microformats2-parsing-brainstorming: Difference between revisions

From Microformats Wiki
Jump to navigation Jump to search
(→‎feedback on more rel info: follow-up on glennjones questions)
(for issues, use -issues)
Line 5: Line 5:
For the microformats2 parsing algorithm, see:
For the microformats2 parsing algorithm, see:
* [[microformats2-parsing]]
* [[microformats2-parsing]]
For filing issues / problems with microformats2-parsing, see:
* [[microformats2-parsing-issues]]


Revision as of 22:04, 1 June 2015

<entry-title>microformats2 parsing brainstorming</entry-title>

This page is for brainstorming, discussion, and other questions and explorations about microformats2 parsing.

For the microformats2 parsing algorithm, see:

For filing issues / problems with microformats2-parsing, see:

Nested h-* objects' "value" property

Status: resolved, awaiting implementation attempt/experience.

Raised 2015-01-06 by User:Kylewm;

If a child element has a microformat (h-*) and is a property element (p-*, u-*, dt-*, e-*), the parser will add a "value" property to the resulting object. The value should attempt to be a useful representation of the object for consumers that do not have semantic knowledge of the particular h-* type. Ref: microformats2-parsing#parse_an_element_for_class_microformats.

To determine the "value", we parse the property element simply (as if it did not have a h-* class), which works well for simple h-* objects, e.g. <a class="u-like-of h-cite" href="...">...</a>

  • To handle more complex microformats, I propose that "value" for a p-* property element take on the first explicit "name" property of the nested microformat, and for a u-* property, the first explicit "url" property. Parsing will fall back on the current rules if an explicit property does not exist.
    • This makes sense to me, and fits with the use-cases and examples I've seen. Tantek 19:31, 6 January 2015 (UTC)
    • A similar (possibly simpler?) formulation would use the implied name and url rules to determine the "value" for p-* and u-* properties respectively
      • I don't think that's needed, as there are already implied rules on a property that should handle that. I'd start with just the "first explicit" scoping to be more conservative, and then see if we find any use-cases that (and existing implied rules) don't/doesn't catch. Tantek 19:31, 6 January 2015 (UTC)

For example:

<div class="h-entry">
  <div class="u-in-reply-to h-cite">
    <a class="p-author h-card" href="http://example.com">Example Author</a>
    <a class="p-name u-url" href="http://example.com/post">Example Post</a>

The nested u-in-reply-to object would parse as

"in-reply-to": [{ 
  "type": ["h-cite"],
  "properties": {
    "name": ["Example Post"],
    "url": ["http://example.com/post"],
    "author": [{
      "properties": {
        "url": ["http://example.com"], 
        "name": ["Example Author"]
      "value": "Example Author"
  "value": "http://example.com/post"

where the outer "value" gets the in-reply-to h-cite's u-url property, and the inner "value" gets the author's p-name property.

  • Because there are no implied properties of the dt-* and e-* types, and no obvious defaults, the value rules for these types would not change.
    • A possibility for dt-* h-*: The dt-* could take either the first dt-* of the h-*, or (perhaps if no dt-* in the h-*,) the first <time> element inside. Tantek 19:31, 6 January 2015 (UTC)
    • First dt-* seems reasonable, predictable, and usable. Consensus at 2015-01-20 meetup.
    • Update 2015-05-29: no known use-cases for first dt-* or first e-*, and implementing that "would require some refactoring" (in mf2py at least per kylewm), thus until there's a use-case for first dt-*/e-* inside, let's treat "dt-* h-*" and "e-* h-*" as before. Tantek . In particular:
      • p-* h-* - value from first "name" as proposed above
      • u-* h-* - value from first "url" as proposed above
      • e-* h-* - value is already defined for e-* parsing, nothing special here
      • dt-* h-* - value from normal dt-* parsing - nothing special.
      • +1 totally agree, let's wait for use cases of e-* dt-* Kylewm 19:44, 29 May 2015 (UTC)

Canonicalization of datetime output

Status: resolved, awaiting implementation attempt/experience.

It would be useful to choose a (more) uniform output format for datetimes to make it easier for users of the parser to consume datetimes. Microformats2 parsers already do sophisticated pattern matching to recognize date vs. time vs. datetimes, so converting this to any specific format should not add overhead.


