I'm really tired of this take. Not because it's necessarily wrong but because it gets used as justification to basically abdicate from doing any up-front design whatsoever. And that, IMO, is absolutely wrong.
This is not an appeal to do "BDUF" and try to design every component, service, button, widget, class, etc. of your application up-front in UML or whatever. It's an appeal to actually do some level of rigorous analysis and design before you start building, and apply the wisdom one has access to (from personal experience, from reading about others experiences, whatever) to construct a design that supports the reasonable expectations for your app beyond what is right in front of your nose right this very minute.
I like to suggest we call this SDUF or "Sufficient Design Up Front". And granted, the exact details of the lines between NDUF (No Design Up Front), SDUF, and BDUF are a bit fuzzy and subjective. But that doesn't mean we can't try.
Sadly, the "agile revolution", while mostly a positive development, kinda sucked all the oxygen out of the room with regards to doing rigorous analysis and design and basically killed software architecture as a practice in many (most?) organizations. I think that was a mistake and that it's time to walk that back a bit.
This is not an appeal to do "BDUF" and try to design every component, service, button, widget, class, etc. of your application up-front in UML or whatever. It's an appeal to actually do some level of rigorous analysis and design before you start building, and apply the wisdom one has access to (from personal experience, from reading about others experiences, whatever) to construct a design that supports the reasonable expectations for your app beyond what is right in front of your nose right this very minute.
I like to suggest we call this SDUF or "Sufficient Design Up Front". And granted, the exact details of the lines between NDUF (No Design Up Front), SDUF, and BDUF are a bit fuzzy and subjective. But that doesn't mean we can't try.
Sadly, the "agile revolution", while mostly a positive development, kinda sucked all the oxygen out of the room with regards to doing rigorous analysis and design and basically killed software architecture as a practice in many (most?) organizations. I think that was a mistake and that it's time to walk that back a bit.