I like how as technologists we talk about "betting" on things. As if technology choices are like gambles.
I like how as technologists we talk about "betting" on things. As if technology choices are like gambles.
If there are two competing technologies, generally one will dominate, or more likely: both will be replaced by something new. You have to gamble that the technology you pick will stay relevant for long enough to be worth it.
Instead I believe choosing the right software platform is predominantly a game of skill and experience.
Remember that old canard, "a poor craftsperson blames their tools"; the corollary is that a good craftsperson chooses good tools.
A good craftsman can rationally choose tool A over tool B, however, for software platforms it is a major gamble whether A v2 will still be preferable to B v2 - it is (a) unknown, not decided by A being "better" than B, and (b) generally outside of your control.
In software tool/platform/protocol/standards selection, the person making the choice is a direct participant in the game, can control the local (and possibly global) quality of the outcome, and has (or should have) a priori knowledge of the characteristics and context that will shape the future viability of their choice.
Sometimes that requires a really deep dive into technical documentation and/or source code; sometimes it means assessing product origins, or the community/ecosystem around it; sometimes it means looking for warning signs (e.g. patent encumbrance, a toxic vendor, unnecessary interoperability barriers, opaque licensing or pricing); it almost always means understanding the values & norms & preferences of the people who'll be bound by your choice. Usually a combination of all these things and more.
The argument that we can't tell, in advance, doesn't wash with me; it is not borne out by my experience of being either the decision maker or decision victim, and just sounds like an excuse for a bad decision and/or poor execution.
What happens later is people saying "I made a bet and lost", a terrible post-mortem that smacks of denial, and one I've heard all too often. But whenever I've make a wrong call - and I've made plenty over the years - there was always, always, always some factor that I missed. Whenever I've made a decision that turned out really well, it has always been on the back of research and analysis and reasoning, possibly (and increasingly with experience) guided by gut feeling. "I made a mistake, what can I learn from it?" is the only hindsight I tolerate for myself and strongly encourage in others, and it is not compatible with framing decision-making as a gamble.
And this is the global market.
Imagine how worse it is if you want to hire local. Yikes.
The pool of engineers who are experienced with your stack will be bigger if you pick a "winning" or "popular" stack. But in my experience at Big Co, the stack is as often proprietary as not, yet Big Co can attract talent anyway.
Having lived through transitions from Popular Stack Today to Popular Stack Next more than once, I have yet to go through the experience thinking that the upgrade path is easy.
Proto-jQuery to jQuery, jQuery to Backbone, Backbone to Angular, Angular to React, React to other React-like things... these frameworks were all "winners," but I can't describe the upgrade path without using the word "rewrite."