  • Choose either 'T' or space as the date/time separator.
    • Prefer space as it is more human friendly/readable, which matters even for syntaxes/formats, as human still develop, debug them. Tantek 19:31, 6 January 2015 (UTC)
  • Choose either +XXYY or +XX:YY as the timezone specification (and convert 'Z' to +0000).
    • Would appreciate some study / input here as to which timezone offset syntax is more human friendly. I lean slightly toward +/-NNNN (without the colon) because in the context of seeing a time, leaving out the colon makes it less likely the offset will be confused for a time. E.g. "07:00-08:00" looks like 7-8am, even if it meant 07:00 in PST. Tantek 19:31, 6 January 2015 (UTC)
    • Space is fine - consensus 2015-01-20 meetup.
  • Parsers should not attempt make datetimes more exact than specified. They should not add time, seconds, or timezone if omitted in the original. Kylewm 04:02, 14 May 2014 (UTC)
    • Agreed. Tantek 19:31, 6 January 2015 (UTC)
    • or month, day per Tom Morris
    • consensus 2015-01-20 meetup
  • Counterpoint: PHP's builtin date parsing does not require strict formatting. And the equivalent functionality for Python is provided by the widely used python-dateutil library. Kylewm 19:02, 14 May 2014 (UTC)
    • However we cannot (must not) depend on either PHP or Python's "smart" "fixing" or Postelian "liberal handling", or any other language/framework's for that matter, as they all differ in how "intelligent" they are. Tantek 19:31, 6 January 2015 (UTC)

Perhaps just provide a guideline for these based on the above consensus.

Add meta http-equiv to microformats2 parsing model

Status: disagreement, awaiting implementation attempt/experience.

Similar to document level parsing of rel attributes, it makes sense simultaneously to parse <meta http-equiv> elements, perhaps treating "Status" in a special way (only using first number (sequence of digits) for its "value").

Use case: IndieWeb "deleted" indication inline in content for static file services that don't support HTTP return codes.

HTTP Header example:

  • Content-Type: text/html; charset=utf-8

HTML equivalent:

  • <meta http-equiv="Content-Type" content="text/html; charset=utf-8">


  • Interesting thought. Are you suggesting a top level "http-equivs:" collection similar to "rels:" in the parsed output? Should we consider "metas:" instead or in addition? Tantek 19:31, 6 January 2015 (UTC)
  • What's the use case for this? Also, http-equiv on its own is useless. http-equiv is only a supplement to the data stored in headers. And headers aren't always there: what happens in the context of someone debugging a page who pastes the source into the textarea of an mf2 parser? Without a compelling use case for including headers (and then over-riding some of them with http-equivs), I'm not sure why an implementor want to do this. —Tom Morris 00:25, 8 May 2015 (UTC)

E.g. from https://gist.github.com/aaronpk/10297489

<meta http-equiv="Status" content="410 GONE"/>
 "items": [],
 "rels": {},
 "http": {
 "status": 410
  • Maybe make this an optional pass in the parser? - Tom Morris 2015-01-20
  • For now, don't bother with metas until someone provides a use-case. Tom Morris
  • Agreed on both counts. Tantek 06:56, 21 January 2015 (UTC)

MIME type

See microformats2-mime-type

Other Interpretation Parsing Notes

Note: most of these need to be written up as separate microformats2-parsing-issues

Author: Ben Ward

Microformats 2 proposes a new, all encompassing syntax modification of prefixes that will allow microformats to be parsed from pages by processors without prior knowledge of a vocabulary. The core components of this model are quite simple, are quite simple to implement, but there are a number of conflicts that emerge with the functionality of existing microformats parsers that need to be handled. This page documents a proposed model to separate these concerns clearly in a way that can be applied to the documentation of generic microformats parsing rules, and the documentation of individual vocabularies.

Collection of other unresolved parsing issues in a generic model:

This is good material for documenting as microformats-2-issues, microformats-2-faq, and perhaps some of the more technical details in microformats-2-parsing-faq.

