Why PHP?

[Builder]

Why are there so many threads discussing PHP usage? There need to be more discussing why PHP really needs to get replaced sooner rather than later, and more people need to read this: http://eev.ee/blog/2012/04/09/php-a-fractal-of-bad-design/

[mvitkun]

Why are there so many threads discussing PHP usage? There need to be more discussing why PHP really needs to get replaced sooner rather than later, and more people need to read this: http://eev.ee/blog/2012/04/09/php-a-fractal-of-bad-design/

Where did you fine these threads exactly?

[nokturrduk84]

So what can replace it?

[Ssoele]

PHP is indeed not perfect, but there is not really something that can easily replace it.

[Builder]

Where did you fine these threads exactly?

http://linustechtips.com/main/topic/149227-creating-a-proper-json-file-with-php/ http://linustechtips.com/main/topic/149172-building-custom-website-with-php/

 

So what can replace it?

Ruby on Rails. Prettier, more predictable, modern, etc.

 

The thread was a bit incendiary. I don't mean to start a flame war, just some constructive discussion on why PHP is used and why it should be replaced. It's not wrong to ask for help with PHP, but when I see people having PHP-specific problems I feel bad for them.

[Ssoele]

http://linustechtips.com/main/topic/149227-creating-a-proper-json-file-with-php/ http://linustechtips.com/main/topic/149172-building-custom-website-with-php/

 

Ruby on Rails. Prettier, more predictable, modern, etc.

 

The thread was a bit incendiary. I don't mean to start a flame war, just some constructive discussion on why PHP is used and why it should be replaced. It's not wrong to ask for help with PHP, but when I see people having PHP-specific problems I feel bad for them.

 

RoR is a lot less lightweight then PHP.

[LukeTim]

http://linustechtips.com/main/topic/149227-creating-a-proper-json-file-with-php/ http://linustechtips.com/main/topic/149172-building-custom-website-with-php/

 

Ruby on Rails. Prettier, more predictable, modern, etc.

I like Ruby on Rails, it is really easy to use and makes for beautiful code, but god damn is it slow compared to PHP...

[alpenwasser]

Speaking from personal experience, I'm not a professional web dev,

but I have done one large-ish project in PHP (my dad's company

website), and by larg-ish I mean I basically implemented a CMS from

scratch (yeah, it took me a while ).

Before starting that project, I looked at Python, Perl, RoR and PHP,

trying to work out which I should go with. I ended up going with PHP

for the simple reason that it was much easier to find beginner-friendly

information in PHP for the things I wanted to do than in any of the

other languages, and that PHP is indeed pretty convenient when you're

new to the whole web dev thing (note: I was not a beginner programmer,

I had done C/C++ before).

I even came across the article you linked and read it back then. But

the technical aspects of a language are not all that matters IMHO.

As a specific example, take development time. The argument that dev

time could be reduced if the language were designed more consistently

wouldn't really have applied to me as a newcomer. I needed to look

up tons of stuff all the time anyway, and the fact that I could

find that stuff more easily for PHP than the other languages I had

looked at because of the many, many tutorials out there more than

made up for the fact that the language could use some improvements.

In all likelihood, dev time would have risen substantially if I had

chosen one of the other languages simply because it would have taken

me longer to find the solutions/tutorials for my problems.

I realize that that doesn't make the language itself any better, but

in my personal use case, it allowed me to get the job done more easily,

and that's not something which I could just ignore. Hence my choice.

I can imagine that many people have been in a similar situation and

ended up going with PHP not primarily for technical reasons, but

because it was the most convenient solution for them, all things

considered.

I am very aware though that for a professional web dev, this argument

might look very different of course.

Personally, before trying to replace PHP, I'd much prefer people

abandon such evils as Java (currently doing a semester course in

that, and dear lord I hate it) and Javascript. Yeah, I know, not

gonna happen, but a man can dream, can't he?

But, I have recently (beginning of this year) started delving more

into Perl, and I must admit that I really like it, so in the future

that will most likely be my preferred solution, at least for the

time being. Besides, as a UNIX/Linux guy, I should know Perl anyway.

PS: Thank you for not derailing the other thread and making a new

one for this discussion btw.

Also, @MG2R might have an opinion on PHP.

[MG2R]

Also, @MG2R might have an opinion on PHP.

