[wvenable]

> PHP is considered extremely weak performance-wise

I don't know, it seems a lot of the very high trafficked sites (Facebook, Yahoo) run PHP. Yes benchmarks do show PHP doing poorly, but when was the last time anyone used PHP to crunch numbers? It's apparently fast at what it needs to be fast for.

> There is poor to no support of multithreading

That is not a flaw. PHP is a share nothing architecture; each request is completely independent. This is a good match for the stateless nature of HTTP. Multithreading isn't at all appropriate for PHP and I can't think of a scenario where it would be needed.

> There is no eventing system

PHP is a programming language not a framework. I'd make more of a comment about the various PHP frameworks but "eventing system" seems kind of vague. Maybe you can clarify this a bit and then I'll comment.

> As a web framework, it sins in mixing code & design (opposed to e.g. Django-Python or Ruby on Rails)

Nobody mixes in code & design in PHP anymore -- you won't find it terribly common among the various PHP frameworks. I do use it from time to time as a debugging aid and my compiling template engine uses it as the compilation target.

[shmichael]

> I don't know, it seems a lot of the very high trafficked sites (Facebook, Yahoo) run PHP

I'm completely for PHP as a simple and accessible language. The previous author wanted to hear about the shortcomings of PHP and I tried to provide some. PHP sites such as Wikipedia probably need more servers (since every request is a separate process). They might be saving a lot of manpower by using PHP, so it might be an OK tradeoff.

By the way, Facebook uses Erlang for its performance-critical parts (such as chat).

> PHP is a share nothing architecture; each request is completely independent.

You could still use an agent-like messaging system for shared-nothing multithreaded/multiprocessed environment.

> I'd make more of a comment about the various PHP frameworks but "eventing system" seems kind of vague. Maybe you can clarify this a bit and then I'll comment.

e.g. When using cURL you must poll the handler until the request is complete. Obviously you'd need multithreading to achieve a better control flow in such a scenario.

> Nobody mixes in code & design in PHP anymore

I've written and debugged some wordpress plugins, and I can't say I came to the same conclusion as you...

[wvenable]

> PHP sites such as Wikipedia probably need more servers (since every request is a separate process).

Citation needed. But every request isn't a separate PHP process; PHP is linked into Apache. Apache starts up a pool of processes and reuses them for requests. It's all highly efficient.

> By the way, Facebook uses Erlang for its performance-critical parts (such as chat).

Actually, it's not so much the performance critical parts but the parts where PHP's share-nothing architecture isn't appropriate. I doubt they're using a front-end web server for chat either.

> You could still use an agent-like messaging system for shared-nothing multithreaded/multiprocessed environment.

If you had the need, I suppose you could. Although I'm not sure what point you're trying to make here.

> When using cURL you must poll the handler until the request is complete. Obviously you'd need multithreading to achieve a better control flow in such a scenario.

You don't really need multithreading. JavaScript has callbacks for everything, for example, but isn't multithreaded. So this isn't so much a problem with PHP as it is a lack of design of cURL. But then cURL is a C library.

> I've written and debugged some wordpress plugins, and I can't say I came to the same conclusion as you...

Wordpress is old and not well designed. (Much like PHP itself)

[shmichael]

Apache starts up a pool of processes and reuses them for requests. It's all highly efficient.

It's highly efficient if you have a small number of requests at any given time; It's not as efficient for peak hours.

If you had the need, I suppose you could. Although I'm not sure what point you're trying to make here.

There are several use-cases I can think of, the simplest being more than one blocking request for rendering your page, and using multi-process to reduce the latency to the maximal block time instead of the sum of all block times.

You don't really need multithreading. JavaScript has callbacks for everything, for example, but isn't multithreaded. So this isn't so much a problem with PHP as it is a lack of design of cURL. But then cURL is a C library.

Javascript on server environments needs solutions such as node.js. On the client side, well, most UI apps have a single UI thread and delegate jobs to worker threads. In javascript, these are usually done via XMLHttpRequests.

[wvenable]

> It's not as efficient for peak hours.

Actually it's quite fine during peak hours. You could, however, say it's a waste of resources during the lull times. But I'm not using those resources for anything else anyway; the entire purpose of the server is to service web requests. If a bunch of processes are sitting there wasting memory being idle, that's not a problem.

> There are several use-cases I can think of, the simplest being more than one blocking request for rendering your page

That's a fair case but I haven't personally encountered it yet. A good example, however, would be building a mashup page of content from a bunch of different web services.

