audio-info-issues

Audio info issues

This page defines the current issues with the hAudio draft specification. Any issues that have not yet been resolved are marked with an open issue! tag.

Contributors

In order of contribution:

• Brian Suda
• ManuSporny - Digital Bazaar, Inc. (Bitmunk)

Problem: Redundant Property Names

The audio info proposal has invented several new properties. These should be condensed, removed and/or discussed further.

image-summary Property

hAudio ISSUE #1: open issue!

image-summary - This is a small graphic image that summarizes the audio piece. It should NOT be a MUST to use the <img> element, (it shouldn't be a MUST anyway) ~BrianSuda. There are several other formats that already use the PHOTO property, including hCard and hReview. This has potential to collapse and remove image-summary.

Possible Solutions

1. Use PHOTO instead.

audio-title Property

hAudio ISSUE #2: open issue!

audio-title - There is an on going debate where this should be something else such as FN or SUMMARY.

Possible Solutions

1. Use FN instead
2. Use SUMMARY instead

published-date Property

hAudio ISSUE #3: open issue!

published-date is a duplicate of the PUBLISHED property found-in hAtom. This can be collapsed to a single property name.

• • or dtstart, from hCalendar (shouldn't the published date be an hCalendar event?) Andy Mabbett

Possible Solutions

1. Use PUBLISHED instead
2. Use DTSTART instead

Problem: Display properties of rel-patterns

hAudio ISSUE #4: open issue!

In order to solve the graphic-only buttons in rel-patterns problem, it was identified that part of the problem dealt with attempting to create relationship quadruples, or incorrect relationship triples. Typically, rel-patterns are created by doing something like the following:

<a rel="RELATIONSHIP" href="URL">TARGET</a>

The final triple that is created can be read thusly: URL is related to TARGET via RELATIONSHIP.

<a rel="brother" href="/people/john_f_kennedy.html">Robert Kennedy</a>

The URL above would be read like so: The URL http://www.people.org/john_f_kennedy.html is related to Robert Kennedy via a relationship of type brother. In other words, John F. Kennedy is Robert Kennedy's brother.

The above rel-pattern is fine, but becomes problematic when images start to be used, thus:

<a rel="brother" href="/people/john_f_kennedy.html" title="Robert Kennedy">
  <img src="/images/robert_kennedy.gif" alt="Robert Kennedy" />
</a>

It isn't defined if the displayable value of the relationship "Robert Kennedy" should be used from

• the @alt attribute on the IMG element
• the @title attribute on the anchor (A element)
• or, not at all.

Possible Solutions

1. Do nothing. The decision should be left up to uF implementors. The market should decide how the information is displayed.
2. Display whatever a text-only browser would display (in this case, the value in @alt would be displayed).
3. Create a rel-value, or other displayable Microformat class that should specify the text value to be used.

Problem: Peeking into child elements to display rel-patterns

hAudio ISSUE #5: open issue!

It has been proposed that peeking into child elements that are not labeled with Microformatted class names goes against the established publishing practices of Microformats. In other words, if there isn't a Microformatted class name on the element, we shouldn't be reading/interpreting the contents of the element.

There have been several proposed solutions, one of which automatically peeks at certain sub-elements, such as @alt in IMG:

<a rel="license" href="http://creativecommons.org/licenses/by/3.0/">
  <img src="/images/deed/by.png" alt="Creative Commons Attribution 3.0" />
</a>

The Microformat parser would take the HTML above and automatically parse out the contents of the @alt attribute.

Another approach would parse out display-able elements based on a pre-defined Microformatted relationship such as displayble, like so:

<a rel="license" href="http://creativecommons.org/licenses/by/3.0/">
  <img rel="displayable" src="/images/deed/by.png" alt="Creative Commons Attribution 3.0" />
</a>

Possible Solutions

1. Microformats should never peek into child elements when displaying information.
2. The community should decide on a set of attributes and elements that are peek-able (for example: any element that contains an @alt attribute should be peek-able)
3. Only elements that are marked with a class of a specified type, such as displayable, should be used to display the text.

Problem: Summary Property is Missing

hAudio ISSUE #6: open issue!

Martin McEvoy did another analysis run of the audio-info-examples data and found that audio descriptions existed for many of the URLs. It has been proposed that audio descriptions should be added to the hAudio Microformat.

Possible Solutions

• Add audio descriptions to the hAudio specification using SUMMARY
• Add audio descriptions to the hAudio specification using DESCRIPTION

Historical: Graphic buttons in rel-patterns

This problem has been split into two parts, per Brian Suda's request. The problems are listed above and refer to this problem statement, namely:

Display properties of rel-patterns and Peeking into child elements to display rel-patterns

It is quite often that a site uses an image instead of a text link to present actions. For example: Instead of using the text "Download", they will use a graphic image with a downward-facing arrow.

In other words, if we have this:

Download:

<a href="http://my.site.com/download/3847293">
 <img src="/images/cool_download_button.png"/>
</a>

How do we present this option to a human being in a non-web-page UI?

