The description for this subforum covers "anything not about webmastering", so it seemed like the right place to discuss rant about PHP.
PHP 5.2.10 was released on Jun 18, and "focuses on improving the stability of the PHP 5.2.x branch". 5.2.x, you will note is the "stable" branch.
5.2.10 wholly breaks PEAR.
It is particularly nasty for at least one class of users -- php-fpm's recommended release depends on PHP 5.2.10. So let's say you also depend on some PECL modules, you'd have to build them yourself, or worse, cross your fingers while statically compiling them with a patched PHP-FPM configure.
Okay, so this isn't the first time a serious regression has occurred in a large project with a QA team and thousands if not millions of users. And yes, I'm sure patches are welcome. However, it would still remain inescapably evident that PHP is a project in distress, especially when compared to the quality of development in, say, the Python camp.
I look forward to migrating yet another deployment from PHP to a better system! 
pr0gr4mm3r posted this at 20:57 — 27th June 2009.
He has: 1,423 posts
Joined: Sep 2006
Fascinating. I rarely have a problem with the language itself. I might learn Python if I have spare time, but PHP fills my needs as an online scripting language, so why change?
PHP Starter
Abhishek Reddy posted this at 04:53 — 28th June 2009.
He has: 3,348 posts
Joined: Jul 2001
If by "needs", you've accounted for the future as well as the present, then there is no reason to change. If not, remember that the Titanic stayed afloat serving the needs of its passengers, right up until it sank.
I don't mean to convince anyone else to change with this thread. Rather, I wanted to log why, in a specific and real circumstance, one might want to. In my experience, this occurs with increasing frequency in PHP deployments.
I named Python only for the software engineering aspect (it's a comparable project in age, type and scale). When I say "a better system", I mean whatever system is best suited to the project at hand, be it Python, Ruby, Lisp, Java, or whatever.
abhishek.geek.nz
pr0gr4mm3r posted this at 13:56 — 28th June 2009.
He has: 1,423 posts
Joined: Sep 2006
I never used PEAR, because not all environments had it installed. I would develop applications on my test server, and then find out that the destination deployment server doesn't have that option. It's so much easier to bundle the classes with the application.
Java is a joke in the web environment (servlets, JSP). I did a web application for a class project, and just getting the server environment working correctly was a huge challenge.
PHP Starter
teammatt3 posted this at 16:28 — 28th June 2009.
He has: 2,076 posts
Joined: Sep 2003
Breaking PEAR is a large mistake, and it shouldn't have happened. But I certainly don't think it's "inescapably evident" that the project is in trouble. Obviously, someone (or some team) neglected to test the version with PEAR. Testing is often not the most exciting thing on a project, and it is something easily skipped over. Sure, that's not a great excuse for what happened, but again, it doesn't mean the entire project is in distress. Well actually, you might say the project was is distress from day 1
.
I have a feeling that some programmers, especially in academia, despise PHP. The language enables un-skilled “programmers” to create rather sophisticated, but poorly designed applications. Some academics might view this as a threat. Formally educated programmers are no longer required for moderately sophisticated web applications. A 15 year old can create a program just as well as a PhD, from the user’s perspective (under the hood, it would be a different story). While I don’t have much (or any) evidence to back that up, that’s the vibe I get from reading people criticize the language.
How much is the Python project paying you to pick on PHP, anyway?
Edit: Actually, I think this particular screw up shows the overall health of the PHP project. When people stop developing newer versions of the language, ignore its flaws, and don’t bother submitting bug reports, the project is in distress. That's the exact opposite of what's going on with PHP.
My Site | Regular Expression Tester
Abhishek Reddy posted this at 09:12 — 29th June 2009.
He has: 3,348 posts
Joined: Jul 2001
My only advice is to gather evidence to test that view. I suspect you're barking up the wrong tree.
Hah, nothing, unless you count all the work they do and code they share freely.
I have no special fondness for Python. It's a relatively good system and currently yields optimal utility for some purposes. No doubt many other systems could also accomplish those tasks acceptably, and for other purposes they may be even better. Typically I make an effort to name multiple alternatives in my critical posts.
"At least it's not dead!" isn't terribly encouraging.
abhishek.geek.nz
JeevesBond posted this at 13:16 — 1st July 2009.
He has: 3,895 posts
Joined: Jun 2002
Excellent thread! Who uses PEAR anyway? I heard somewhere that the way PEAR and PECL relate to the PHP project is confusing, particularly due to the amount of stuff in PHP core that should be in libraries.
I personally think PHP is the new Visual Basic. In short: easy, but a bit shit.
There is a lot of anti-PHP snobbery, just read Slashdot or Reddit some time and see how many people believe all PHP developers are retards, without exception. It is possible to know and understand the limitations of a language, but still use it. I believe this is what Matt is getting at, if it is, then he is probably not barking up the wrong tree, although I do suspect he's extracting the excretory fluid.
Out of interest Abhi, is there a way of deploying Lisp or Python applications as easily as PHP ones? PHP is still stupidly fast to deploy and setup, and many of the bad parts of the language can be glossed over.
a Padded Cell our articles site!
Abhishek Reddy posted this at 20:59 — 1st July 2009.