[jrockway]

You don't really need multithreading.

It's true; many languages allow you to handle concurrent downloads by letting the OS call a function when the socket with the data is readable. (This is part of an "event loop".)

Many language implementations call this threading, however, because it is. You don't need one heavyweight OS thread for every thread of control in your application; many applications that make heavy use of IO easily handle thousands of threads inside one OS thread. It's a very useful abstraction.

[jrockway]

Multithreading isn't at all appropriate for PHP and I can't think of a scenario where it would be needed.

Not blocking the entire PHP process while waiting for database results, waiting for a file to download, waiting for memcached to respond, etc.

I would (and do) use an event loop for this sort of things, but lightweight threads are technically a better abstraction. Web apps would use a lot less memory (and hardware) if people were not so afraid of threads.

(This happens a lot; a bad implementation of something taints the whole category. pthreads, Java threads, Perl ithreads, etc., are broken, so all threading is bad. C++'s object model is bad, so all object oriented programming is bad. It's sad.)

BTW: some good thread implementations include Haskell's and Perl's Coro.

[wvenable]

> Not blocking the entire PHP process while waiting for database results, waiting for a file to download, waiting for memcached to respond, etc.

The thing is, I need those database results -- I can't do anything else until I have them. I received a single request from the user and I'm spitting out a single HTML response back. That's what PHP is all about. And mulithreading as optimization itself wouldn't improve performance because I'd be taking away CPU from other requests that are running concurrently.

> Web apps would use a lot less memory (and hardware) if people were not so afraid of threads.

Fundamentally I don't disagree; there's a lot of cool work going with event-based web servers. However, I think those designs are great for very specific tasks like chat servers and but require too much fiddling when doing traditional stateless web work. However, it may indeed be the future.

[jrockway]

The thing is, I need those database results -- I can't do anything else until I have them.

You could start requesting the database results for a user that is now waiting in the tcp connect queue, cutting latency significantly.

Most people handle this with load balancing between processes, but your blocked-for-DB-results-process is sitting idle using system resources, like memory.

If you only have one user using your site at a time, then this isn't a concern. But for everyone else, blocked processes are wasteful.

but require too much fiddling when doing traditional stateless web work

Well, yeah, because you don't have threads. When you do, everything is handled by libraries for you; you can think you're blocking, but actually not block the process.

[wvenable]

> You could start requesting the database results for a user that is now waiting in the tcp connect queue, cutting latency significantly.

That's already how it works. If one process is blocked waiting for results, another process will get it's turn. Apache is one process per HTTP connection (when not running threaded).

> but your blocked-for-DB-results-process is sitting idle using system resources, like memory.

Either way it will need to use system resources. I can't throw away the resources of the user waiting for results, I need that once the results come in! Processes really don't take up that much memory and typically you're reusing a small number of them over and over anyway.

> If you only have one user using your site at a time, then this isn't a concern.

If somebody is blocked then someone else's request gets a turn at the CPU. Blocking a single process is completely inconsequential to multi-process web server serving multiple users.

[jrockway]

Processes really don't take up that much memory and typically you're reusing a small number of them over and over anyway.

They take up several orders of magnitude more memory than lightweight threads. I did some benchmarks a few months ago showing this; check searchyc if you care. Edit: here you are: http://news.ycombinator.com/item?id=794409

Anyway, when you start thinking about "web applications" instead of "web pages", lightweight threads make a lot more sense. And, you'll find that treating a web application like it's just a bunch of web pages leads to many problems down the road. (Performance is the least of your worries.)

Finally, "I definitely don't need to know about this abstraction because my favorite language doesn't have it" is dangerous thinking.

[wvenable]

> They take up several orders of magnitude more memory than lightweight threads.

Premature optimization; when a 1st grader can count to the number of processes you need for large scale web application optimizing it any more is a waste of time and effort.

> You'll find that treating a web application like it's just a bunch of web pages leads to many problems down the road.

Really? I find it refreshingly helpful. Instead of your large application being a large application, it actually just a series of very small applications. If a page crashes, it's pretty much no harm at all. If your whole application crashes, you're screwed. You want to add some functionality, just add it on.

> Finally, "I definitely don't need to know about this abstraction because my favorite language doesn't have it" is dangerous thinking.

"Oooh shiny" is also dangerous thinking. Is it really better to waste your time using up as many threads as possible or actually solving real problems for a few hundred processes?

Sidebar: Why is always you, Jonathan Rockway? Just how many posts do you make on Hacker news in a day? :)