category-brainstorming

category is a field used in many microformats, originally taken from hCard.

Type name: CATEGORIES Type purpose: To specify application category information about the vCard.

— RFC2426

It is closely related to rel-tag, but has no dependency on URL schemes, it simply marks up the inner-text as a category keyword.

Brainstorming on this page may be applicable to parsing of category and/or rel-tag.

Supported Languages

An object — a person, or item (in a hProduct or hListing context) — may be consumable in multiple languages. A person may communicate in multiple languages, a set of self-assembly shelves may come with an instruction manual in multiple languages, a web application (also a product) can be localized to support use in multiple languages.

In a loose form, supported language is a form of categorization, but it is useful to filter by language separately from other tags/categories (‘Only show me people that speak languages I speak’, ‘Only show applications available in my language’).

Localization vs. internationalization need to be distinguished. The term "language" is too abstract/vague/ambiguous in this context.

examples of language publishing

Language on the web are indicated in multiple forms on the web:

Note: at least a live URL example for each of these would help make these examples appear less theoretical. — Tantek 23:41, 1 September 2009 (UTC)

• Names (e.g. ‘Français’, ‘Deutsch’, ‘English (British)’
• Abbreviated local names ‘En’, ‘Fr’, ‘De’
• Abbreviated country names: ‘GB’, ‘UK’, ‘ES’
• Local names (e.g. ‘French’, ‘German’, ‘British English’, ‘Dutch’)
• ISO 639 language codes (e.g. 'en-gb', 'en', 'fr', 'de')

existing formats for marking up language

HTML already has a lang attribute for marking up actual content. This uses ISO 639 codes.

brainstorm requirements

Categorizing languages has the following requirements:

• Differentiate from other types of category
  • Why? Based on what use case(s) and what real world examples? — Tantek 23:41, 1 September 2009 (UTC)
• Allow the language to be presented in local name form
  • Needs citing of real world examples that do so. — Tantek 23:41, 1 September 2009 (UTC)
• Allow the language itself to be ascertained reliably and interoperably
  • Based on what use case for ascertaining? — Tantek 23:41, 1 September 2009 (UTC)

Proposal 1: class='category language'

• Use an additional class-name language to indicate that a category is marking up a supported language for the object.
  • See above re: localization vs. internationalization. This choice of term is too ambiguous and does not relate directly to a particular use case. — Tantek 23:41, 1 September 2009 (UTC)
• class=language also instructs parsers to read the ISO code from the lang attribute of the element, in place of the inner text, where present.
  • Parsers are *already* reading the lang attribute of the element. — Tantek 23:41, 1 September 2009 (UTC)
  • If you want a generic mechanism for pulling a microformat property value from the "lang" attribute of the element rather than the contents of the element, then it might be better to follow the pattern set by the value-class-pattern's "value-title" and introduce a term like "value-lang". — Tantek 23:41, 1 September 2009 (UTC)

e.g.

<ul>
    <li class="category language" lang="en-gb">English (British)</li>
    <li class="category language" lang="en">English (American)</li>
    <li class="category language" lang="fr">Français</li>
    <li class="category language" lang="de">Deutsch</li>
</ul>

In HTML, the strings should be marked up with a lang attribute to ensure that the local named languages are pronounced correctly by screen readers (for example).

With rel-tag:

<ul>
    <li><a class='language' rel='tag' href='/apps/lang/en' lang='en'>English</a></li>
    <li><a class='language' rel='tag' href='/apps/lang/en-gb' lang='en-gb'>English (British)</a></li>
    <li><a class='language' rel='tag' href='/apps/lang/fr' lang='fr'>Français</a></li>
    <li><a class='language' rel='tag' href='/apps/lang/de' lang='de'>Deutsch</a></li>
</ul>

In this case, the URL scheme is shown to also use ISO codes. This may not be the case in production environments (they may choose to use /apps/lang/english, for example), so parsers look explicitly for language information should be instructed to again look for for the lang attribute where rel-tag is used in combination with class-language. --BenWard 01:16, 1 September 2009 (UTC)

Questions

• JonathanMalek 15:19, 1 September 2009 (UTC): I've been looking for ways to express more information about a particular category. For instance, when categorizing a particular hAtom entry as "about Los Angeles, CA", it would be helpful to also mark up that particular rel-tag with the geo data for LA. Do you see this category proposal as a way to support richer tags?

issues

• I think the main implied use case here would be better researched, analyzed, and developed through the creation of:

localization-examples, localization-formats, localization-brainstorming. — Tantek 23:41, 1 September 2009 (UTC)

see also

• hCard
• rel-tag