Affected Parties

Any site that uses images for links. 'rel-sample', 'rel-enclosure', and 'rel-payment' are affected. Most of the examples also contain images instead of text for samples, downloads and purchase links. This is a demonstrable, widespread problem.

If there is no text to display, then how does one place the item into a menu/display for Operator/Firefox? Grabbing the image and placing it in a UI is a difficult argument to make - there are a variety of image sizes that might not do well in the Operator UI (or Firefox 3 UI).

Potential Solutions

Use the TITLE Attribute on the rel-pattern Links

Using @title was proposed in the following post to the mailing list:

http://microformats.org/discuss/mail/microformats-new/2007-July/000590.html

Here is an example markup:

<a rel="rel-enclosure" title="Download My Song"
   href="http://my.site.com/download/MySong.mp3">
 <img src="/images/cool_download_button.png"/>
</a>

Benefits

1. It POSH-ifies the website.
2. It works well with Operator, Firefox 3 and other uF parsers/UIs.
3. It fixes the accessibility/screen reader problem.

Drawbacks

1. There are arguments stating that the @alt attribute would be a better choice: http://microformats.org/discuss/mail/microformats-new/2007-July/000593.html

Use the ALT Attribute on the IMG Elements

Using @alt was proposed in the following post to the mailing list:

http://microformats.org/discuss/mail/microformats-new/2007-July/000593.html

Here is example markup:

<a rel="rel-enclosure" href="http://my.site.com/download/MySong.mp3">
 <img alt="Download My Song" src="/images/cool_download_button.png"/>
</a>

Benefits

1. Valid HTML requires that the ALT tag is specified anyway, thus we're helping people write proper HTML.
2. Because of the item above, this is in-line with what most text-based browsers and screen readers expect.
3. There is no danger of @title and @alt both being specified and read by a screen reader.

Drawbacks

1. The point was raised that ALT tags are not used properly on the web: http://microformats.org/discuss/mail/microformats-new/2007-July/000598.html and http://microformats.org/wiki/hcard-faq#Why_is_IMG_alt_not_being_picked_up

History

1. ALT was proposed on the mailing list: http://microformats.org/discuss/mail/microformats-new/2007-July/000593.html
2. The point was raised that ALT tags are not used properly on the web: http://microformats.org/discuss/mail/microformats-new/2007-July/000598.html
3. Research was started on ALT tag usage on the web: http://microformats.org/discuss/mail/microformats-new/2007-July/000614.html and http://microformats.org/discuss/mail/microformats-new/2007-July/000596.html
4. Analysis on ALT tag usage on the web was reported: http://microformats.org/discuss/mail/microformats-new/2007-July/000624.html and http://microformats.org/discuss/mail/microformats-new/2007-July/000627.html
5. A very long argument started on the use/mis-use of ALT: http://microformats.org/discuss/mail/microformats-new/2007-July/000603.html
6. More research was performed and presented to the list asserting that ALT is not being mis-used on the web: http://microformats.org/discuss/mail/microformats-new/2007-July/000629.html
7. It was asserted that the research completed was not acceptable: http://microformats.org/discuss/mail/microformats-new/2007-July/000635.html
8. Things got nasty: http://microformats.org/discuss/mail/microformats-new/2007-July/000637.html
9. Posting stopped due to an End of Thread request: http://microformats.org/discuss/mail/microformats-new/2007-July/000644.html
10. The issue and discussion was documented on the wiki.

Separating the Issues

There are two actually issues here, which need to be untangled.

• 1 the yearning to have the rel attribute define 2 different things at once, an @href and a title.

<a href="foo/" rel="rel-value-here">bar</a>

rel has a very specific definition, it can only form a triple based on the href and the rel X is related to Y via rel-value-here

This proposal was to make it do a secondary thing, which it is not semantically designed to do.

X is related to Y via rel-value, and that rel-value has a value of TITLE (@alt or @title).

this should be fixed in 1 of 2 ways... either with a value unto itself, class="rel-value-here-title" because rel-value has a value or TITLE doesn´t make much sense semantically. -or- not at all. Currently, this is how rel-license works. Operator (and other applications) are free to attempt to give a LABEL to that link how they see fit - there is no reason to define this in the spec, let the market define how to do this in the best way.

• 2, the whole "lets peek inside child elements not directly labeled with microformated properties"

The problem with peeking into child elements is, where does it end? if we do @alt for IMG then we should do @title on abbr as well, ...

I feel that peeking inside child elements is a bad idea. Microformats should only be concerned with elements that are explicitly marked-up as so. XSLT, XPath and RDFa also do NOT "peek inside child elements" so this gives us other technologies that have look into this issue and avoided it as well.

I hope this helps direct any of your further documentation. My suggestion would be to split this into 2 issues and discuss each on separately.brian suda 13:11, 2 August 2007 (UTC)