Hacker News Reader

Comment by Udo | Hacker News Reader

    mysite.com/games/test_game

  instead of something like

    mysite.com/games?id=test_game
  
  Which would be the PHP solution?

I fail to see what this has to do with using PHP or not. The entire question has a background untouched by any knowledge about either Node or PHP. The answer, however well and reasonably written, is about Node only. The title implies a discussion about why to prefer Node over PHP which is just not there.

Contrary to the title, the subject here is not PHP at all.

View on HN

Heh, agreed - and to think that the "PHP solution" is the latter automatically means the author doesn't know much about writing web apps in PHP.

One line of mod_rewrite in an .htaccess file to rout everything to a 3 or 4 line controller script is all it takes.

That being said - Node looks cool and I want to build something in it.

rta, he specifically mentions mod_rewrite

I noticed that as well. I think he was implying that the first URL structure isn't possible in PHP, which of course is not true.

Likewise, you can pass and parse URL parameters in an Express app (or any framework for that matter) if so inclined

Or just split the $_SERVER['REQUEST_URI'] by '/' and map the request to pages.

Yes, but that can still be a bit complex. I have used https://github.com/chriso/klein.php prior for doing exactly this, but it was a long time ago, the codebase now seems a bit more advanced, and mature.

url routing is one of those things about PHP frameworks I never really understood until I wrote my own. I think if you work with PHP, it's probably something you kind of should know how to do.

Imho, '/' should be used to map a specific resource, like a file or a collection of objects in the database.

URL arguments with '?' and '&' should specify a secondary "meta" mapping for things like sorting or filtering, that still acts on basically the same resource.

That's my opinion, as well. However, what about paginating things? A different page is a different set of the same type of object, so it's basically sorting and filtering, but you're returning different content.

    /posts/page/2

    /posts?page=2

    /posts?limit=25&offset=25

I like the third option because it allows the must functionality to the user if they so desire, but it kind of feels like a violation of the "only sorting and filtering" rule.

The person asking the question is mistaking get/post variables as a php language construct, instead of a web app construct.

The question seemed to be from a novice web developer.

Many PHP solutions rely on mod_rewrite to re-write URLs from the first to the second. Without those rewrite rules, you would get the second form.

Because PHP doesn't implement the web server itself but instead relies on Apache/Nginx/whatever, it has to use the info the server provides for routing. Once you configure the server to pass this information along, there is really not a lot of difference how routing takes place inside the web app.

Of course, there are a lot of sites that use .htaccess to implement routing rules, but I'm not a huge fan of that. All things being equal, I believe just passing the virtual path to an /index.php endpoint is preferable - and it's pretty much what other languages and frameworks do in these cases as well.

Reply on news.ycombinator.com