Unless you're Automattic (Wordpress.com). Or Acquia (Drupal).
"Acquia… today announced that its co-founder Dries Buytaert was named CTO of the Year at the 2014 MassTLC Leadership Awards, hosted by the Massachusetts Technology Leadership Council. Acquia led all companies with four finalists at this year’s awards…" [1]
Those examples show that PHP can be used, in a core way, by great technology companies.
Personally, I think you can be suboptimal in all sorts of ways which cost you, yet nevertheless succeed.
Though one problem: people use a tiny percentage of successful companies to justify their mistakes. One ad company manager told me, "Well Doubleclick's UI was slow!" But the obvious retort, if I thought he'd listen: "You're not at Doubleclick, are you? No, your company's failing miserably.
Same with PHP: "Don't imagine you're Facebook. You'd be lucky to be Myspace." If your tradeoff analysis suggests PHP is the right choice, so be it. But comparing oneself to Facebook suggests commitment to irrationality and monkey-see-monkey-do.
It'd be more reasonable to evaluate PHP in terms of tool quality, but that's too much to ask for in the tech industry.
I fully agree with you. Facebook is also a heavy PHP user, and that's the biggest startup IPO of the last decade. PHP can definitely be used at a great company -- the data simply showed that the better the company, the less likely it was to be using PHP.
I think the causal effect works more like this: People in the startup scene (and particularly in SV) are unlikely to pick PHP as the language of their choice -> Not many PHP companies started by experienced entrepreneurs -> Lower quality.
The fact that PHP is correlated with lower startup quality (whatever that means) has probably nothing to do with PHP itself. If you forced all the smart people to use PHP, their companies would probably work out just as well.
I was with you until the last line. Forcing really smart people to take a project in a direction they are fundamentally opposed to is suboptimal, at least.
I think you misinterpreted was I was trying to say. I didn't mean you should force people to use PHP (obviously that would be suboptimal) , but that the choice of technology has generally little to do with startup success. People who can execute well in terms of distribution and product can do so using any technology. It's just that this group of people is less likely to choose PHP for their projects because of cultural circumstances (SV, press, peers, etc).
A programming language is just a tool, it's architecture and design that matter. A competent developer can write a great product in PHP, many have done so. The negative correlation is trivially explained, the startup world is like a big fashion show.
I quickly browsed the article and I'm not sure how they defined quality. To answer the original question I'd say pick what you're familiar with or what you can easily hire for in your location.
Rails. In w10, we were one of two non-rails users, though it seems that has changed to larger diversity.
Most YC companies use CircleCI, so we get to see some of the diversity. While I haven't got concrete stats on this, I think it leans slightly more heavily rails than usual (and usual is about 50%). Bear in mind that that's skewed in some ways: if they were using C# for example they couldn't use us.
Via a complete and total misreading of his answer, actually! The initial "Rails." stuck in my mind even after he directly noted that they weren't using it.
I wouldn't assume there is a "YC tech stack". From individual blog posts I've read there is great diversity among YC companies' tech stacks. I also wouldn't assume that rails + angular is, or is even considered to be, the cutting edge of web stacks.
Maybe. On the other hand, Paul Graham has strong views on what great hackers are like, which are likely to color his evaluations of candidate companies. It would be surprising to see a YC company running an Azure/Windows/C#/Visual Studio/TFS stack. And there are good reasons why someone might do so.
I believe Loopt was on C# stack and its founder is currently leading YC, so I'm sure Paul didn't take stack too seriously when evaluating companies. There are a lot of good reasons to use Microsoft stack, and I suspect more than a handful YC companies use it.
Paul said it is advisable to not use enterprise technologies if you're willing to attract high quality engineers only. He never said it can't be done and it is never implied that high quality small number of workforce is important to every business. If you're in consulting, for example, you'd want a larger workforce you can get for cheap. But I think you already mentioned that point, just throwing it out there.
