Yahoo! switching to PHP for backend scripting

[pippo]

I'm sorry if I'm out of the scope of this thread,
maybe not at all, but they use freeBSD as OS.
I didn't see any notice from them of switching to another OS,
so every considerations have to keep in mind that.



pippo

[firepages]

So, before you come in flaming me for whatever jynx you think I have against PHP RTFM.

In light of your comments I do not think that was even a light toasting yet alone a flame.

[vudu]

First of all, I want to say nobody can talk for Yahoo!, except Y! themselves. And I have no crystal ball, like almost anybody here.

My opinion about all this : this is good news for opensource *and* php. Another "big" company adopting php will give opensource and php more exposure, and show to those who are still in doubt that they can be used in some enterprises. I do hope Y! will give back and help in any ways.

A subscriber wrote that they were picking a technology that didn't fit their needs. Even if that was true, I don't think that would be a major problem, because we're talking about opensource software. That would definitely be a problem with a commercial solution.

They made a good choice adopting an opensource solution with a huge number of users because they'll benefit among others contributed code (even if they don't use it), a better support, and overall a better solution "out of the box" (if I can say) as compared to one that was developed in-house.

Basically, I think this is good news for opensource and php.

[Jeremy W.]

Originally posted by firepages
In light of your comments I do not think that was even a light toasting yet alone a flame.

*L* Again, in light of what comments? If you are going to toast/flame me, then toast/flame the guy who wrote the presentation.

As far as OpenSource somehow being better than commercial I'd love to know how. The cost of deploying a project of this size is massive no matter what the technology, and the relative cost of using a commercial solution wouldn't have been much higher.

My point was that if they had been smart and produced this document "back in the day" they could have saved hundreds of millions of dollars becuase their proprietary script would have perfectly suited their needs. Proprietary in house solutions are always better than a generic solution because it fits the company's needs perfectly.

[Richi]

do think they will developing the php ?

and if yes, can they keep it in-house as YSP/2 ?

just one 1 question on that site (Y! eng. man site) , i think that i read that they have 612+ develoer .. is that right ... maybe they need it for 4500+ server !

agian 4500+ ... who can say more than WAAAW

[firepages]

"They really need to get their act together and pick something that actually does satisfy all of their operational requirements. PHP doesn't."

one assumes they are not total idiots ?
one assumes that any deficiencies that they may expect to find would be addressed beforehand, can we get over this/that/whatever obstacle.(they have been using PHP for some time in various departments its not like they read one of HarryF's posts and thought 'ohh this looks good lets drop everything'(OF course they should listen to HarryF anyway!!))

And with PHP the answer IME is normally yes, of course I have never worked on that grand a scale, but at worst you write it in C++ & compile it in ! <<EOF

[Jeremy W.]

Who said they were total idiots? They made a choice and will have to live with it.

Does PHP fulfill all their operational requirements? No. One of their operational requirements was complete separation of design and content. Later on, it talks specifically about the deficiences of PHP and lists this one as one that will give them headaches in the future.

So, no, they aren't total idiots, no the issues haven't been dealt with beforehand and no, I don't know why you keep harping on me for saying exactly what Yahoo is saying.

[firepages]

Your first post implied (at least to me) that they are idiots Jeremy W.

......

Separation of design and content in PHP is easily acheived in several ways from template systems , XSLT to full custom-tag like schemes (ugh), whether or not its a good idea depends on the scope of your application and the abilities of your coders and your general philosophy about such things, but you do, as a php developer have the option to skin the cat whichever way best suits your needs.

Further investigation reveals a host of options and core functionality that give PHP various cat-skinning options,if you have done more than a casual investigation of PHP then you could appreciate this.


(No cats were hurt during the typing of this thread)

[Gryff]

Jeremy

- They need to pick something which meets all their objectives, or don't switch. PHP doesn't


Please - where doesnt it meet their objectives, quote from the slide, or another source.