  • The include pattern references other elements from elsewhere in a document. A generic parser needs to track IDs and fill them in after walking the DOM. (also, itemref if adopted.)
  • Will itemref always map to an item property name?
    • No, itemref maps to one or more elements by ids, and their children. Those referenced elements may have property class names themselves, or they may contain elements that do. Tantek
  • hAtom implies author from an hCard in a page that uses an address element. This requires format knowledge, but a generic parser does not currently track the element type of a property node. Should it?
    • It should not. element-specific handling (e.g. using "alt" from img, and "title" from abbr) is completely done at parse time. The JSON data model does not reflect which element type or attribute the value came from. Additionally, hAtom is an example where we created far too many vocabulary-specific rules, in practice they're not necessary, and only complicate the microformat for both publisher understanding and parser implementation. Tantek
  • hAtom defines that the highest level heading within an entry implies entry-title. This particular optimisation might be better off dead.
    • Agreed, this is gone in microformats 2. Tantek
  • hAtom defines that permalinks be parsed from rel attributes, not class
    • In practice this has been one of the more problematic/error prone aspects of hAtom implementations, and it's also inconsistent with other microformats (although hReview tried to use both rel permalinks and "url"). The dependence upon rel-bookmark for permalinks is dropped in h-atom in preference to re-using "u-url" and "u-uid". Tantek
  • XFN is entirely built on rel (although, has various other differences from structural microformats, as do vote-links, so perhaps are excluded from this discussion and will always be handled by dedicated parsers/queries regardless?)
    • The best (easiest and most reliable) use of 'rel' microformats in practice is when they are orthogonal to 'class' microformats. This is true both with XFN and some newer rel values like rel-author. In addition, it was very clear at the recent schema.org workshop's syntax session that RDFa's decision to apparently arbitrarily mix use of 'rel' and 'property' attributes for specifying different types of properties (it wasn't clear to people in the room when you use which for what) has caused a high degree of confusion among publishers and thus high error-rates. Thus if anything we should learn from both the mistakes of RDFa and our own experiences with even very deliberate/specific mixing of rel microformats in class microformats, and keep them defined as separate orthogonal building blocks that work together, but don't depend on each other. Tantek
      • Relatedly to this: rel-tag in hAtom. --BenWard 06:50, 5 October 2011 (UTC)
        • Yes, and two related things here. First, despite my (and others') objection and (past) interoperable post/entry-specific treatment by Technorati and Ice Rocket, Hixie has redefined rel-tag in HTML5 to mean applying to the whole page, not a single post. Second, I've explicitly added 'p-category' to the draft 'h-atom' vocabulary in microformats-2. Tantek 07:12, 5 October 2011 (UTC)
  • HTML's time element includes an optional pubdate attribute. Simply: We should parse this as dt-published. --BenWard 06:12, 10 October 2011 (UTC)
    • *If* there is even some reasonable data on actual use of the "pubdate" attribute (I don't think there is, frankly, especially with the removal of the algorithm to produce Atom from HTML5), then we could consider parsing "pubdate" as backwards compatible option for "dt-published". As a general rule, however, it is bad (demonstrably/experienced) design to depend on additional attributes (c.f. RDFa confusion over "property" vs. "rel"), especially for an instance where no additional attribute is necessary. I would leave this out for now until there is non-trivial (more than just test pages or folks who've written HTML5 books, ahem) use in the wild. When there is such use in the wild, it should be documented on a wiki page. We don't want to encourage more complex (additional attribute) publishing as a result of supporting it. Tantek 12:12, 10 October 2011 (UTC)
  • value-class-pattern: In microformats-2, since there are no sub-properties, there will presumably no-longer be a 'value' property in any parsed model. Properties such as 'tel > type' in hCard are, as I recall, deprecated due to underuse anyway, so 'tel > value' becomes redundant. (There's also potentially some clarification around 'price > value' in hListing, whereby value was used in a pattern. So, what does this mean for value class parsing, with regard to value-title patterns and date separation patterns. Are we looking for a 'p-value' and 'p-value-title' classname, but treating them specially (excluding them from regular property parsing.) Or, are we giving them a special prefix (v-text, v-title? That seems confusing, but could be a concept.) I'm fine with p- for both, and just having the parser ignore them since they're special, but need clarification and naming confirmation. --BenWard 09:35, 10 October 2011 (UTC)
    • A few things:
      • 1. Yes, no more subproperties. 'tel' becomes just 'p-tel'. If there is demand for a structured 'tel' value, then we can use that demand (and research into publishing in practice) to brainstorm and create an 'h-tel' structured telephone number (with perhaps fields like 'type', 'extension', some indication of it being local dialing (an extra 0 in some countries) or international dialing, etc.) Or, we address the different 'tel' types as their own flat properties (again as justified by research), e.g. perhaps 'p-tel-fax', or 'p-tel-mobile'. Something for hcard-2-brainstorming.
      • 2. For prices, e.g. hListing, either we're going to need to encode how to parse monetary amounts including monetary symbols, or consider creating an 'h-price' structured price. Not sure what the right answer is here, again, will need to be informed by analysis of documented actual price publication practices.
      • 3. We should avoid introducing a new prefix 'v-' just for value-class-pattern. As we've noted elsewhere, each new prefix adds complexity and should be avoided without substantial advantage.
      • 4. Using 'p-value-title' is strange, as it would be an exception to 'p-' parsing, since it would get the value from the 'title' attribute whereas 'p-' properties don't normally do that (exception: abbr).
      • 5. Using 'p-value' is also strange, as it wouldn't generate a 'value' property in the JSON data model.
      • 6. Class name 'value-title' is already sufficiently prefixed - we've found or even heard of no collisions in practice.
      • 7. Class name 'value' can, by its simpler naming nature, be expected to potentially collide with other web designer class name usage though we have no documentation/mention thereof. We could consider a renaming, or providing of alternative, such as 'value-string', or 'value-content', etc. However, let's keep that as a backup plan to use only if/when evidence is presented that we need to.
    • Conclusions: for now, in microformats-2, keep using 'value' and 'value-title' as defined in the value-class-pattern, and add the additional (obvious) interpretation that value class pattern: date and time parsing applies to all 'dt-' properties. - Tantek 12:12, 10 October 2011 (UTC)

