Crazy Code (A ReBlog)

Ever feel like the world is filled with crap developers who think too much of themselves sometime?

You may not be alone.

There seems to be consensus: there are too many "cool" bits of software out there
(which may in fact be just crap written by Monkies sleep deprived developers).

This is why time proven frameworks are amazing; they make code that goes, that works with other code, and works with large bodies of employable developers. I'm beginning to see the point behind consistent, proven, legacy systems, as much as bleeding edge might present its visceral allures.

This is NOT an argument against writing your own shims/libraries. In fact:

1. Its an amazing learning experience. Every library you don't write is in fact a set of problems you never learned to solve, which you may need to solve still because said library wasn't general enough.
2. Sharp vs Dull -
   1. Generalities are hard, and tend to be too weak to handle most cases effectively, only powerful enough to handle most cases just well enough
   2. General libraries also tend to be large, especially if they're advertising how small they are. They may also be difficult to pick up
   3. In house libraries can be incredibly light weight, and efficient, however they aren't going to make the toast or mow the lawn for you as well.
   4. In house libraries can be easily picked up by new recruits, and since they're easier to maintain in the first place, may be better documented

Its really whatever's best for the job. Monolithic libraries save everyone lots of headaches, but then so do agile, discrete service orientated libraries. Perhaps the discrete service libraries should be twice as carefully out-housed.

Me? I'm sticking with DOM+CSS+jQuery, Java/Python/Perl, and C++/C. Here's to another fifty years of consistent software!

Refs:

• Crazy World of Code (Original)
• Discussion

(Btw, love the new inline html editor, Google! - you may have noticed the atrocious inline styling ;) )