Secondly i didnt see them mention mysql, i thought postgres would have been a better choice.

and ALSO, with their budget i think they'll be using the ZEND IDE + accelerator.

[Gryff]

oh, and also jeremy, if an open source library doesnt fulfill their needs, they can write a PHP extension module/library that does fulfill their needs - in house solution eh?

PHP can probably do 95% of what they want, without them having to develop their own scripting language and maintain/bugfix/create an IDE.

What they CANT do with PHP they can either get an open source library at no cost, or code an extension themselves, I think thats a much better solution then using a costly in house language.

[neil100]

Originally posted by Gryff
Please - where doesnt it meet their objectives, quote from the slide, or another source.

From the slide titled Language Criteria :

"1. C/C++ extensions
2. loops, conditionals
3. complex data-types
4. pleasant syntax
5. runs on FreeBSD
6. high performance
7. robust, sand-boxed
8. interpreted (or dynamically compiled)
9. low training costs
10. i18n support
11. clean separation of presentation/content/appsemantics
12. doesn’t require CS degree to use"

Notice point 11.

From there slide titled Coding PHP Takes Discipline :

"mixed app/presentation problematic – PHP code and HTML forever intertwined"

Clearly PHP doesn't fufill language requirement 11.

[Owen]

...and what other language comes closer to fufilling their requirements?

Owen

[Edit: dang I can't spell]

[Gryff]

I see you havnt heard of the include or require functions then

[M. Johansson]

Hmmm.... It seems to me that they really should have went with java instead, since it fulfills their needs much better. They basically just listed one con against java:

"But you can’t really use Java w/o threads and threads support on FreeBSD is not great."

Seems like they simply didn't choose it because FreeBSD:s support for threading is bad.

As for why Java would have been a better choice:

PHP main strength and simultaneously main drawback is that it’s very non-strict. It’s loosely typed, it doesn’t force OO and it offers no protection of class members. This is bloody awesome of small projects and beginners. BUT it also makes it a bit difficult to use in a situation where there are lots of code (and coders) involved. It’s not impossible, but for it to work, you have to develop an extra set of guidelines for all coders to learn and use, so that all code follows the same model. It’s not that PHP can’t do it or anything – it’s just that Java seems more suited for the job, in my opinion.

Java also has lots of very advanced IDEs – Good IDEs for PHP is pretty rare in comparison.

[M. Johansson]

Originally posted by Owen
...and what other language comes closer to fufilling their requirements?

Java.

[neil100]

Originally posted by Gryff
I see you havnt heard of the include or require functions then

Yes I have and am sure the guys wrote that presentation have aswell, but includes and requires don't enable you to achieve the same level of seperation as you get with some other languages such as asp.net.

[Owen]

It doesn't meet #5 on their list (though I agree that it may be a little silly.) Plus I think it doesn't quite meet #9 or #11. I guess it's a little harder to learn and use than PHP. Still, for a project the size of Yahoo, I will say that Java makes sense.

Owen

[firepages]

"Separation of design and content in PHP is easily acheived in several ways from template systems , XSLT to full custom-tag like schemes (ugh), whether or not its a good idea depends on the scope of your application and the abilities of your coders and your general philosophy about such things, but you do, as a php developer have the option to skin the cat whichever way best suits your needs."


just thought I would post that again in case it was not visible in the post above in some browsers

"PHP code and HTML forever intertwined" - I know that's their quote but its still incorrect.

And yes JAVA is the main competitor and I am pleasantly surprised that it appears not to have won the day here, thats where I would have placed my bets.

[pas]

Why people would want to put all sorts of HTML in their code and/or reinvent caching, security, et al is beyond me.

[pootergeist]

Java?

So with 4,500 servers.
Your J2EE app server will set you back aroun $85,000.
You .not server runs about $36,000.

The logic is pretty darn simple if you ask me.

