Image’s alt attribute is not a (fig)caption

[riddle]

Hello, I understand that HTML5 brings a lot of freshness to our old wooden table, but unfortunately transforming alt attribute in <img> element to <figcaption> is not correct and not just in some abstract, nerdy way.

That alt attribute is supposed to be used as a replacement for the image, in case the image can not be displayed.

Read more about its usage: http://dev.w3.org/html5/spec/Overview.html#alt

Or imagine something like this: you have an article where you describe how good business relationships look like. You add an illustration of two shaking hands. The alt attribute of that illustration isn’t “Two shaking hands” and should not be “Copyright Getty Images 2011” either. It should represent a mental connection between seeing an image and understanding its context inside the article – so “Agreement is a great tool in business relationships” would be much better.

Now, <figure> and <figcaption> are used only when their content is specifically mentioned in the article. So it could be a graph, a listing of code or an image. But it doesn’t mean that every image with a caption automatically is a figure.

Multimarkdown is a great stuff and I appreciate the amount of work you put into this thing. I just want to make sure HTML5 is supported by modern tools like the spec says it should.

[paulirish]

@title attribute i could imagine would make for a better figcaption value, yah?

[fletcher]

I'll think more to give you a more detailed response, but briefly:

true - alt tag is now limited to being a stripped down version of the caption.

@title attribute doesn't really support MMD syntax, whereas the former alt does. Ironically, alt can't contain markup whereas a caption can.

Therefore:

![This will be the *caption* since it can have `markup`](image.png "This could be the alt, which can't have *markup*")

One possibility would be to get rid of the title attribute all together, and simply have alt and a caption, if desired. But basically, we have three pieces of information, and only two slots to put them in.

[fletcher]

PS> Just saw the "thread" on twitter between you and Merlin

It's not that alt became caption.

It's that the caption is also copied to the alt. But I can change that so that the alt comes from the "former title" bit, and title no longer exists.

In other words, my intention is not that alt and caption are the same thing, if that clarifies things a bit.

[fletcher]

I have added more detail about this to the User's Manual, that might further explain what's going on.

Captions are not going away.

I guess the question is --- do we keep the current approach for generating the alt for inline images to be consistent with Markdown, knowing that "figures" will have the caption duplicated to alt? Or do we use the former "title" to instead create an alt attribute, and remove the title attribute all together? I suspect the current approach is better, but I am open to discussion.

[riddle]

I don’t use MultiMarkdown so often to really know how it influences the workflow. I only know that if something is default it will most likely be used over and over. The question is, is Markdown used only by people who can tell the difference.

If you want to create caption, do so. But I think alt cannot be generated or copied in any way. When it’s duplicated, it’s useless to people who will use it – screen reader with read a phrase two times. Alt attribute has to be written by a caring author. Title tooltips tend to upset me because people overuse them.

[fletcher]

Like I mentioned above, one option is to get rid of title altogether (I think it's useless), leaving two pieces of data - alt and caption.

[fletcher]

No further thoughts on whether title should go away, and we should only use alt, and the new figcaption?

[fletcher]

Since this thread seems dead, I'm going to leave things the way they are.