L'architecture progressive - Matthieu NAPOLI - Forum PHP 2019

[@bernard-ng]

merci pour cette conf

[@renarsdilevka6573]

I understand that you would not use DDD for Wordpress website and use custom SQLs for reporting, but business logic is the heart of the software. At least having precise description of the business or Ubiquitous Language sets the bar of quality. I have seen enterprises today, which claims the perfection of being professional and attentive to detail is not just the way how the approach works. Well the question what works then? Why developers are neglected in their challenge every day? What is the point then to become a professional? Just coding is not very interesting, solving exact business problems is far more interesting, seeing how it all works together and be proud of being part in it. I have seen a lot, now it is about DDD, Clean and business domain. I do not see other way to approach the problem, and tools we use today the reason why we chose them has to be challenged. I am not just a code monkey, and we will not ship shit, right? And getting to your point in the talk, sometimes also not having PRs work, sometimes not having tests work, sometimes sloppy code works. I think the quality of the outcome can be as high as you and your peers set it. Just need to find a team, like minded people which sets the bar. Robert Martin, Martin Fowler, Kent Beck those are examples which have seen different results of quality in their long experience. We should not neglect it. We should read and understand what is said. Hype is not an option. Being a developer does not mean yet that you are conscious. Pragmatic approach is good but there are limits which have to be challenged. Over engineering is not recommended, but software changes faster than we can anticipate. For those people who write tests first or cover the code fully with tests, for those people who do rigorous PRs and for those who just do not accept sloppy code and anticipate by experience, they do not claim that is too much classes to write. The problem in industry today, that we are promised to get more time later for a sake of creating new business product, but sadly, later never comes. So i feel neglected, feel misunderstood. And you are paid to deliver the obvious. You are paid for bad news. You should not lie, you should not just close the eyes to favor the business needs. And companies cannot promise that we will have time later. Because they need to move faster too. So you have vicious cycle, sorry, but that is very sad. The time while you get to the point that maintaining the quality is needed now to go well, this just sucks energy and free will to collaborate. And yes, they throw the product in the bin after some time, and phase starts to do refactoring, and they loose the faith, because it is so sloppy, so undone in the first place. Three levels of refactoring, you ar lucky to get only one. It is kind of nobodys fault, but dev side is one to point to, like we have never told so. And that is not a problem of technologies chosen, it is about that dev side is neglected, something to fix later, no time now. This is problem of consciousness, what we are in business for. Right approach comes with many years of experience. And the decision should be conscious and based on facts. Legacy code is the sign of being unsconcious what you are writing, is it easy for others, is it maintainable, is it changeable. And code without tests, i would say, there is only little value. There is a value as much as the developers were conscious at that point of time, what they were doing. And feels like you are trying to justify the sloppiness of code to the business needs. Well you would be favourite, but does it stand long enough without headaches. Or we just nod anemically and cannot do anything about it. Because you have to have guts to take control over legacy code. About management would not agree that team composition should decide what architecture to put in place. We would never learn anything. The question today, who wants to learn, and who wants to earn.

[@yannickg5788]

instructif, mais un peu bizarre, comme si c'était un plaidoyer pour la nullité comme argument économique pragmatique, oui possible mais nivellement par le bas

[@renarsdilevka6573]

I have seen multiple times, you just do not write code in native language, like French. Business language in computers is english. Ubiquitous language should be english, code should be in english, comments should be in english.