incorporated 2015-05-28

The following brainstorms were incorporated 2015-05-28.

more information for alternates

Raised 2015-04-24 by Kevin Marks

The existing alternate parsing is omitting title - that should be added. The text would make sense to add here too.

Use-case: labels for presenting alternates

  • +1 Makes sense. Tantek 03:41, 25 April 2015 (UTC)

more information for rel-based formats

Raised 2015-04-18 by Kevin Marks

Related github test suite issue: https://github.com/microformats/tests/issues/16

Several rel-based formats have additional information that is useful beyond the link itself, which is all we capture at the moment. As I am trying to update the Universal feedparser to support mf2 based I will show examples from the testcases there.

The main change is to add a rel-urls entry for more information about the attributes and text of the urls pointed to by rel's in the document

A fork of mf2py that implements these changes is at https://github.com/kevinmarks/mf2py


<a rel="tag" href="http://del.icio.us/tag/tech">Technology</a> 

currently parses to:

{"rels": {"tag": ["http://del.icio.us/tag/tech"]}, "items": []} 

This loses the link text, which is useful as a label.

We add a rel-urls element to the parsed output with this extra data that can be looked up from the rels, which doesn't break backward compatibility and works better with xfn (see below)

    "rels": {
        "tag": [
    "items": [], 
    "rel-urls": {
        "http://del.icio.us/tag/tech": {
            "rels": [
            "text": "Technology"


<a rel="coworker" href="http://example.com/johndoe">John Doe</a>

currently parses to:

{"rels": {"coworker": ["http://example.com/johndoe"]}, "items": []}

This loses the link text, which is the person's name. Suggested output using the urls object:

    "rels": {
        "coworker": [
    "items": [], 
    "rel-urls": {
        "http://example.com/johndoe": {
            "rels": [
            "text": "John Doe"

with multiple xfn values

<a rel="coworker friend" href="http://example.com/johndoe">John Doe</a> we get this:

    "rels": {
        "coworker": [
        "friend": [
    "items": [], 
    "rel-urls": {
        "http://example.com/johndoe": {
            "rels": [
            "text": "John Doe"


<a rel="enclosure" href="http://example.com/movie.mp4" type="video/mpeg" title="real title">my movie</a>

currently parses to:

'{"rels": {"enclosure": ["http://example.com/movie.mp4"]}, "items": []}'

This loses the link text, which is the title and the attributes which give type. Suggested output:

    "rels": {
        "enclosure": [
    "items": [], 
    "rel-urls": {
        "http://example.com/movie.mp4": {
            "rels": [
            "text": "my movie", 
            "type": "video/mpeg", 
            "title": "real title"

This generalises to other rel's too, such as rel-feed and rel-alternate that have type, lang etc attributes.

