On March 2017, Eugene Gusev had an interesting talk (in Russian) at HolyJS regarding the technological singularity as whole in Front-End, and JavaScript ecosystem in particular. As a thesis taken, he noted an enormous number of 350k packages in NPM.

Few points being stated over the talk:

  • Quality of the packages (aka younglings publish low quality tools)
  • Hype over the ecosystem (frameworks and libs go up and down in popularity scale).
  • Business forces the choice, leaving minimum time on technical decision-making process.
  • Stop writing code, or “stop publishing your code”.
  • In order to program you should be a computer scientist.

All the points are quiet controversial. Though, accepting some of these statements as a potential problem, most of them have a reasonable explanation.

Prelude

Back when the trees were tall and the grass was green, I graduated my BSc in Computer Science, I’ve started working as a Junior Web Developer. By that time, a programmer who knew only JavaScript, or solely was coding on Perl/PHP/Python was considered lazy/unemployed/unicorn. The time flew by and the industry started splitting technological stack, so from Programmer/SysAdmin we forked into Front/Backend developers, SysAdmins to DevOps etc.

University, really?

We still have certain professions that require an educational degree. By paying for your education, you’re prompted to access research laboratories, expensive equipment that’s required for your research. Think of physicists or chemists. Stating that diploma (for Computer Science) somehow justifies your profile, or what you should be doing for living left behind with the era of industrial revolution. “The spice must flow” - that’s where we get into the almighty Internet for self-education. Udemy, Coursera - it’s right there, most of it is free, just read it, learn it.

Technological Singularity

Technological singularity is a hypothesis that the invention of artificial superintelligence will abruptly trigger runaway technological growth, resulting in unfathomable changes to human civilization.

Are we there yet? Yes, we are. It just arrived silently. Clustering professions to narrow the specialization is one of the countermeasures to prevent the informational noise that we get once diving into IT-sector. That’s where we step into Hype effect.

Hype Effect

Hype Cycle diagram Hype Cycle by Ember in the Real World

350 thousand package in NPM. That’s massive, but hold on a minute! CPAN didn’t have the same issue in Perl community? PEAR and Pecl repositories stacked by repetitive packages, written by people just because they could. And they did! Every year, some big player like Google, Facebook, name a few, presents a new approach towards a common problem. Angular, React, EmberJS, VueJS. It automatically triggers a hype.

Architects in the companies start massively migrating to trendy frameworks, as it’s backed up by one of the above companies, which promises stable development, stable versioning and ongoing support. We get a drastic shift of packages developed for these frameworks. And the story continues until something new arrives on the market.

As companies do not operate in vacuum, all these transitions correlate with business decisions. Packages appear in almost-ready-to-us state, and remain loosely maintained just to get your **npm/phake/rake **running smoothly during the deploy process.

Packages & Plugins flood

Slowly the picture of 350k packages come into place. Each of the packages is a reflection of a person behind it and the problem he or she was trying to solve. Good solutions become trendsetters itself, become community defacto standards. What happens with others? - Well, it’s a pure darwinism, which triggers professional boost among the developers.

To sum things up

Your knowledge derives from those who dared, and published something. If you cannot choose out of a hundred packages one or two that might fit in your application, maybe you should question yourself what exactly you trying to solve? Diversity of approaches in IT and its openness is our strength. You found something wrong, write a bugreport, or send a pull request. The rest is just excuses in most of the occasions!