Oh you just had to bring me into this, didn't you?

 

 

First of all, that blog post in the OP is hi-la-rious! Oh man, did

I laugh. Granted, I didn't read the complete thing, but  did come

across something I personally had trouble with only yesterday.

 

I'm not a professional web dev (and I wouldn't want to be), not

even slightly, but I do have some experience in PHP and

programming in general. Enough to understand what the OP

and the writer of that blog post are on about.

 

 

Yes, PHP has its gripes, sure. Do I hate it sometimes? Yes!

Does it do freaking weird stuff sometimes? Sure. Do I need

to scratch my ahead at a problem for an hour sometime?

Certainly! But do I get things done with it? You bet your ass!

 

I could learn something else (and, honestly, I've been wanting

to lear Python for a long time). In fact, I know (or used to know)

something else: Perl. The fact of the matter is, learning some-

thing new takes time. Time is simply something I don't have in

abundance.

 

I should probably note here that I actually like loosely typed

languages. I can't really explain why, I just like them. Maybe it's

because I don't think the language should limit what you can

do. The same thing goes for C. Not because it's loosely typed,

which it isn't, but because it doesn't really limit programmers in

what their program is capable of. A great quote by one of my

teachers that shows this capability is:

 

 

C is a language that's built to do things which it isn't built to do.

 

If that translation from Dutch even makes sense...

 

 

PS: I had a bug yesterday which had to do with the fact

that the ternary operator doesn't really do what you would

expect it to do.

[Prömetheus]

PHP do the job and work perfectly with SQL

And it's kinda easy to learn without any long letters that is hard to remember.

PHP does the job

[leodaniel]

Why are there so many threads discussing PHP usage? There need to be more discussing why PHP really needs to get replaced sooner rather than later, and more people need to read this: http://eev.ee/blog/2012/04/09/php-a-fractal-of-bad-design/

 

I agree with you. Personally I just love Ruby but I m still learning it (and know PHP quiet well...)

 

And for your question: Because most websites use PHP, thats why it's being discussed so much.

[GoodBytes]

If you want an alternative to PHP, you have ASP.NET, but you need a Windows Server, I believe.

[WanderingFool]

If you know Java really well (and haven't learned PHP by that time, which actually happened in my case), then JSP is actually a good solution to PHP.  With that said, PHP is just so easy, and the mix of html and php makes it so much easier in my opinion (Unless you are wanting to do a lot of processing tasks on the server side, then JSP might be better just because if properly written I believe it is faster than PHP...and can do a bit more)

[rashdanml]

Personally worked mostly with PHP and dabbled a bit with Python/Cherrypy. 

All languages have their flaws, and it's also a matter of personal preference. A good number of people might absolutely despise FORTRAN, but someone might turn it into a Web dev-friendly language and actually use it as such*.

 

* Extreme example. 

[alpenwasser]

A good number of people might absolutely despise FORTRAN, but someone might turn it into a Web dev-friendly language and actually use it as such*.

 

* Extreme example.

Now that is something I'd like to see.

[Builder]

-snip-

 

-double snip-

 

Sorry I'm so tardy in my reply, finishing up a term paper.

 

 

 

When the best argument for a tool's use is "I can do something with it," it shows how dire the situation is.

 

Rails to me seems like one of the best approaches to replacing PHP. Node.js and whatever that Haskell compiler/thing for web apps is called (dammit I'm a Haskell guy I should know this) to me are far more preferable solutions.

 

If you can function with PHP's ambiguity, more power to you. I was kind of hoping LTT was a place for radical new opinions on web and backend design, but I suppose not.

 

To me, PHP is kind of like x86. There is a core group of people who are convince that it's the best thing in the world, some "meh" guys, and the totally opposed guys. Personally, I am pro abolition of PHP and x86, but those seem to be very radical opinions here.

 

PS I'm a *nix guy myself, which ATM is OS X. Yes, even hardcore geeks can use Macs.

Edited May 15, 2014 by MG2R

[MG2R]

-snip-

Little request: please do shorten long quotes, it makes the forum WAY cleaner. Thanks!

 

 

Personally, I am pro abolition of [...] x86

Now, here it's getting interesting. Please elaborate.

 

 

I was kind of hoping LTT was a place for radical new opinions on web and backend design, but I suppose not.

