I dislike using lorem ipsum or other non-sensical text for mockups, but I admit that it fills a very real need for designing content-centric websites when the content isn't ready yet.

I've encountered just too many cases where the design looks great as a mock-up but when you start plugging in real data the layout start breaking apart, either because the content is too short, too long, or the designer didn't consider long usernames.

My dream for a filler content generator would be a large database of real content types, so you would have a list of real twitter handles, facebook names, blog post titles, world news articles and so on. You could then do stuff like h1 class="fixie worldnews title" or class="fixie twitter username".

Designers often argue that the purpose of Lorem Ipsum is to help the client focus on the look and not the content. I think this posture is wrong because you need the client thinking about content as early as posible.

Front-end developers also need the randomness of real world data to test the sturdiness of their layouts and account for edge cases as early as posible.

If we can get Samuel L. Ipsum integrated, I'm in. Either way, I think this is great if for no other reason than the amount of time it will save.

Would you object if I tried to implement this? Seems like a fun weekend project.

No objection at all, I'd be willing to design the landing page if you come around to implementing this :)

Agreed, lorem ipsum can become a crutch - good copy is integral for good design. That said, this script will give you random length paragraphs, articles, and what not, giving you some room to test your layout, and makes it easy to throw in temporary content until you've written something yourself.

awesome. maybe an alternative api would be to allow for something like

  fixie.init([".array", "#of .selectors", ".that", ".should", "#be .populated .with .lorem"])

or just a string like

  fixie.init(".string, #of .comma .separated .selectors, .that, .should, #be .populated .with .lorem")

I think that would address the class proliferation issue and still be explicit about which dom nodes should be populated. easier to maintain / comment out.

Nice! I added it as an issue just a second after you.

I'll definitely give it a try, though I invite anyone else to send a pull request.

This is fully implemented now. Woohoo :)

This would be so much more awesome if you could assign the class to a container div. Then on each page reload it would randomly add content - be it paragraphs, headers, imgs, lists, tables etc.

This would be a fantastic tool for testing templates against a wide variety of random content. When making a template for clients to use you can never be too sure what they are going to add so you do your best to style everything. Getting random content on each page load would really solidify this testing.

Adding a class to individual elements... I don't find this too useful.

You can now assign the class to a container div - you must still provide some empty tags to specify what kind of content you want.

For example, if you need a couple paragraphs of text, you can use `<article class="fixie"></article>`. If you need exactly 1 header followed by 3 paragraphs, you can use ``` <article class="fixie"> <h1></h1> <p></p> <p></p> <p></p> </article> ```

Depending on what browsers you're supports it's probably far easier to use generated content (:before, :after). Maybe combined with calc or CSS expressions.

Sure, but it's hard to get a variety of content in that way. If you have a paragraph that begins with lorem ipsum over and over again, you'll quickly notice.

Then add some :nth-of-type() wizardry to it.

Good point.

While there are some situations where that makes more sense, doing that for things like paragraphs is kind of silly - you're just moving junk copy to your CSS. You either have a giant CSS file, or some rather repetitive paragraphs.

I like how the words generated are totally hipster!

Thanks! It's actually kind of funny how small the dictionary is...

Great idea!

I'd use this on client projects if:

- I could pass a selector rather than adjusting my markup (maybe I can add 'fixie' via jQuery?)

- There's an easy way to swap out the word list - I've had Bacon Ipsum, Hipster Ipsum, and Gangster Ipsum all slip through before and it's embarrassing (depending on the client).

I don't this is a good idea. Because even if you want quick design but later you will have to fix them all. And nowadays designers are familiar with html ( the most basic skill ). But it's nice btw

The point isn't to write content for you - it's to automatically add temporary content so you can test out your design before writing copy.

Thought this might be something like http://jsfiddle.net/MNFV8/1/ but I like the addition of the paragraph generators too.

Very cool. However, I don't particularly like the fact that I need to add a class to each applicable container.

Maybe a configuration file that allows for selection of specific classes/ids instead?

Now you can use a specific class instead. Just call fixie(myclassname);

This has been replaced by `fixie.init(["CSS-selector"]);`

This looks pretty cool. It would be a great companion to http://placehold.it (which fills a similar role for images).

It uses placehold.it, actually: http://ryhan.me/fixie/sample.html

I can't remember the number of times I've wished for something like this. I'll be sure to check it out for the next web project :)

Like it. The only nuisance is the class attribute you have to repeat across all tags that you need to fill in.

On a side note, you can now change the class name to whatever you want (thanks Luke Holder via github).

this is fixed: https://github.com/rthprog/fixie/pull/18

What is the better alternative?

don't know. But the smart folk behind should be able to figure something out.

Just auto-filling all <p>'s and <h_>'s might be an interesting option. You could also have the API allow you to list tag names to be auto-filled. It can even be smart and first check whether the tag in question already contains text content, and if it doesn't, only then auto-fill it.

Nice suggestion. I made some modifications so this is the new behavior. Child elements are targeted, and content is never overwritten.

``` <div class="fixie"> <p></p> <p>Hello <a></a></p> </div> ``` The above will cause only the first `<p>` tag to be filled, and the `<a>` tag.

In addition, you can now use `fixie.init(["CSS-selector"]);` to target arbitrary tag names, class names, and ids. If the tag in question contains text content, it will not auto-fill it (though it will auto-fill children, if need be).

I considered doing that, but the problem is that I've noticed that some people use <p> tags as breaks between paragraphs. Requiring a class makes sure that the developer/designer knows where their filler content is going.

If you're going to require classnames anyway, it might be more flexible to use those to define the type of content to insert. Personally I'd prefer adding a simple bit of js to a page to define the fixie behaviour, something like :

  fixie.fill('.my_header_class','head');
  fixie.fill('#my_id','paragraphs');

etc, as otherwise I have to have a special layout with lots of 'fixie' everywhere and perhaps a special structure using certain tags in order to use the js, and can't define what sort of content goes where except by changing the type of html element which contains it.

I'd prefer to start with an html document which has already been made, and consider how to add fixie to it in the least obtrusive way - and how to remove it later when it was not required, without impacting the document structure or classes etc at all. So if I can add a bit of js which adds fixie in an unobtrusive way to any existing page, and then remove it by simply commenting out that js, that would be preferable to having to set up a special template just to use fixie, or having to consider fixie use from the start. Fixie should really IMHO adapt to the html it is used on, not force that html to be adapted to it.

As mentioned elsewhere, this is now implemented as `fixie.init(["CSS-selectors"]);`

yep, we added fixie.init(query); functionality. coming up soon.

Won't work with javascript disabled, obviously. This makes it unusable for me.