Software development, as often as comparisons are made to engineering or science is neither. It's a craft. The tools differ from person to person, each piece of software is different, a lot of practises are learned and tacit and transmitted from one pracitioner to the next.
Writing code is more like being a blacksmith than being a scientist, as much as people yearn for formalistic answers and peer reviewed studies so that they can have the appearance of objectivity, a lot of software development is fundamentally subjective.
My electrician father doesn’t base his craft on thought leading blogs. He has actual textbooks on electrical equipment, and guides with examples on using tools and solving certain common problems, full of reference material to other books.
I happen to think software is something like 70% craft, 25% science and 5% art, and to me this suggests not putting much stock in armchair blog opinions (and yes that includes Fowler’s) and instead seek scientific or data-backed analyses, even for craftsmanship such as any of the references in the comment above yours.
Embracing this seems misguided, because as software becomes more complex, it's no longer enough to have software blacksmiths banging away with their software hammers and unit testing the end result into a releasable shape.
Teams working on high-availability, high-reliability software know this and take a rigorous approach. Too many take the BDD approach (blog-driven development) and when the thing inevitably falls apart, the customers pick up the broken pieces.
I am not sure of that. A lot of what we see today is just adding an abstraction on top of other abstractions. People use five frameworks for a static webpage that could have been easily made in half a day using editor, css and html without ever writing a single line of javascript.
In a evolutionary sense it works for them, but the reason isn't that html and css wouldn't also work, the reason is that they believe they have to use it.
So I'd be careful with the darwinian comparison here, this is more a mimetic or meme-like evolution of ideas and once the layer below is abstracted away it is quite easy to just forget about it alltogether.
Not to speak for them, but I think it's more that despite decades of ridicule, there's still a lot of COBOL out there humming along. Some of it might even be deciding whether or not your health insurance is going to cover that thing that you went to the doctor for the other day. See also: Visual Basic, ColdFusion and so on.
Approaches that are, nevertheless, still more akin to philosophy than engineering.
Asking what you mean by what you wrote in a public discussion forum isn't seasoning. You aren't being asked for evidence for you opinion, but what that opinion is, and the context of a discussion forum implies you mean to discuss the topic. Also, I can't see the "persistence" or bad-faith that would make it trolling.
The study of the fundamental nature of knowledge, reality and existence is unrelated to writing software.
Software development is applied philosophy. Wittgenstein would’ve been good at it.
But that doesn't mean it has to be "craft" or other highly subjective fields, nor that we can't aim for similar standards as other engineering disciplines.
> Writing code is more like being a blacksmith than being a scientist ... a lot of software development is fundamentally subjective.
But writing code is more like engineering than science, so why not compare those? There is subjectivity to engineering design too, but plenty objectivity in the constraints - what make something "fundamentally" subjective versus some other kind of subjective?
Is it possible we don't need to compare programming to any other profession? We can instead identify there is subjectivity in software development, and that some (but maybe not all) it it could be more objective for a benefit.
He did not say that. He said, that he wants concrete criticism, what ideas are wrong and if possible, why ...
* Code Complete - McConnell
* The Economics of Software Quality - Capers Jones
* Facts and Fallacies of Software Engineering - Robert Glass
* Making Software - Oram & Wilson
... and there are many others. Not all references have survived the test of time, L. Bossavit describes in his book " Leprechauns of Software Engineering" how several software development "facts" are in fact myths.
The approach you suggest suffers from a chicken & egg problem because you first have to be experienced enough to be able to evaluate such reference-free content, as the blog written by Martin Fowler. It's precisely the inexperienced that are likely to internalize such advice and perpetuate personality cults.
Your objections are invalid; one does not need to be better at a task than an individual in order to criticize them. Asking for references and proof is perfectly legitimate, at least if we want to approach our discipline more as engineering and less as "Strunk & White" as someone comically suggested in another comment.,