Well, you have to look at both sides of the fence. A lot (if not all) in this world is driven by economics. Yes, there may be other things better suited to build server-side logic for websites. Then again, if the whole world uses X, and Y can provide a 5 % better solution, is it worth it to convert the whole world to Y?

[alpenwasser]

Sorry I'm so tardy in my reply, finishing up a term paper.

 

 

NP, kinda familiar with that.

 

When the best argument for a tool's use is "I can do something with it," it shows how dire the situation is.

 

There are two sides of me regarding the issue of PHP's shortcomings:

One side is the pure pragmatist who does not care too much about the

technical side of things: He has a job to get done, he finds a tool

which gets the job done within a reasonable amount of time and that's

that. As long as the end result fulfills its job reasonably will (and

in my case, it does so pretty well actually) and the result can be

achieved within a reasonable timeframe, he'll go with that tool, even

if it's far from perfect.

Basically, he's an economist, as @MG2R has alluded to. He makes a list

of criteria, weighs them according to their importance and then makes

a decision based on that. Technical aspects, while not irrelevant, are

in the end not the most important thing to this guy.

The other side, well, he's a technical purist and would like to write

web apps in ANSI C. He's very annoyed by all the technical impurities

in PHP, the inconsitencies, the strange stuff, and bloody hell he misses

his dearly beloved pointers!

But in the end, when you want to put something into production, what

matters much more is money and time (well, one could argue that they're

the same thing ), and all things considered, that's an area where

PHP's package is very appealing, despite its technical shortcomings.

I think that that's also one of the reasons Java became so popular,

and will continue to be. IMHO, it's an utter crime of technology,

but when you consider the entire ecosystem, it has a certain appeal

to certain people.

 

Rails to me seems like one of the best approaches to replacing PHP. Node.js and whatever that Haskell compiler/thing for web apps is called (dammit I'm a Haskell guy I should know this) to me are far more preferable solutions.

Well, personally I would have loved if we had just stayed with Perl,

but who listens to me...

I've looked at Haskell (and functional languages in general) a bit,

although not in the specific context of using them for web apps.

I really like the concept behind it and at some point (when

I have the time...) I definitely want to delve into that side of

things more, sadly so far that has not been the case.

Node.js is really interesting, though I haven't applied it in

practice yet (as said, I'm not a professional). How much of

Javascript's syntactic horribleness does it carry over (yeah,

I'm not very fond of JS)?

 

If you can function with PHP's ambiguity, more power to you. I was kind of hoping LTT was a place for radical new opinions on web and backend design, but I suppose not.

Well, as said above, I'm not a professional web dev, and neither

is MG2R. I kinda doubt you're going to see cutting edge ideas for

next-gen web apps from a bunch of people who've only done the

occasional site TBH.

I get your argument, I really do, and it's not that I don't care

about the technological shortcomings of PHP, as some of them have

annoyed me to great lengths.

But, as a non-professional web dev, those technical shortcomings

are just not really that significant to me compared to other

criteria. I have a job, I need to get that job done. I'm

currently learning Perl, and I really like it, and yes, I'm

very much intending to make Perl my go-to language for this

kind of stuff, but back then that just was not a realistic

option due to all the things outlined in my previous post

(I should mention that I'm not just learning Perl for web

stuff, but primarily because it's very handy to know Perl

in the UNIX/Linux world in general, as that is very much

where I am at home).

For a professional, these considerations may well look

completely different, and I have absolutely no issue with

that.

But, I am not, and as emphasized, my posts are specifically

written from a non-prof perspective. You asked why we believe

PHP is so popular, and that's a big part of it from what

I can tell: There are many people out there who aren't professional

web developers, but do the occasional site, and from my

personal experience, at least for the time being, PHP is

a pretty attractive option for those kinds of people.

Although, yes, it also still seems to be very popular with

professionals, but I can't speak for them.

For me, personally, there is only really one thing that

would have needed to be different to possibly sway me to

an alternative: Easily accessible tutorials for beginners.

There are just a ton of those available for PHP. If/when

the situation changes in that area, things might change for

people like me pretty quickly though.

 

To me, PHP is kind of like x86. There is a core group of people who are convince that it's the best thing in the world, some "meh" guys, and the totally opposed guys. Personally, I am pro abolition of PHP and x86, but those seem to be very radical opinions here.