He has: 3,348 posts
Joined: Jul 2001
If you want to do something PHP doesn't have builtin support for, there's not much choice but to extend it with PEAR or PECL. Arguably, you shouldn't be using PHP for anything it doesn't support by default. (The counter, then, is that you shouldn't be using such a rigidly impoverished system to begin with!)
Certainly I agree with your statement. Even with my itchy migration-trigger finger, I continue to use and support PHP deployments. Mostly, my criticism is with the PHP project and its founding developers' work. I reserve some of it for users who thoughtlessly, irrationally or negligently choose or advocate PHP for tasks when better alternatives are available.
I guessed that he drew from my posts a sense of academic elitism. This cannot be the case, however, as I am not an academic. My opinions are based on experience in the "real world" (whatever that is). In fact, I'm critical of academia for having abandoned such high standards that, ironically, I've found are still valued in industry.
This is not a simple situation anymore, and I think it warrants some discussion.
In the case of shared hosting, PHP is still king. You can generally expect your Drupal, phpBB, WordPress or other substantial application to install and run anywhere.* But this is of declining value and truth.
VPSes and "cloud" nodes are cheap and only getting cheaper. On a private machine, you can freely set up other systems as easily as PHP. In fact, if you chose to use PHP, say, via FCGI and nginx rather than Apache, you could have a somewhat harder time. On dedicated servers, this was always the case.
At the same time, models for deploying web applications have matured. It's common practice to manage even a small web app server-side with a build system or other sanity framework. Likewise, to use and distribute modules at a platform level -- not just application level. PHP environments lag in these respects.** The demand for these capabilities also exceeds typical shared hosting setups, since they can't really support them.
Even so, shared and managed hosting with Ruby and Python -- sometimes so specific as Rails or Django application hosting -- is growing too. For many cases well within the shared hosting threshold, I would expect to find an economically persuasive alternative with one of the above.
Recently, in another thread, I linked to nanoc, a static website compiler in Ruby. This is another vector for replacing PHP if your use case is simple enough. As a bonus, you can get away with even more basic hosting.
The point is that it is often just as easy to use something other than PHP. Sometimes it's not, and sometimes it's even easier. If you're prejudiced towards using PHP for your next project, it's probably worth re-evaluating your options. The playing field is more level than before, and it continues to tilt daily.
* Well, not exactly. Some applications like dokuwiki provide scripts to be run from the command line. Others don't cope with not uncommon resource restrictions (I have had phpWiki and Drupal separately hit memory limits and crash).
** PHP 5.3 introduces phar modules -- let's hope that works out.
abhishek.geek.nz
JeevesBond posted this at 22:28 — 3rd July 2009.
He has: 3,895 posts
Joined: Jun 2002
I take your point, but am not sure of the amount by which it's declining.
They're getting cheaper in terms of the cost to rent them, but not in the cost of having to learn to sysadmin your own GNU/Linux server.
There are still plenty of designers and Web code monkeys who can turn out a half-decent, dynamic Web site but who have no idea how to setup their own Web servers. Probably why PHP is getting such a bad reputation: many people of ability have moved to something else. This does not detract from your point on shared hosting, more on that below.
Code monkey's use FTP though! Their testing stategy works in two stages:
Also the overhead of having to learn Ruby Gems, Rake, Capistrano etc. may make life easier in the long-term, but presents a steep learning curve in the short-term. Compared to PHP's 'Upload-and-Go' philosophy, there's a lot to learn.
I take your point about the growth of shared hosting, but following these instructions would be daunting for a newbie compared to, 'Download XAMPP; run it', in order to get a local development environment going.
Interesting. I was working on a Perl Web site that used an equivalent Perl templating language a while ago. Everything works in cycles.
My argument is simply, 'Buh, buh, buh PHP's easier!!11!one' Which I'm sure you've heard before, and while you do make a good case that the others are getting easier, nothing beats the simplicity (and stupidity) of PHP. Whilst I agree the playing field is leveller than before, it is likely that PHP will remain the easiest to pickup and work with; therefore it will remain the language of choice for noobs, code monkeys and those aiming at those markets, for some time to come.
a Padded Cell our articles site!
Abhishek Reddy posted this at 14:28 — 4th July 2009.
He has: 3,348 posts
Joined: Jul 2001
We're pretty much in agreement on those points.
PHP is incumbent and will likely remain for some years. My posts are laced with a normative flavour because I would like that to change. 
abhishek.geek.nz
davecoventry posted this at 09:08 — 30th July 2009.
He has: 114 posts
Joined: Jun 2009
I have a feeling that some programmers, especially in academia, despise PHP. The language enables un-skilled “programmers” to create rather sophisticated, but poorly designed applications. Some academics might view this as a threat. Formally educated programmers are no longer required for moderately sophisticated web applications. A 15 year old can create a program just as well as a PhD, from the user’s perspective (under the hood, it would be a different story). While I don’t have much (or any) evidence to back that up, that’s the vibe I get from reading people criticize the language.
How much is the Python project paying you to pick on PHP, anyway?
Well, I do for one
However, not for the reasons you put forward. Indeed, were a language to be despised for ease of use and readability of code, I'm pretty sure it would be python, which I'm actually pretty fond of.
I've had to learn PHP in order to run a Drupal site and I must admit I'm hating it more and more.
I'm not recommending my hosts anymore until they sort out my hacked sites