Live data from Hacker News

Ask HN: Please tell us how you started a project in PHP and are sorry for this.

news.ycombinator.com

31–36 of 36 posts

Re: Ask HN: Please tell us how you started a project in PHP and are sorry for this.

#32

You wont be jobless in a few years even if you continue writing PHP. So don't worry. The scare tactics are nothing but marketing usually. When Rails was initially marketed DHH used Java as a scapegoat. Ironically the Java stack traces he posted as proof in the early days of Rails have nothing on the monstrosities that exist in modern Rails apps. I would recommend playing around with some other technology though, not…

Thank you for your encouragement. Yes, you are right, I know JS and I think that time_invested/result coefficient will be highest with Angular and Node.

Re: Ask HN: Please tell us how you started a project in PHP and are sorry for this.

#33
post #26
post #6

Thank you people! My lesson: I will stick with PHP and learn more about organization, architecture and best practices. And Composer and PSR.

You have 10k hours of php under the belt and have not used composer much?

Composer's what, barely 2 years old? You can have 10 years of PHP experience and not have used that specific tool much, yes. It's not the first PHP package manager.

Re: Ask HN: Please tell us how you started a project in PHP and are sorry for this.

#35
I am myself heavily invested in PHP. And yes all "PHP Sucks" articles scare me, and I have seen some of the worst code written in PHP. But I think PHP should be with us for a long time. With modern frameworks like Laravel, we can write great and easily maintainable code.

However, I still have Java as my backup language. I use Java for my side projects. I think there will always be demand for Java because of heavy use in big enterprises.

Re: Ask HN: Please tell us how you started a project in PHP and are sorry for this.

#36
10 years ago I was working at a European telco, they needed a way to monitor traffic flow on customer connections but the tools that existed didn't work for half the population. So, given that I had paid my way through college programming the bosses gave the task of scoping the project to me (a communications engineer).

I looked at the available tools, got a price for a very expensive commercial tool, and got a price from an independent programmer to use rrdtool and build a custom tool. I wrote a paper which set out the options in a paper for management, explicitly saying it should be a fixed price contract with only supervision from our team, no effort.

The Engineering Director came to me and said he had a contact at another company, they already had a solution that would be suitable and they would be paid (by the hour) to adapt it to our needs. A month later they arrived with a Cisco PIX firewall, two HP Xeon servers and a pile of PHP and Perl code.

There was supposed to be two days of training and some integration. It didn't work. They had done nearly nothing and we had two weeks before the system had to be ready. So I knuckled down, rewrote the entire thing virtually from scratch and while learning PHP and Perl (I had exposure before but you could see my evolution of understanding in different parts of the code).

Two weeks later we had an undocumented mess that actually worked. I left shortly afterwards because I realised I might actually have to support this spawn of bile. I think it was two years later that I got a call asking me why it had slowed to a crawl. Given that my mind had blanked everything to save my sanity I couldn't help. But I suggested they look to see if the filesystem was full and to check the database. Apparently after a database and file system clean-up everything worked fine. I don't know how long it remained in service, but I fear it might still be functioning.

I know you wanted to opinions on different methodologies, but this is really just me sharing my scars and explaining why I shouldn't be let near code. ;-)

Post reply on HN