Originally Posted by Wolf
*chuckles* You'd think with my signature that nobody would miss it.
They're rigid, not forward-thinking, kludgey, and inefficient. For some reason they can't do anything quite right, either - they do things that work and what they've been taught by someone else, but if it changes, they can't recover and it becomes extremely fragile (you should see what it takes to create a "development environment" or install some software I have to deal with that was obviously designed by and is maintained by a bunch of rigid "it just works this way" STJs, who are so inflexible that it drives me (and my department full of NTs) mad). Wonderful for regurgitation and re-ordering of existing theories and structures, but horrible at solving a problem they've never seen before, considering for the effects of their decisions, or even comprehending why something is as it is (or being bothered by it enough to figure out how to do it right).
It's funny because we NTs make more small mistakes and work around problems, yet come up with better results, while STJs make fewer basic mistakes, but more systemic mistakes because there is too much rigidity in their patterns. You might find an STJ trying to apply poorly-suited concepts from unrelated processes to the task at hand because they can't come up with a solution, while an unrestrained NT will start writing a massive framework for doing the task (but not document it). I've seen it more times than I can count - you find an NT (or a group of NTs) and they will have been developing things that are more flexible than they ever needed to be, you find an STJ (or a group of STJs) and they will have been pounding all the square pegs they learned in college into all the round-hole problems they have been given.
Based on my knowledge and background, I have come beyond the NT tendency to write frameworks and toolkits, because I know it's best to just solve the problem with an eye to the future so it'll be just flexible enough to handle the problems that are likely to come up. But I don't go to extremes on this because I also know that only a small fraction of these cases will ever need to be handled. I also try to apply known patterns when they are reasonably-effective, because it's faster and does the job (it's probably the hardest thing to overcome).
I also inherited an undocumented behemoth, the brainchild of a few NTs that no longer work at my employer. The code is completely undocumented, exceptionally-complex, and almost impossible to learn. It was written in a long-outdated language for over six years, touted to be capable of everything, and did I mention that it's overly-complex? The thing is more a huge toolkit and software development framework than an actual software product, but there are a number of applications that were made with it. It only works with some very simple (and peculiar) configuration of the development PC, yet the final produced software is very reliable. Watching people that worked on developing it work with it, it's amazing - simple tweaks to the framework solving a huge array of problems. However, if you have no clue how it works, it's nearly impossible to comprehend because it is written like a video game, massively multi-threaded, and in case I forgot to mention, undocumented.
|