ImagesInRss

MarcCanter wants to be able to put an image of the blog author in a blog's RSS feed.

The RSS spec [specifies an optional <image> element] that can point to an image that describes the channel.

This has been around since [RSS 0.9]. You can see an example of its use on [Slashdot's RSS 0.9 feed]:

<image>
  <title>Slashdot</title> 
  <url>http://images.slashdot.org/topics/topicslashdot.gif</url> 
  <link>http://slashdot.org/</link> 
</image>

However, this doesn't say anything about the blog's author. So here's a proposal: extend the <image> element in a namespace to let people provide multiple images, each of which describes one specific thing. Keep using the existing <image> mechanism for channel images (so they show up in existing aggregators), and use extended <i:image> blocks for other images.

Proposed format

The idea here is to basically take the RSS <image> element spec, redefine it in a namespace, and add some more options. Either that or just add extra namespaced elements to <image>, but I'm not sure if that will work, as you are probably only allowed one <image> per <channel>.

<channel xmlns:i="http://purl.oclc.org/NET/rssimg/">
        [... title, link, description, etc ...]
        <image>
                <!-- this is a 'standard RSS image' - describing the channel -->
                <width>42</width>
                <height>42</height>
                <url>url-of-broadband-mechanics-logo</url>
                <title>broadband mechanics logo</title>
                <link>url-of-marc's-blog</link>
        </image>
        <i:image>
                <!-- this is an 'extended image' - describing something else -->
                <i:describes>author</i:describes>
                <i:width>36</i:width>
                <i:height>36</i:height>
                <i:url>url-of-image-of-marc's-face</i:url>
                <i:title>a picture of marc</i:title>
                <i:link>url-of-marc's-blog</i:link>
        </i:image>
        [... items and stuff ...]
</channel>

Alternate syntax

I like this way of representing it better, and it would be easier to write a parser for, but it's not very RSS-ish, and existing parsers already know how to parse <image> blocks, so this is probably not such a great idea.

<channel xmlns:i="http://purl.oclc.org/NET/rssimg/#compact">
        [... title, link, description, etc ...]
        <i:image i:describes="channel"
                i:width="42" i:height="42"
                i:href="url-of-broadband-mechanics-logo"
                i:title="broadband mechanics logo"
                i:link="url-of-marc's-blog" />
        <i:image i:describes="author"
                i:width="36" i:height="36"
                i:href="url-of-image-of-marc's-face"
                i:title="a picture of marc"
                i:link="url-of-marc's-blog" />
        [... items and stuff ...]
</channel>

Things you can describe

(values of the i:describes element).

• "channel": the logo for the channel (same as RSS 0.9 - see the Slashdot example)

• "author": a picture of the author

See Also

[RSS 1.0 proposed mod_image]