A slug is a part of the URL when you are accessing a resource. Say you have a URL, such as the one below, that displays all of the cars in your system:
http://localhost/cars
When you would want to reference a particular car in your system, you would provide the following URL:
http://localhost/cars/audi-a6/
Notice how the URL is still very logical, and very SEO friendly. In terms of using the slug, that's at your own discretion. The audi-a6 string above may be a unique identifier for a car in your system — let's say that you have a relational database with the following fields:
id
car_name
car_brand
car_unique_identifier
The field car_unique_identifier would then be used to store the values that get displayed in the slug; in the example that I've specified above with an Audi A6 car, this is where your audi-a6 string would live.
You may use it in other ways as well — for instance, if you have a posts controller that functions like a blog. The title for a page might be the slug for it, if it is URL encoded. For our article named "Best ways to make SEO better", you may provide the following URL:
http://localhosts/posts/best-ways-to-make-seo-better
You would then run url_decode() on the slug, and you would obtain the string best ways to make seo better, which you can use in order to find a post via its title.
It doesn't need to stop there — you may decide to have multiple slugs to represent something — let's take a look at how BBC is doing it. I've taken a random article from today, which has the following URL:
http://www.bbc.co.uk/news/world-africa-24506006
This links to an article named: African Union urges ICC to drop cases against leaders. The way that BBC are doing it is that they use the last part of the slug world-africa-24506006, which is 24506006, to identify a unique entry in their system. They then most likely use world-africa to denote the category that a post belongs to (although this may only be an assumption, it's still an educated guess).
Finally, let's imagine the following DB table, for research papers.
id
category
title
You may have an example that works like the one below.
http://localhost/papers
This URL represents all of the research papers currently in the system. You would then be able to access all of the research papers on physics via the following slug:
http://localhost/papers/physics
Our slug is physics, and our database select currently looks something like:
SELECT *
FROM papers
WHERE LOWER(papers.category) = 'physics'
You may then expose the following URL:
http://localhost/papers/physics/gravitation
Now our slug is composed of physics and gravitation. Our query behind the scenes may look something like:
SELECT *
FROM papers
WHERE LOWER(papers.category) = 'physics'
AND LOWER(papers.title) = 'gravitation'
This allows us to uniquely identify an entry in our system.
So we've used slugs repeatedly in order to filter out our information. In the example, when we ran the URL without any slugs:
http://localhost/papers
We wanted to list all of the research papers available. When we ran the URL with the physics slug:
http://localhost/papers/physics
We wanted to list all of the research papers on physics, thus narrowing our results. Finally, when we provided two slugs, we could uniquely identify an entry in our system.
http://localhost/papers/physics/gravitation
Could we have modeled this differently? Of course! Depending on our system's requirements, we can normalise and denormalise our relational tables. We could have had a permalink system in place, so that our posts table looked like this:
id
title
permalink
We might then have had the following entry:
| 20013 | Gravitation | physics-gravitation-breakthrough |
Thus exposing the URL:
http://localhost/papers/physics-gravitation-breakthrough
In the example above, the slug physics-gravitation-breakthrough allows us to uniquely identify a post via:
SELECT *
FROM papers
WHERE papers.permalink = physics-gravitation-breakthrough
