What happened to LOP?

Wouter attended me to this interesting article by Sergey Dmitriev. Blunt axe summary: Language oriented programming will allow the programmer to naturally create DSL's as the need arises. Using the OO paradigm where applicable, using other paradigms in other domains. This will keep the code look more like the design and less like the unreadable crap that we normally write. Be honest: your beautiful creations are not exactly poetry to the untrained eye, right?

There are some more articles about it primarily the one by Martin Fowler, but nothing new really since 2005.

My theory is that getting an application (damn thing) to work it is just too much fuss to use something like MSP. The potential maintenance hell will lead managers to the conclusion that sticking to skills that they have sent their programmers to trainings for will be better. Who hasn't been to that design meeting where somebody started by drawing a very familiar picture on the whiteboard and said: "This is the service layer, this is the business layer, this is the data layer ...."? After that meeting everybody takes their template design documents, SLA's, status updates... and starts filling in the name of the project.

Creativity should be done by humans, filling in templates should be done by computers. But in the end humans tend to favor familiar territory. To be perfectly honest, most of my hours are spent doing stuff that should be automated.

Anyways, it would be nice if this LOP made it's way into mainstream software development. Time to polish up those true programming skills and forget about the compiler in the back of your head that you shouldn't rely on anyway.