5 ms·
And what I still don't see is anything saying they are taking their performance problems seriously. Still nothing saying they are fixing their documentation. No
by dkhenry 13y ago
And what I still don't see is anything saying they are taking their performance problems seriously. Still nothing saying they are fixing their documentation. Nothing addressing their hodgepodge of legacy code and mixed methods of implementing features. Its like their upset that people have moved on, but instead of providing a better core for people to build products on they keep adding chrome to the fringes of the language hoping to get more people sucked into developing for it only to realize that once they do anything of significant size in PHP they will need to do a full rewrite into a language that can actually scale or turn to a PHP to C++ lexer to actually get a product that doesn't also server as a Data center toaster oven.
- RossM 13y agoOut of interest, what do you find wrong with the documentation? I've always found it pretty comprehensive and relatively easy to navigate.
- Zergy 13y agoI personally found it very incomplete. The comunit makes up for it but quite often there were large gaps exspecially regarding edge cases and details on how functions would behave if they failed. The comments did a really good job of filling in these gaps but the documentation alone was often inadiquate to understand how the code would behave.
- Navarr 13y agoGenerally very little, but there are things like http://php.net/trader_stoch http://php.net/trader_stoch
- polyfractal 13y agoTo be fair, that is an optional extension and not bundled into PHP by default.
- dkhenry 13y agohttp://www.php.net/manual/en/internals2.apiref.php http://www.php.net/manual/en/internals2.apiref.php
- juice13 13y ago5.3 and 5.4 had significant performance increases. 5.5 has integrated opcode cache. IMO, PHP has one of the best documentations of any programming language. Many bad bits have been deprecated in recent versions and there are PSR recommendations which make it easier to write decent interoperable code. PHP is growing up. However, if you don't like it, nobody's forcing you to use it. Just walk away.
- dcc1 13y agoThis ^ I find the people doing the moaning about PHP are the noobs who are regurgitating what they read or heard elsewhere in order to appear smart, Having moved to PHP 10 years ago after C and Java and .NET development I am happy for having changed direction and have made a bootload of money and saved alot of time thanks to PHP over the years.
- verelo 13y agoGreat response, i love the PHP documentation. Writing clean PHP is not hard, its just easy to write bad PHP. In general it is one of the simplest languages to get started with and it'll stick with you a very long time because of its flexibility.
- cwt137 13y agoThe PHP team does take speed seriously and just about every minor release gets a little bit faster, sometimes alot, like 20% faster. I'm not even including stuff like the opcode caches that are built in now like APC or Zend Optimizer+. Lots of times, minor releases also get more efficient with memory usage too. If you are not happy with the speed, PHP is an open source project and you can contribute patches to speed it up. Documentation gets fixed all the time. PHP has been historically known as having one of the best docs and the docs have contributed to its success. As with the source code, the docs are open to contributions if you feel the need to modify or add something. That's funny you say PHP can't scale. Facebook, Yahoo, Etsy and a whole lot of other big PHP shops get more traffic than most of us will ever see in a few life times. So, PHP scales. Even pre-Hiphop Facebook got more traffic than most of us will will ever see. Many times scaling is more about architecture design/implementation than language. App Engine is all about scaling and PHP runs on it. Oh by the way, at scale, every language is a data center toaster oven. PHP people aren't mad that people are leaving. They are happy that it is still the #1 language used for web development. The only valid argument you have is the mixed methods and stuff. It would be cool to fix, but it would mean seriously breaking the API and would cause a lot of confusion. Every minor release breaks the API a little bit and people are unhappy. So, this might never change. I don't see it as a big deal because if you are a great programmer, you know the quirks of your language and they don't throw you off. Also, there is auto-completion and docs for you to look up things if you don't know stuff off the top of your head. It is true that lots of developers have left PHP, but it is a falsehood that everybody is moving away from PHP. If that was the case, PHP wouldn't be running on 75% of the public facing web hosts out there. There would be a business incentive (because they are using another language) to get PHP off their box. They could devote those resources to the new language they are using. The 75% number shows that their is no mass exodus, but definitely evidence not everybody is using PHP. Oh, by the way, PHP is open source so you can help contribute to its continued success if you like. If somehow you have already contributed, don't stop till the work is done.