(updated to include changes from feedback below) Kevin Marks 22:13, 26 April 2015 (UTC)

attributes parsed

Attributes currently parsed are:

  • hreflang for alternate and enclosure
  • media for alternate and enclosure
  • title for alternate and enclosure
  • type for alternate and enclosure

Attributes we may consider parsing if we have a use case are

  • sizes for icon - need use-case documentation
  • coords for area - possibly for people tagging - no examples yet, and unnecessary as people-tagging requires using h-* microformats
  • shape for area - possibly for people tagging - no examples yet, and unnecessary as people-tagging requires using h-* microformats

In addition there is a special attribute name text which is the text contents of the link, which is useful in rel-tag rel-enclosure and xfn, and in alternate when used for feeds. It's also clarifying for rel-me links.

Tantek suggests we use textContent for this instead, and make it a single string, not a list as name is elsewhere in mf2 parsing

  • Update: "text" is good enough, and "textContent" is ugly camelCase. Tantek 04:39, 29 May 2015 (UTC)

feedback on more rel info

  1. "name" is bad because it misleadingly conflates with use of "name" elsewhere in microformats2.
    • Suggested alternative: textContent - since that's literally what is being returned there. Tantek 02:35, 25 April 2015 (UTC)
      • as all other mf2 keys are lowercase-with-hyphens, Tantek suggests 'text' as that isn't going to be an html Kevin Marks 07:28, 25 April 2015 (UTC)
  2. no need for array for "name"/textContent - since there is always only one at most
    • E.g. should be "textContent": "my movie" Tantek 02:35, 25 April 2015 (UTC)
    • Update: "text": "my movie" Tantek 04:39, 29 May 2015 (UTC)
  3. "urls" key is misleading - implies all URLs in the document, which is neither true, nor desired (takes much more parsing time and work and code)
    • Suggested alternative: "rel-urls". And open to better alternatives too. Tantek 02:35, 25 April 2015 (UTC)
      • If we are trying to extend the number of properties retuned from a rel without breaking the old structure why don't we call the new structure something like "rels-extended" Glenn Jones 12:29, 1 June 2015 (UTC)
        • Extension is not the point, but rather to use them complementary. One structure for look-up of any rel value, hence "rels", which returns you a list of URLs. Then you can lookup those URLs in the new mapping, by URL, hence it is called "rel-urls" - that's the point to use them in conjunction and that's why rel-urls is named what it is. Tantek 22:03, 1 June 2015 (UTC)
  1. Why is the structure of "rel-urls" different to the "alternates" structure. Should the "url" not just be added as a property and not as a key. Creating two data structures for one type of object seems inconsistent. It adds cognitive load to anyone trying to understand the JSON structure Glenn Jones 12:29, 1 June 2015 (UTC)
    • I was trying to avoid breaking the existing rels structure and use of it - I did implement a variant that put the structure inside rels, and it became cumbersome and repetitive where there were multiple rels on a url (xfn cases). Denormalising as properties of the URL made more sense. It also dedupes if there is repetitive linking to the same URL, eg a series of posts with rel-author on each. Kevin Marks 20:05, 1 June 2015 (UTC)
  2. If the rel is a "tag" then the main value we need to return should be the last path component of the URL, not the link text? Should we add another output property ie "tag" Glenn Jones 12:29, 1 June 2015 (UTC)
    • No need to return last path segment of the URL, because the URL is already there - and that's just a library/framework utility function to get the last path segment of a URL. Tantek 22:03, 1 June 2015 (UTC)
  3. As currently described, the URL from alternates is repeated in the rel-urls structure. If we are doing this, surely alternate should be in rels too? I assumed a mapping between them. Kevin Marks 20:05, 1 June 2015 (UTC)
    • Yes it makes sense to drop "alternates" assuming the backcompat impact is low, put alternates in "rels" along with everything else, and direct people to use rels and rel-urls for alternates functionality. Evidence this is an acceptable even preferable approach.[1] Will add an issue accordingly. Tantek 22:03, 1 June 2015 (UTC)

see also