I have many project­s. I al­so have a plan.

In the past, many of these projects would not ex­ist. They would be a source file in­side some oth­er projec­t.

• I would have a file that con­fig­ures log­ging, rather than re­quir­ing Oplog
• I would have done a long, com­plex CLI pars­ing func­tion rather than us­ing Poly­do­copt

But since I have been cod­ing in Crys­tal for my hob­by pro­ject­s, and cre­at­ing and us­ing Crys­tal "shard­s" is so easy...

Well, that caused a rad­i­cal change on how I ap­proach things. In the past, I would just code in a fea­ture in­to what­ev­er lar­gish project I am work­ing on at the time, and maybe, some­times, I would spin off some­thing in­to a li­brary.

By maybe and some­times I mean I can't re­mem­ber a sin­gle time I did that.

Maybe it was be­cause I was us­ing Python, and pack­ag­ing Python code in a way that oth­ers can use it is sig­nif­i­cant­ly more painful than do­ing it for Crys­tal, or that I was just do­ing it wrong.

When I want a fea­ture, I do the thing ev­ery­one does: LOOK FOR A LI­BRARY right? Then why when that li­brary does­n't ex­ist was my first in­stinct to just code it in­to the project I was work­ing on?

If I want­ed a li­brary, why was I cod­ing a fea­ture in­stead­???

So now, when I want a fea­ture, I look for a li­brary. If it does­n't ex­ist, I CRE­ATE THE FREAK­ING LI­BRARY

So, I want­ed Base16 theme sup­port for some­thing. I looked for a li­brary. It did­n't ex­ist. I cre­at­ed Six­teen.

Not on­ly is Six­teen smal­l, self­-­con­tained, and easy to use, but it is re­us­able, so af­ter I use it in Nicol­i­no I can use it in Mark­ter­m.

I think as open source de­vel­op­ers we are leav­ing a lot of val­ue in the ta­ble by not do­ing more li­braries.

So, that's the plan. The plan is to keep cod­ing for fun, but al­so to keep cre­at­ing li­braries when it makes sense, so I leave crumbs for fu­ture (prob­a­bly in­ex­is­tan­t) de­vel­op­ers to pick up.

Have fun!