- dkhenry 13y agoI do contribute to PHP. I have had a few patch submission fixing bugs here and there, but then you run into this http://www.php.net/manual/en/internals2.apiref.php This is PHP's documentation for fixing things internally to it. Which would be fine if the internals made any sence whatsoever, but between the implementation of steams and the abmismal design of zval's there isn't much to logically make sence of. And this isn't just well if you find it lacking its open source so make it better. We have we made better PHP's and named then things like Node.js and Python. Why would I implement a JIT in PHP when I could just use V8 and have an awesome runtime right out of the box ? Imperically PHP isn't nearly as fast or efficient as other languages, that’s not just an opinion it's been shown in multiple benchmarks and tests. I know it can be used for doing useful things and can even be used for large projects. I work on one of those daily, the issue is no one seems to care that as a platform PHP is _not_ a good choice any more, and increasingly the community isn't making a point of steering the platform to where it is a good choice. They just rattle off the same talking points of "Well facebook uses it" and "XX% of websites are using PHP so clearly it is good" You want a good PHP release? Provide documentaiton on the entire platform (http://www.php.net/manual/en/internals2.apiref.php http://www.php.net/manual/en/internals2.apiref.php) Fix the zvals to make thbem memory efficient Speed up the core ( not a opcode cache a real tracer and JIT ) Then add in yield and continuations, because without the other parts no one should care about language niceties, because no one should be using the language.
- jtreminio 13y agoAnother person who knows nothing about PHP, yet still has an opinion on it? Color me shocked.
- j_col 13y ago> And what I still don't see is anything saying they are taking their performance problems seriously. Hmmm, you have heard about PHP-FPM right? And the built-in APC cache? And HipHop? Besides, in my experience performance problems will arise with other parts of your architecture a lot sooner that your application code (e.g. database if you're not clever about load balancing and caching), regardless of what language you are using. Network calls and disc I/O are a lot more expensive than CPU cycles in your app code. > Still nothing saying they are fixing their documentation. Define fixing. It works fine for me as-is. > Nothing addressing their hodgepodge of legacy code and mixed methods of implementing features. Again, define addressing. Deprecating? They're doing that. What else do you suggest, breaking existing popular applications by removing legacy code that you find offensive? Let me know how that works out (see Perl 6, Python 3...). > Its like their upset that people have moved on, Define people. Seriously. > but instead of providing a better core for people to build products on they keep adding chrome to the fringes of the language hoping to get more people sucked into developing for it only to realize that once they do anything of significant size in PHP they will need to do a full rewrite into a language that can actually scale or turn to a PHP to C++ lexer to actually get a product that doesn't also server as a Data center toaster oven. Wow, that needs a comma. I hate these ranty posts attacking open source projects. Have you any idea how many hours of dev time went into PHP 5.5, free of charge? If they work that hard to build something this significant for the community free of charge, the least they deserve is your respect. In the meantime, feel free to go use Scala/Ruby/Node/whatever excites you.
- dkhenry 13y agoI do respect the work they have done to the language, and I have also spent many hours contributing back to PHP and the various projects in the PHP community, however that doesn't change the fact that the _community_ collectively has a hard time dealing with the hard issues that need to be solved. Proof of that is the response you see right here on this forum. Mention how PHP's internals are lacking and get told how I must be a n00b and should just turn on APC. Until PHP starts to take its core defects seriously I don't think it will ever make significant progress as a platform.
- 13y ago
- Udo 13y agoCould you be more specific? What's your performance problem? Where do you think the documentation is bad? In my experience these are two areas where PHP is doing pretty well. I share your reservations about legacy API (the naming scheme and parameter order of many standard PHP functions should have been cleaned up years ago) and I think adding yet another slew of OO sugar coating isn't going to encourage developers to write better PHP apps, quite the contrary. > Its like their upset that people have moved on, but instead of providing a better core for people to build products on they keep adding chrome to the fringes of the language hoping to get more people sucked into developing for it only to realize that once they do anything of significant size in PHP they will need to do a full rewrite into a language that can actually scale or turn to a PHP to C++ lexer to actually get a product that doesn't also server as a Data center toaster oven. Wow. This big run-on sentence about not being able to do "anything significant in PHP" is just not true, and that's me parsing it generously. Big projects can and do run on PHP. It's a valid choice. If you're using PHP for genome sequencing analysis, you're not using the right tool to begin with - but the same applies to pretty much any other web language as well.
- dkhenry 13y agoWhen I say documentation I mean the internals of PHP. Here is their documentation page http://www.php.net/manual/en/internals2.apiref.php http://www.php.net/manual/en/internals2.apiref.php
- Udo 13y agoI must confess my knowledge about PHP internals and the perspectives of being a core runtime developer is pretty much non-existent. All I have to go by are the API and language features exposed to any web app developer - and that paints a different picture. Are you, by chance, a core contributor?
- dkhenry 13y agoI have contributed patches to the PHP core, but I am not a core developer.
- 13y ago
- newbie12 13y agoThe performance differences between all of the major scripting languages -- Ruby, Python, PHP -- are insignificant, unless you are operating at Facebook scale (and Facebook, which still uses PHP, supports the HipHop project). Your web app bottleneck is not going to be PHP. Plus the PHP community is coalescing around Symfony2, which is a best-of-breed framework. Things have never been better in PHP land.