Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

The CRUD IDE/Stacks of the late 90's were closer to the domain and that was a large reason for the productivity I used to see: more done with less key/mouse-strokes and less code. Web stacks make one waste time fiddling with janky stack parts & layers instead of domain logic, where it SHOULD be spent.

It's hard to match their productivity with current web stacks, often requiring specialists and/or big learning curves. (Sure, the 90's IDE's had their warts, but were getting better with time, until web murdered them. And I'll limit this to small-and-medium apps rather than enterprise/web-scale.)

It's usually assumed the features that made them productive are mutually exclusive with web-ness. I'm not convinced; the industry just stopped experimenting. What would help is a state-ful GUI markup standard so that GUI's don't have to be re-re-reinvented via HTML/DOM/JS, which is inherently ill-suited for rich CRUD GUIs.

Having your language/tools/standards being a close fit to the domain matters. It's why COBOL lives on despite being even older than me. I'm not saying we should mirror COBOL as-is, only that it still carries important lessons, including FDD: Fit the Damned Domain.



Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: