While working on a refactoring task of a legacy app, I came across a scenario where a series of cause->effect type events are happening and currently everything is being handled inline.
To give you an idea, a user uploads a file, then it is converted into a number of format, and on completion/failure of each format conversion task, an email is sent to some predefined addressees and on email delivery success, it is also logged to some file.
On first glance it seems like a typical use case of an observer pattern or pub-sub but I have came across this post and the linked paper and it got me thinking if I am going to end up into a mess for which I might not be ready.
I have been kind of living under a rock for quite a while now when it comes to the latest changes in software industry. If observer is not (anymore) the right or proper way to decouple the problems belonging to the type mentioned above, then has the industry (by industry I assume the mainstream of software engineers) already come up with an alternate that I don't know of?