Flutterby™! : Insane? 2014-07-20 10:01:03.680074+00

Insane?

2014-07-20 10:01:03.680074+00 by meuon 2 comments

Am exploring retooling my productive coding environment, again. One driving factor is updating everything to the "improved" MySQL connectors in PHP, or using a database abstraction layer to be able to support MySQL/Maria 5.6 or other databases. Another is better supporting some of the good stuff in HTML5. I've been browsing various PHP frameworks this morning and I still find them overly abstract. I might as well learn a new language as well. Am seriously considering writing, from scratch, again, everything.

How often do most of you "retool"? ie: Start over with the same tools using what you have learned, or with a completely new language/toolset?

Or maybe I can make a living as a metal sculpture artist?

You could probably make a living as a metal sculpture artist, though most of the working artists I know have a lower standard of 'living' than I suspect you're used to....

Every new project is an opportunity to retool for me. I'm still happy with Python, but I just chucked Django in the bin and have been enjoying Web2Py (once I understood it and chucked a bunch of it in the bin, too) for web development.

Database abstraction layers usually just get in my way -- not as powerfuL as writing SQL, not as flexible as writing my own models from scratch, harder to read.

#Comment Re: made: 2014-07-20 21:35:20.467454+00 by: meuon [edit history]

dexev: I've spent some time today playing with Symfony and Zend, and already have experience with CodeIgnitor and other friends.. and am now 300+ lines into a fresh look framework with some ideas from that world, and my own. for example: I don't like what the world now calls "routing" for the kinds of things I'm looking at doing, but I use it for CMS style apps. I'm calling it "Glass" (GeekLabs Acrononym's all start with GL..) and am using raw mysqli (improved). My first step is to create a small simple reference application and set of functions I know I'll need. I think the thing that has surprised me is I found myself writing "tests" of what/how I wanted things to work first. I'm not sure of the current name for this thought, I don't want to use the words "MVC" or "Model View Controller" but I want to try to enforce the practice of having code that does work not print out HTML in the middle of doing work. It's a bad habit I have, but it sure is powerful. I've got some time to do this well, no deadline for getting this done, so I'm trying hard to come up with my personal "best possible practices". If nothing else, it's forced me to look into the mirror hard. I might even post this to github as it fleshes out.. I've certainly seen worse there.

Side note: I think I finally almost grok arrays, especially associative arrays and how to use (abuse) them well.