Semantic elements and attributes provide a higher level of communication. The semantics of an a href alone does not provide a clear, direct, explicit, and strong semantic for an image long description. An a href is a generic link. It is not an explicit link to a long description. Adding rel="longdesc" would provide explicit semantics. However, this technique is riddled with problems.

Impossible for One Link to Have Two Destinations

Use Cases

No use cases of rel="longdesc" have been presented that are not already satisfied by longdesc. Whereas use cases for longdesc have been presented that are not satisfied by rel="longdesc".

The fact is rel="longdesc" syntax only works on a limited subset of use cases. It is impossible for one link to take a user to two locations. As Julian Reschke has explained,

<a href=* rel=longdesc href=URL><img src=* alt=*></a> only works when the <img> element doesn't already have a parent <a> element, which is something which is used a lot.

rel="longdesc" is not programmatically determinable when an image already has a link which is mapped to go to another page or a larger image. Expecting all <img> elements on the web that need long descriptions not to have a parent <a> element mapped to a different task unrelated to a long description is unrealistic, i.e., a thumbnail <img> that links to a larger .jpg as evidenced by sites such as the National Institute of Dental and Craniofacial Research. Attempting to force dual functionality into single functionality as proposed in the Zero Edit/Obsolete longdesc proposal is unworkable.

For example, if an image already has a link that takes the user to another page (i.e., a logo image that links to a home page, one cartoon in a series of cartoon images that links to the next comic strip page, a thumbnail image that links to a larger version of an image, etc.), it is impractical and most times nonsensical for the destination page to shoehorn in a long description. It would provide a confusing and inferior user experience.

Removing the possibility for direct, dual, programmatic access to both a long description via longdesc and a separate a href on an <img> element would be (a) needless authoring impediment, (b) a step backwards for HTML, and (c) a barrier to providing accessible content. Including longdesc in HTML5 affords authors a way to provide both an a href as well as an accessible programmatically determinable long description.

Forces a Visual Encumbrance

rel="longdesc" fails because a href forces a visual encumbrance on sighted users unless additional code (in the form of CSS or hidden) were added in order to hide the visual indicator. Please refer to the Cascading Style Sheets and hidden or details of why this is a bad idea.

Reinvents the Wheel

rel="longdesc" attempts to reinvent the wheel by duplicating a basic method that has already previously been created. This proposed replacement is new syntax for the same thing, or for a limited subset as It is impossible for one link to have two destinations.

No Examples in the Wild

No examples in the wild of accessible long text alternatives with rel="longdesc" have been presented for any of the use cases. No evidence exists that authors have or will use it for accessible long text alternatives.

No Improvement

No evidence has been presented that rel="longdesc" produces more accessible content for long descriptions than longdesc or that more authors would use it correctly.

No Implementations

No user agent takes advantage of rel="longdesc" semantics. None inform users that a rel="longdesc" link leads to a description page.

No evidence exists that authoring tools will implement rel="longdesc".

In contrast user agents that support longdesc differentiate between longdesc and an a href. Moreover the new spec text for the rendering section will help even more user agents do so.

Lacks Backwards Compatibility

rel="longdesc" is new. It has no legacy or provenance. Whereas, longdesc has legacy content and legacy support. It lacks all backwards compatibility. It is not:

No retrofitting is required for longdesc as it is an existing HTML feature not a new one. For further rationale please consult the backwards compatibility document.

In sum rel="longdesc" is unworkable and not an excuse to drop longdesc from HTML. Its existence does not negate the need for longdesc.