The late, nice Fred Brooks wrote many phrases about software program engineering. You will have heard of Brooks’s Legislation or learn his seminal guide, The Legendary Man-Month. Even should you haven’t, you’ve in all probability heard at the least one colleague proclaim:
9 girls can’t make a child in a single month!
This can be a highly effective metaphor for 2 causes. First: it’s so vivid that it’s onerous to overlook (and straightforward to recall in a time of want). Second: it makes apparent a not-so-obvious conclusion. Which is:
Including manpower to a late software program undertaking makes it later.
That’s a tough capsule to swallow, particularly when your undertaking is late. However it’s a reality capsule, which makes it price choking down. Right here’s one other reality capsule that Brooks wrote in his paper No Silver Bullet–Essence and Accident in Software program Engineering:
There isn’t any single growth, in both expertise or administration method, which by itself guarantees even one order of magnitude [tenfold] enchancment inside a decade in productiveness, in reliability, in simplicity.
For individuals who don’t perceive the silver bullet reference… it’s pulled from folklore the place a bullet solid from silver is the one weapon that may kill a werewolf. Over time a silver bullet has come to signify “a easy, seemingly magical, resolution to a troublesome drawback.” Brooks may have simply reached for panacea as a substitute, however I believe the shiny attribute of silver makes it the proper analog to software program options.
I’ve discovered myself referencing this a couple instances just lately. Each of these conversations had been within the context of React, however the precept applies to each over-adopted expertise.
So, this put up is a delicate reminder to my fellow software program engineers:
- Does a brand new instrument promise to unravel all of your issues? It received’t
- Do you suppose you’ve discovered the “one true tech”? You haven’t
- Have a shiny hammer that appears nice for hitting screws? It isn’t
Please humor me whereas I quote myself for a second: (Hey, it’s simpler than writing)
We are usually type of lazy and simply take the large hammer and hit all of the nails with it. Like, “Oh, I’ve discovered the panacea. That is gonna remedy all my issues”, as a result of it does remedy a few of your issues big-time.
However it’s is also gonna create different issues, and it’s additionally not gonna slot in each place that you could shove it. I imply, I do know there’s React-based command line builders… And it’s like, okay, should you like parts, cool. That is sensible. However why is React constructing your command-line app? I simply don’t perceive.
After I extemporaneously mentioned “I simply don’t perceive”, I meant that the selection confounds me. However I do perceive why individuals may make it.
Saying “use the correct instrument for the job” is simple, however truly deciding on the correct instrument for the job is something however. Good instruments are onerous to search out, onerous to judge, onerous to study. We’ve got constraints, we’ve got biases, we’ve got shortcomings.
However that’s all a part of the work.
And should you “simply use Go” or “simply use React” or “simply use Postgres” for each drawback that crosses your keyboard, you’re simply not placing within the work.
(And I say this as a man who makes use of Postgres for many issues! 😆)
In case you loved studying this, you’ll in all probability get pleasure from Changelog Information, my free weekly podcast + publication overlaying developer information price your consideration. I preserve it transient (~8 min), entertaining & all the time on-point. Not so positive? Try a current subject!
