# Does changing to\_param() have any side effect?

**URL:** <https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083>\
**Category:** Rails\
**Created:** [May 24, 2010, 10:08am UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083 "2010-05-24T10:08:15Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![winterheat](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@winterheat](https://www.ruby-forum.com/u/winterheat)\
**Post date:** [May 24, 2010, 10:08am UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083/1 "2010-05-24T10:08:15Z")

</div>

In some book, it is recommended that to\_param is changed to

```
class Story < ActiveRecord::Base

  def to_param
    "#{id}-#{name.gsub(/\W/, '-').downcase}"
  end

end

```

so that the URL is

```
http://www.mysite.com/stories/1-css-technique-blog

```

instead of

```
http://www.mysite.com/stories/1

```

so that the URL is more search engine friendly.

So probably to\_param() doesn’t need to be used by other parts of Rails  
that changing it may have any side effect? Or maybe the only purpose is  
to construct a URL for linking?

Another thing is, won’t it require to limit the URL size to be less than  
2k in length – will it choke IE if it is more than 2k or maybe the part  
more than 2k is just ignored by IE and so the URL still works. It might  
be better to be limited to 30 or 40 characters or something that will  
make the URL not exceedingly long.

Also, the `ri` doc of to\_param:

```
       class User < ActiveRecord::Base
         def to_param # overridden
           name
         end
       end

```

if to\_param is changed like that, then the link actually won’t work, as

```
http://www.mysite.com/stories/1-css-technique-blog

```

will work, but

```
http://www.mysite.com/stories/css-technique-blog

```

will not work as the ID is missing. Are there other ways to change the  
to\_param method?

**Update:** on second thought, maybe

```
http://www.mysite.com/stories/css-technique-blog

```

won’t work well if there are many webpages with similar title. but then

```
http://www.mysite.com/user/johnchan

```

will work. Will it be params[:id] being “johnchan”? So then we will  
use

```
user = User.find_by_login_name(params[:id])

```

to get the user. So it just depends on how we use the param on the URL.

```
C:\ror>ri ActiveRecord::Base#to_param
--------------------------------------------

```

ActiveRecord::Base#to\_param  
to\_param()  
------------------------------------------------------------------------  
Returns a String, which Action Pack uses for constructing an  
URL to  
this object. The default implementation returns this record’s  
id as  
a String, or nil if this record’s unsaved.

```
     For example, suppose that you have a User model, and that you

```

have  
a +map.resources :users+ route. Normally, +user\_path+ will  
construct a path with the user object’s ‘id’ in it:

```
       user = User.find_by_name('Phusion')
       user_path(user) # => "/users/1"

     You can override +to_param+ in your model to make +user_path+
     construct a path using the user's name instead of the user's

```

id:

```
       class User < ActiveRecord::Base
         def to_param # overridden
           name
         end
       end

       user = User.find_by_name('Phusion')
       user_path(user) # => "/users/Phusion"
```

---

<div class="post-metadata">

**Author:** ![winterheat](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@winterheat](https://www.ruby-forum.com/u/winterheat)\
**Post date:** [May 24, 2010, 1:42pm UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083/2 "2010-05-24T13:42:26Z")

</div>

Jian,

This was a common practice a while ago but I would recommend against  
changing to\_params. It will work but my experience has been that code  
gets messy as you find fringe cases. I HIGHLY HIGHLY HIGHLY recommend  
you look at norman’s friendly\_id gem. It is extremely easy to setup  
and all you do is drop one line of code in your model and everything  
works like magic.

> **[GitHub - norman/friendly\_id: FriendlyId is the “Swiss Army bulldozer” of...](https://github.com/norman/friendly_id)**
>
> FriendlyId is the “Swiss Army bulldozer” of slugging and permalink plugins for ActiveRecord. It allows you to create pretty URL’s and work with human-friendly strings as if they were numeric ids fo...

Matt

---

<div class="post-metadata">

**Author:** ![winterheat](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@winterheat](https://www.ruby-forum.com/u/winterheat)\
**Post date:** [May 25, 2010, 1:24am UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083/3 "2010-05-25T01:24:58Z")

</div>

On May 24, 2010, at 4:11 PM, Marnen Laibow-Koser wrote:

> Huh?

The only case I come across after changing to\_param is when I pass  
params[:id] (= “123-foo-bar”) into find\_by\_id and forget to call to\_i on  
it. find() by itself will do the right thing, but find\_by\_id won’t.  
Not a big deal though.

-philip

---

<div class="post-metadata">

**Author:** ![winterheat](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@winterheat](https://www.ruby-forum.com/u/winterheat)\
**Post date:** [May 25, 2010, 1:11am UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083/4 "2010-05-25T01:11:45Z")

</div>

M Daubs wrote:

> Jian,
> 
> This was a common practice a while ago but I would recommend against  
> changing to\_params.

Why? I can’t think of a good reason.

> It will work but my experience has been that code  
> gets messy as you find fringe cases.

Huh?

> I HIGHLY HIGHLY HIGHLY recommend  
> you look at norman’s friendly\_id gem. It is extremely easy to setup  
> and all you do is drop one line of code in your model and everything  
> works like magic.
> 
> [GitHub - norman/friendly\_id: FriendlyId is the “Swiss Army bulldozer” of slugging and permalink plugins for ActiveRecord. It allows you to create pretty URL’s and work with human-friendly strings as if they were numeric ids for ActiveRecord models.](http://github.com/norman/friendly_id)

Yeah, there are several gems like that. They look great.

> Matt

## Best,

Marnen Laibow-Koser  
[http://www.marnen.org](http://www.marnen.org)  
removed\_email\_address@domain.invalid

---

<div class="post-metadata">

**Author:** ![winterheat](https://www.ruby-forum.com/letter_avatar_proxy/v4/letter/w/7993a0/32.png) [@winterheat](https://www.ruby-forum.com/u/winterheat)\
**Post date:** [May 25, 2010, 8:42pm UTC](https://www.ruby-forum.com/t/does-changing-to-param-have-any-side-effect/189083/5 "2010-05-25T20:42:20Z")

</div>

I don’t enjoy reinventing functionality when there’s already a good  
gem to get the job done. Changing to\_param will work but my  
preference is to not change it because then I have to modify all of my  
calls to find for every model class involved. And if I’m going to  
change to\_param, you might as well rip out the integer ID as well so  
that you are SEO and human-optimized at the same time. With that  
comes the responsibility to ensure another unique column which might  
be trivial or may be a pain depending on your existing data  
requirements. Nothing here is a deal breaker but the way I see it  
it’s all a lot more work than just using friendly\_id. Less code to  
test and less code to break. That’s just my preference, though.  
Everyone’s different.