You can use AngelList search to get some approximate relative counts of technologies at YC companies. For example, here are the counts for Java and Python:
I use Rails in my company, which is not a startup anymore, but we are using java in a new venture.
Our stack in this new company is based on spring (spring-boot, spring-data-rest, etc) and angularJs. We are very happy with it. Spring security and spring-data-rest are amazing.
I love Rails with all my heart since it's what we use since 2005, but I have to say I'm impressed by the advances in Java in the last year or two. That said, if you are doing SPAs Java can be more productive than Rails.
Yes. We use PHP and we love it! It has its, er, quirks from time to time, but allll that negative stuff that you're always reading about PHP doesn't really affect the bottom line: PHP is great at building websites and scalable web apps, warts and all.
(Edit: a few others in this thread are observing that most PHP-ers got their start in the early 2000's. To add another datapoint, that's true in my case as well. Active since 1998, I started with perl, got familiar with PHP, moved on to MS tech for a few years, then became a RoR dev in ~2005-07. Eventually I got frustrated with Rails' performance, decided to refactor an app in PHP and was blown away by the two-orders-of-magnitude performance difference at the time. I've stuck with PHP since then, though I still play with Node and Python. All great tools. I hate the language wars, use whatever works for you.)
As an anecdote - in my circles I observe that PHP is very popular among people who started working in web ~2000, while ruby is popular among those who started in ~2007. Nothing to do with seniority etc, I think it's simply the default choice of whatever was in vogue at the time when the person started hacking, and by now many PHP guys skew older and are out of the startup game (i.e. working in established companies on established products).
As another example, I'm still astonishingly productive with PHP for backend APIs and rapid development projects. The symfony/Doctrine community has done a phenomenal job at bringing world-class tools to the language.
As a young person, when I was examining possible choices for development, I chose what was fast and stable. At the time that was C & PHP, and maybe even perl.
I don't see how the young people of today decide to go with Ruby, with it's terrible performance, and especially last year with it's slew of vulnerabilities. If you like the beauty of Ruby, use it on your own stuff or for learning. But using it with intention for a production environment? I just don't get it...
It's not Ruby. It's Rails. Ruby on its own is just a slightly more flexible Python.
Rails was embraced because it provided a clear way to write DB-driven web apps with minimal configuration, following a strict pattern. And a lot of people sort of like strict.
I started out with PHP in ~2000 as that's what people used, but then migrated over to Python around 2007 and web.py. Now I've experimented with node and Go, Java and Scala, but I prefer Python and Django. There's so much that Django gives you, it feels stupid not to use it. I haven't tried Ruby or Rails, I just don't feel like I need to.
It doesn't really matter what you use if it gets the job done.
Laravel is indeed the thing I am learning next from PHP, and I love the Startup ecosystem (and working in one), so I guess I'm not the only one. Also, take into account different countries; I'm from Spain and the Rails ecosystem is really weak here.
Pure conjecture, use salt liberally: PHP was 'in vogue' in 1998 to the mid 00's, after that a ton of other (and better) options appeared (I say that as a 'PHP guy'). So if you look at younger people you're likely not going to find a very large number of programmers there that will have PHP as their first language of choice. The older ones though might find that their experience with that particular platform will give them an edge over first learning a new stack.
So to complete the guess I think if you focus on the teams with older technical founders you might find some PHP but in the teams with younger technical founders I doubt that that would have a high incidence.
One's proficiency in a language is more important than using a newly learnt shiny thing out there. Given the people who get in the YC batches are deemed to be smarter of us, I would not expect them to be judgmental about these things.
If you don't change the default styles, sure. It looks decent enough that you can get an MVP into production without having to do much other than a light top coat for branding. Later on, you may want to invest in more heavily customizing some elements, but Bootstrap will pretty much stay out of your way.
Source: Senior front-end dev that puts around one app per month into production using Bootstrap (some more heavily tweaked than others).
