Tag Archives: URL encoding

Why is not a good idea to pass data via parameters via URL / GET

Several motivations discourage to use GET to pass information as parameters via Url.

The first is security reasons. Proxies keep results saved and could keep confidential info, and even serve to another user. Browsers keep history as well and cache.

The second is practical: limitations on the amount of bytes that could be sent via GET. Limits of servers use to be small.

The third is funny, because a single # character could break the thing. If you send to the server a request like: http://blog.carlesmateo.com/#2013/02/22/cloud-must/ the browser will send only to the server the http://blog.carlesmateo.com/ part. The #2013/02/22/cloud-must/ is considered to be information that is the browser that has to process.

So imagine a contact form that send by Javascript/JQuery by GET where the user adds #, for example: 7th street #4, to refer to number. The data sent to the server will be a mess.

Even if you encode the character # as %23, that will be sent by the browser to the server correctly and understood by the server as #, in the long run many problems will come from this.

And imagine problems you could have with other characters: + decoded to space. % decoded when the user was trying to send a common percentage, and in unicode.

There are many more reasons to avoid sending data through GET calls. Use POST instead.