Hacker News Reader

Comment by SergeAx | Hacker News Reader

The most interesting thing about PHP is it is, AFAIK, the most community-driven language today. Go, Python and JavaScript are all under heavy influence of Google, C#, VB.NET and TypeScript is Microsoft's turf, Java is Oracle's, Rust and WebAssembly - Mozilla's.

С and С++ are holding their grounds, but they are more and more niche today. Facebook tried to domesticate PHP, but failed spectacularly. So it looks like PHP is most democratic and liberal programming language of today) I am glad it's getting steadily ahead.

View on HN

Rust has a lot of core team members that no longer work at Mozilla, so this isn't really true today.

Also, C and C++ are some of the most widely used languages in the world. I wouldn't call them niche.

Edit: according to the TIOBE C/C++ are ranked higher than PHP

... and it's been that way for a long time; we operate on consensus, and in 2014, non-Mozilla folks were added to the core team.

Today, it's extremely diffuse; there are ~200 people on the various teams that hold decision making power over the project, and only like six are Mozilla employees. The core team has 3 Mozilla employees and 9 people total.

Thank you for those numbers, it confirms my intuitive conclusuion: 1/3 of core team is a significant influence.

Sure, there’s no doubt that Mozilla has influence. What’s important to me, personally, is that they cannot dictate what happens. And we have non-theoretical examples of where that’s the case; Servo would have really liked some form of inheritance to model the DOM, but the team overall said no. Also, note that the core team doesn’t make most of the decisions, unlike core teams of other languages. They also have a minority of representation on the language team as well.

And, as someone who was involved before, during, and after working for Mozilla, Mozilla just plain doesn’t have interest in really guiding Rust development. If anything, it’s diminished over time, after all, when Rust started, it was sponsored by the CTO and a director, both of whom eventually left, but also that thought it was important for Rust to be as independent as possible.

Google does not care that much about Python, at least not when I was working there. I would certainly not mention it together with Go.

Google does not care about python is not the same as Python being controlled by a small no of Googlers. I don't know much about whether that is true or not though.

Aren't Elixir and Ruby primarily community driven too?

I know almost nothing about Elixir short of there is a language with this name and it related somehow with Erlang. It's a very niche language/VM good for high-load realtime events processing, thus only noticable project on Elixir is Discord (and on Erlang - WhatsApp, Facebook Messenger and RabbitMQ, which kinda gives a hint too).

Ruby is a brilliant language (pun intended), but from where I am (web- and mobile software development for last ~20 years) it seems to be steadily declining. Headhunters having a hard time looking for Ruby devs now.

> Headhunters having a hard time looking for Ruby devs now

Sounds like the supply is lower than the demand for Ruby devs, which sounds to be like the opposite from "steadily declining".

Also, one should look further than usage when it comes to evaluating a language.

I see your point. My point is that to be called "community driven" the language should be relatively popular. Otherwise there would be an absence of community on the first place.

IMO Ruby/Rails became legacy tech. Are you aware of any fresh significant projects on Ruby?

I thought "community driven" also meant that the language itself was primarily developed by the community and not by someone with monetary interests. And in that regard both Ruby and Elixir are very much community driven — and both regularly see new releases too.

And in terms of numbers. If you look at e.g. number of repos on GitHub with at least 10 stars[0], Ruby comes in as #7 ahead of C# (#8) and Go (#9) and faaar ahead of Rust (#17).

Interesting side-note. C and C++, which you called "niche", clock in at #6 and #5 respectively.

[0]: https://github.com/oprogramador/github-languages#10-stars

This is because Rust is a very young language, and C and C++ are very old. Number of projects on GitHub is more proportional to language's age than to it's current popularity. To check popularity we have to take a first derivatve of these numbers over time.

Thanks for the lunk though, it does shows some interesting stats.

I don't think age is that much of a factor because TypeScript also places higher than Rust and is two years younger, not to mention Go and Rust are only a year apart but they're pretty far apart on the chart.

More data here: https://githut.info/

You may think of TypeScript as a modern dialect of JavaScript. I don't think it applies to pair Java-Kotlin though.

By that account isn't Rust then a modern dialect of C++ and Go of C?

(That web page has data from 2014)

PHPs feature development are driven by a relative few “core” people who understand C, and who seem intent (despite what the larger community wants) to keep making PHP more and more like Java.

IMO Some stuff is great of course but things like their planned JIT additions (which effectively make PHP faster at calculations, which few people use PHP for) and type enforcement (in a dynamic language with no compile step) seems to really miss the mark of what makes PHP great and so widely usable.

> keep making PHP more and more like Java.

I know this is meant as an insult but it's a poor one. There are programming languages features out there that are good on their own merits and PHP is incorporating those features. If they also happen to also exist in Java, or C++, or Haskell then they do. Adding those features might make a language more like X because X also has that feature but that doesn't mean it's the goal.

In many ways PHP does things better than Java by virtue of having hindsight on how those features exist in other languages. The way PHP handles nullable types is better than both Java and C#.

PHP feature development is driven by a rather considerable commitee of core devs and representatives of major frameworks and libraries. All features are undergoing a voting procedure.

PHP JIT is intended to reduce latency and large codebase start-up delay, two biggest problems of modern PHP as for me. It's not about calculations.

It will also help with calculations, maybe the intention is to grab some of python's share.

That would be a very long shot IMO. But we'll see eventually.

What would you like to see instead improvement wise? The typing features are optional, so everything that makes PHP simple to use is still there and available.

I have some patches to improve the DOM API extension for example, simplify its usage which would make it much simpler to use.

> keep making PHP more and more like Java

I have not need to create a class to encapsulate a GET request parameter in PHP. In Java it was common the last time I used it.

Ruby?

Also, why do you say that Python is controlled by Google?

See above: Ruby is good, but it's on decline and IMO soon going to be a niche legacy language for projects too big to rewrite, like Github and Gitlab.

Python was influenced by Google and that influence is hard to wear. Until now Python is definitely not community-driven. Also, look at this version 2/3 snafu.

Probably because they employed GvR from 2005 to 2012.

Yeah, but 2012 was 7 years ago.

And then Dropbox.

Reply on news.ycombinator.com