I'm not really sure if the abolishment of x86 is such a

radical idea, purely technologically speaking. But, here

too, same argument: The world is ruled by economics, not

technological considerations. It doesn't need to be the

best tool for the job, it just needs to be a tool which

gets the job done for acceptable cost.

Replacing x86 would cost billions if not trillions of

dollars and likely be a process taking decades.

I could see x86 fading out of many people's homes in the

not too distant future simply because they don't have

desktops and normal laptops anymore, but instead tablets,

mobiles and other devices in which ARM (or similar alternatives)

are strong. But for servers and places where you still

need a normal PC, I think x86 is going to say with us for

a while yet.

Since this is a very intersting topic (to me, at least):

What would you propose as x86's replacement?

 

PS I'm a *nix guy myself, which ATM is OS X. Yes, even hardcore geeks can use Macs.

 

Never really used OSX. From what I've seen it seems pretty decent,

although I'm not very fond of closed ecosystems. Also, I need my

tiling window managers, not those normal ones (out of curiosity:

could one run alternative window managers on OSX?).

PS: Terribly sorry for the wall of text, I kinda tend to do that.

[Hazy125]

 

PS: Terribly sorry for the wall of text, I kinda tend to do that.

You provided me with tonight's reading material and only women really care about length anyway

[Lumi]

1: PHP isn't that complicated

2: It works

3: If you know what you're doing then you can make some cool ass shit.

[Builder]

Little request: please do shorten long quotes, it makes the forum WAY cleaner. Thanks!

 

 

Now, here it's getting interesting. Please elaborate.

 

 

Well, you have to look at both sides of the fence. A lot (if not all) in this world is driven by economics. Yes, there may be other things better suited to build server-side logic for websites. Then again, if the whole world uses X, and Y can provide a 5 % better solution, is it worth it to convert the whole world to Y?

Sorry about the long quotes. I tend to leave them long if lurkers jump in on a certain page but I'll cut them from now on. I would say so, because it would save programmers a lot of time and bugs to use something more expressive and less ambiguous like Node.

 

-snip-

I'm going to try and address these one by one.

 

First off, if you understand JS, chances are you will pick up Node.js pretty quickly. I'm not a C purist by any means, I'm polyamorous with programming languages. They all have their niche, it's just some of them fit some niches better than others do. PHP doesn't seem to be the perfect or even close to the best tool for anything, whereas if you "get" JS it is great for building interactivity into web apps and Node extends that capability to server side operations including database manipulation. I think most people would also argue that Ruby is much better than PHP when it comes to how expressive and beautiful the language is.

 

Java is awesome. I'd take it over C or even PHP any day. Quarrels with Oracle as a company aside, Java runs on basically anything you can get a JVM on to, and runs damn fast in general too. That's the magic of Java and when you add in garbage collection I think it's nearly perfect for performance intensive applications that need to be used consistently across several platforms.

 

As for your non prof vs prof comparison, what I would say is that better tools are better for everyone. If you compare Ruby and PHP side by side it will be clear to you what I mean. For example, http://hyperpolyglot.org/scripting

 

On top of being very expressive, you will notice how concise it is. Considering how well implemented Rails is I think it could easily take over for PHP.

 

With regards to x86, there are very few parts of technology that move slowly. Intel has decided to make x86 one of them. The fact that we're using an architecture that has its roots in the 8080 is quite frankly shocking and I really don't think it's appropriate at all. Considering Intel has actually used ARM chips in a product demo, I'd say that's conclusive evidence that x86 and even CISC chips in general are on the way out. When the biggest proponest of the CISC philosophy uses a RISC chip, even for a demo, it shows how desperate their situation is.

 

As for window managers, there's probably a haXie for it. I'm not entirely sure, I use BetterSnapTool to tile windows. Very flexible and easy to customize.

 

PHP is by far not perfect, and in fact it never really intended to be a language, and the developers keep having to come up with different solutions, rather than fix the past solutions (for instance, we went from mysql functions, to mysqli functions now PDO).

 

But it is a brilliant language just to rapidly develop something, and it is some what fast - but yeah, I hope to see a new solution sometime soon as the web is continually developing.

It might have been brilliant at the time, but any CS major will tell you that it is far outclassed by more imperative and concise languages such as Ruby or even JS/Node.js.

