Hacker News Reader

Comment by rms | Hacker News Reader

I guess most coders around these parts would rather not develop in PHP which can make attracting talent harder, but from a pure product standpoint PHP is just as good as Python or Ruby as long as it's developed using modern best practices. Let's not language war right now.

View on HN

Modern best practices include "not using PHP"

Could you explain why? I'm also curious why Facebook, arguably the definition of a modern startup success, chose to remain powered by PHP after they obtained more than enough capital to switch to say Ruby on Rails.

Keep in mind Facebook actually wrote their own PHP interpreter, which doesn’t support eval() (treated like a core language feature by countless crappy developers). And that they are hardly all-PHP, having created and open-sourced a mechanism to shuttle data around between languages.

Yes, Facebook rewrote and released a faster PHP runtime, which means they embraced rather than shunned it.

With comments like the GP you'd think FB hobbled along with PHP on its way to becoming as big as it is now.

I recall reading that they wished they could move to Python, but they decided that rewriting their entire base of front-end code would be a bad decision. It was too large of a task (I believe a Facebook engineer covered this when they released HipHop). So, they are basically stuck with PHP and have made the best of it.

Yes, many companies will review all their options once they have plenty of capital, and make decisions based on what's best for their situation. For example, Twitter was reported to be switching from Ruby on Rails to PHP or Java amidst all their uptime troubles.

http://techcrunch.com/2008/05/01/twitter-said-to-be-abandoni...

My point is that choice of language is not what will likely stand in your way of becoming a monumental startup success. Yet people will insist on promoting language wars as if it will.

Paul G. has given Lisp a lot of credit for enabling ViaWeb to push out new features in minimal time, compared to competitors who were using languages such as Java. While your choice of language might not be a make-or-break decision, it's good to have all the advantages you can get.

(I think Twitter ended up using Scala, didn't they?)

We could always do an "Ask PG" to clarify, but my understanding of what happened with ViaWeb is you had a group of smart young people innovating on a problem against bulky institutional competitors. When asked who had an offering even close to what they did they said no one. (You might even say they were the first YC modeled company!) Viewed this way I'd say the success was about the people, not the language; they simply used their tool of choice.

I don't know what's the "official" (or real) story on current Twitter architecture.

Right, their success was not purely about the language, I'm just saying it helped (rather, he said it helped). This is drawn from my recollection of reading an article pg wrote. I could find the article if anyone is interested (it is an article about how great lisp is, not specifically the history of ViaWeb, I think).

I agree that some people take the choice of language and architecture a little to seriously. Of course you can start a blockbuster business on PHP and MySQL - people prove that frequently!

Yes, I agree any creator needs an able tool, but that's a clear strength of PHP when it comes to Web deployment. That's a fact. I believe people should use whatever they feel they can get the job done with, even if it's BASIC (http://news.ycombinator.com/item?id=2257866), because focusing on a language more than what it is you're actually trying to accomplish IMO is missing the point.

Why PHP is not a good idea:

http://www.quora.com/What-are-the-crawling-horrors-of-PHP

Why facebook has stayed on PHP:

http://www.quora.com/Why-hasn-t-Facebook-migrated-away-from-...

Reply on news.ycombinator.com