I'd suggest that we just leave all the business mechanics up to the people at Yahoo (who've probably spent more than an hour thinking about this).
If their move to a PHP based internet system fails, you can start arguing in this thread.

Until then I would suggest you pay a little respect to the decision makers of the world biggest internet site and remain mute.

[johnn]

It's good news. I think a lot of biz will switch to PHP in the next several years and more jobs for PHP developers. BTW, why don't they use Redhat Linux instead of freeBSD?

[Phil.Roberts]

Originally posted by johnn
BTW, why don't they use Redhat Linux instead of freeBSD?

Because FreeBSD has a proven track record of being the most stable under high traffic and heavy loading perhaps? Check netcraft.com....

[Technosailor]

Originally posted by pootergeist
Java?

So with 4,500 servers.
Your J2EE app server will set you back aroun $85,000.
You .not server runs about $36,000.

The logic is pretty darn simple if you ask me.

I'd suggest that we just leave all the business mechanics up to the people at Yahoo (who've probably spent more than an hour thinking about this).
If their move to a PHP based internet system fails, you can start arguing in this thread.

Until then I would suggest you pay a little respect to the decision makers of the world biggest internet site and remain mute.

I totlaly agree.

Jeremy and the rest of you. Chill.

Mattias, yeah. MySQL does seem like a stupid idea for a databse backend....

Aaron

[Jeremy W.]

Sketch,

I'm not excited. Not once have I bashed PHP or Yahoo's decision. All I have done is reply based on what they themselves have said in their presentation. Really as simple as that.

As far as the "massive" figures of 36K and 85K for the server software, I think it's funny considering that Yahoo themselves said they have lost "hundreds of millions of dollars" already with their open source projects mingled with proprietary.

At the end of the day it isn't about open source versus closed source or whatever PHP is "capable" of doing various things. I'm not arguing any of that. I am simply saying they had certain objectives. If they had either realised these objectives anytime in the last 10 years, or if before they chose PHP they had realised that it doesn't fulfill all their objectives they would have saved money on development instead of spending it.

Someone above said MS technologies are just too expensive. Sorry, but even after servers, SQL Server, workstations, extensions, etc, we will save the companies I work for more money then they will spend on us. In fact, we will save them 2-3 times what they spent on us this year.

Fact is, that that would be possible with PHP as well. Heck, it'd be possible using Fortran. It isn't the language, it is the needs as matched up with the language. For Yahoo, this was a decision they consciously made. They said "PHP can't do everything we need it to, but we think it will be okay". It is obvious that they will commit to this "for a time", but it should also be obvious that, as in any corp, over time the amount they are willing to "take it on the chin" will slowly decrease and they will switch technologies.

All of this is in their presentation and, again, none of what I said is against PHP. To think PHP can do everything and be the best at everything is foolhardy. Not even C++ or Java, arguably 2 of the best languages ever written, can do that.

I'll say it again, as it appears I've had to over and over: PHP is a great language. If people choose to use it that's fine with me, heck I use it every week and have likely made more money using PHP than half the people in the PHP section. I choose the best tool for the job, and sometimes that's PHP. Everytime you choose a tool it's a judgement call. Sometimes I've been wrong using PHP. Sometimes I've been wrong using .NET. Sometimes I've been wrong using Delphi. It's the name of the game. You make a call and live with it, which is all Yahoo's doing.

J

[weirdbeardmt]

Originally posted by Sketch
I totlaly agree.

I don't That's what these forums are for. Intelligent discussion of hot web-related topics. This is blatantly quite a hot topic and it is good to receive other people's opinions. The fact is Y! have made their decision and probably don't give a **** what we say about it, doesn't mean we can't dicuss it. If it falls on it **** in 2 years time, then well, hey... it'll give us something else to talk about.

But let's remember that what everyone says here is an opinion, so debate it in a mature and orderly fashion; if you're incapable of that, THEN you should stay mute.


/go PHP! PHP RULZ!