 

1: PHP isn't that complicated

2: It works

3: If you know what you're doing then you can make some cool ass shit.

1: Ruby is less complicated, JS more but more powerful as well

2: Plastic bags work as contraceptives but do people use them for that? No.

3: You can make cool-er ass shit in Node and Ruby, and do it faster and maintain it for longer as well.

[alpenwasser]

First off, if you understand JS, chances are you will pick up Node.js pretty quickly.

Cool, I'll need to look into it at some point then. Can't really

say I'm fond of JS itself (the syntax is just horrible IMHO, but

then again, that's more a matter of taste than anything else),

but I've been hearing people reave about it (Node.js), so there

should be something good about it.

 

I'm not a C purist by any means, I'm polyamorous with programming languages.

Hehe, yeah as said, the technical purist side of me isn't really

the boss. I love C, but I also love Perl, and I definitely want

to look into proper functional programming when I finally get the

time for that. C will probably always hold a special place in my

heart though.

 

Java is awesome. I'd take it over C or even PHP any day. Quarrels with Oracle as a company aside, Java runs on basically anything you can get a JVM on to, and runs damn fast in general too. That's the magic of Java and when you add in garbage collection I think it's nearly perfect for performance intensive applications that need to be used consistently across several platforms.

Hehe, I just recently had this discussion with somebody in another

thread. I'm currently doing a semester course in Java, and dear

lord I hate it. But, as you say, there are niches to fill and tools

to fill those niches, Java seems to be one such tool.

 

As for your non prof vs prof comparison, what I would say is that better tools are better for everyone. If you compare Ruby and PHP side by side it will be clear to you what I mean. For example, http://hyperpolyglot.org/scripting

 

Yes, better tools are better, but that's kind of my point: My

decision to use PHP back then was not primarily about the tool,

it was about its ecosystem and project economics, specifically

how easily I could find beginner-friendly tutorials. I was well

aware that there might be better tools out there, but I just

could not find beginner-friendly tutorials for all the stuff I

needed. The tool's quality was but one factor among many in my

decision-making process. By now though, the situation might

actually have improved, don't know.

Since this specific point is possibly irrelevant for professionals,

their decision-making process could easily look different I could

imagine.

Also, now that I have some experience, the information I need

to pick up a language can be less beginner-friendly and I can

still find my way around, as I have discovered when delving into

Perl.

Very interesting link, will make note of that for future reference,

thanks!

 

With regards to x86, there are very few parts of technology that move slowly. Intel has decided to make x86 one of them. The fact that we're using an architecture that has its roots in the 8080 is quite frankly shocking and I really don't think it's appropriate at all. Considering Intel has actually used ARM chips in a product demo, I'd say that's conclusive evidence that x86 and even CISC chips in general are on the way out. When the biggest proponest of the CISC philosophy uses a RISC chip, even for a demo, it shows how desperate their situation is.

 

Yeah, x86's core is ancient, I wouldn't mind seeing something better

come along and replace it. As said, I'd estimate it'll take quite

a while though.

 

As for window managers, there's probably a haXie for it. I'm not entirely sure, I use BetterSnapTool to tile windows. Very flexible and easy to customize.

 

Cool. I must admit that I've grown very attached to my tiler (i3,

specifically). Every time I need to use Windoze I get annoyed by

manually needing to arrange my windows.

[tbc]

let me start things off....trying to replace JS is an unbelievably wasteful endeavor.  JS is ALREADY a compile target for c++ (emscripten), dart, go(not entirely sure bit pretty sure go is what i think it is), typescript and probably a whole bunch I don't know about.  this is aside from the fact that es6 and es7 are making some pretty big differences.

 

language replacement in a browser setting is absurd.  language evolution is the only way to go.

 

 

that said, PHP is used by startup companies because it is cheap compared to asp.net/IIS and the required licenses.  As well, it's easy to teach and errors in the language can be overcome by the million monkeys + million typewriters strategy.

 

larger companies use PHP because they want to hire experienced devs (and a large portion of those devs will come from startups using PHP).

 

make sense?  it's the exact same thing microsoft does with windows.  if people use windows at home, they will demand their employers/employees use windows at work.

