentation. UML is flexible in specific implementation, with many variations of the visual representations in degree and detail, and that is a boon of UML.
This document attempts to establish within this development community the visual design standards we want to follow with our diagrams, so we're all on the same page when looking at the code.
Once a UML class diagram and the other documentation are in place, it is easy to have someone volunteer to take on creating the code for that class/feature in a much more consistent way. And if they don't finish, or get stuck, make it much easier for others to come in and help.
If you wish to contribute coding to this project, then we ask you to please make yourself at least somewhat familiar with these OO and UML principles.