 

 

this is why there is 'bizspark' by microsoft; it's the counter to PHP's cost advantage

[Builder]

Cool, I'll need to look into it at some point then. Can't really

say I'm fond of JS itself (the syntax is just horrible IMHO, but

then again, that's more a matter of taste than anything else),

but I've been hearing people reave about it (Node.js), so there

should be something good about it.

 

Hehe, yeah as said, the technical purist side of me isn't really

the boss. I love C, but I also love Perl, and I definitely want

to look into proper functional programming when I finally get the

time for that. C will probably always hold a special place in my

heart though.

 

Hehe, I just recently had this discussion with somebody in another

thread. I'm currently doing a semester course in Java, and dear

lord I hate it. But, as you say, there are niches to fill and tools

to fill those niches, Java seems to be one such tool.

 

 

Yes, better tools are better, but that's kind of my point: My

decision to use PHP back then was not primarily about the tool,

it was about its ecosystem and project economics, specifically

how easily I could find beginner-friendly tutorials. I was well

aware that there might be better tools out there, but I just

could not find beginner-friendly tutorials for all the stuff I

needed. The tool's quality was but one factor among many in my

decision-making process. By now though, the situation might

actually have improved, don't know.

Since this specific point is possibly irrelevant for professionals,

their decision-making process could easily look different I could

imagine.

Also, now that I have some experience, the information I need

to pick up a language can be less beginner-friendly and I can

still find my way around, as I have discovered when delving into

Perl.

Very interesting link, will make note of that for future reference,

thanks!

 

 

Yeah, x86's core is ancient, I wouldn't mind seeing something better

come along and replace it. As said, I'd estimate it'll take quite

a while though.

 

 

Cool. I must admit that I've grown very attached to my tiler (i3,

specifically). Every time I need to use Windoze I get annoyed by

manually needing to arrange my windows.

Nothing makes me angrier than buying three new wireless cards to try and replace my first one and then realizing I only had to reinstall the drivers for the first one. Windows Update appears to have bricked all the external device drivers. Thanks, Microsoft.

[MG2R]

I hope this won't drive this thread too far off topic, but I really need to

respond to the whole x86 thing you were mentioning. (sorry it took this

long, I kinda forgot I still had to reply)

 

RISC vs. CISC is an argument that's been wildly debated in the past,

and probably will continue to drive a debate in the future. It is a subject

similar to AMD vs. Intel, AMD vs. nVidia, and other such comparisons

where there are evangelists of each group. In this post, I'll try to keep

my opinion on the matter as neutral as possible.

 

You argue that x86, and CISC architecture in general, is something

from the past and on it's way out. You base this argument on the

statement that Intel has used an ARM processor for something and

that they're supposed to be a CISC proponent. I would argue that

those two thing aren't mutually exclusive. Depending on what they

were demoing, I can see lots of valid reasons to demo it on an ARM

chip.

 

I've also read allegations of Intel making the x86 platform a slow-

moving one. I can assure you that the x86 platform in itself isn't

really slowed by Intel. Rather, R&D money has been shifted to

making their chips less energy consuming and increasing their yield.

 

Also, I don't see how it's bad to be using a processor that has it's

roots in a product from the 70's. The loom that's used to produce fabric

for the shiny clothes you are (or aren't?) wearing right now has it's

roots in the late 1700's. How old can it be? Yet, it's still getting used

today. Why? Because it works reliably and, by extension, is economi-

cally viable.

 

Intel is getting closer and closer to getting the x86 platform in a power

envelope that used be reserved for ARM processors, without sacrificing

performance. Now, if you can get a CISC architecture, that can execute

the instructions from it's RISC opponent in the same time or faster, all

the while using the same amount of or less power, I don't see why we

would need to replace x86, or CISC in general.

 

Right now, both have their merits. The gap, however, is closing fast. By

making statements like "x86 needs to be replaced because its roots are

in the 8080", I'm getting the feeling that you have no feeling for the

economic side of the world. Now, please be aware that I don't mean that

as an insult. I don't know your background, but I do think that you could

do good by thinking with an economic mindset a little more often. It

could help you understand some decisions that get made by companies

better.

 

PS: we're at a point in time right now where IC designers are having

trouble with designing more transistors into CPU's at a rate high enough

to keep up with Moore's law. Starting over completely just to get around

the old roots of x86 would thus be something that is not even slightly

economically viable.