Page Summary Formats
This page describes existing formats that enable an author to publish a summary description for a page. Whether or not this is worthy of doing research (e.g. page-summary-examples, page-summary-brainstorming) for a microformat is open to debate. Nonetheless, it doesn't hurt to at least have a place to document existing formats.
HTML4.01 provides meta description as an example in Appendix B: Performance, Implementation, and Design Notes of the HTML 4.01 spec. E.g.
<META name="description" content="Idyllic European vacations">
Google search results display the value of the meta description element for a page when that page is listed in Google search results.
- Invisible. meta description is invisible metadata on the page and therefore violates microformats principles.
- Depends on
headelement. This summary format depends on the author having access to the head element which in many (most?) web authoring/ publishing scenarios the content author has no access.
Mozilla has introduced a proprietary format called "microsummary", which is used in Firefox 2.0's support for bookmarks.
- that microsummaries are not intended simply to deliver a summary of a page; they may be used, for example, for a "deal of the day", "last updated" date or time, a new headline, or other timely and dynamic information.
- a "microformat" consisting of a
class="microsummary"attribute would allow a piece of content (a span including a "last updated date", for example), to be extracted from a page, by software.
- if multiple tags included
class="microsummary", we'd need to decide whether they each represent different microsummaries or they all represent a single microsummary whose value is the concatenation of each tag's content. The former seems saner, but sometimes microsummary content is scattered around the page (f.e. an eBay auction item microsummary might contain the name of the item, the auction end date, and the current high bid, but that information appears in three separate places in the page).
Examples of discussion
- Invisible. Though different from meta description in that the
linktag is used to relate to a summary at another URL - this doesn't actually markup the summary content itself. Since the content is not necessarily visible on the page itself, an author could easily change the page content without noticing or even knowing to bother to change the referenced "summary" page/ resource (except, of course, where the microsummary is automatically generated, such as by extracting the first "H2", or the content of a div or span with a specific ID, or the last-updated date). Thus this method has all the same potential flaws of typical invisible metadata publishing, which is that overtime it may rot, become out of date, unreliable and untrustworthy in general.
- Depends on
headelement. This summary format may depend on the author having access to the head element which in many (most?) web authoring/ publishing scenarios the content author has no access.
- Haven't we already solved 99% of this problem with RSS and the
titleelement? Ianloic 17:11, 10 Dec 2006 (PST)
- No, that's not the case. Andy Mabbett 10:29, 11 Dec 2006 (PST)
titleelement doesn't solve this problem, because it can only contain plaintext, whereas microsummaries are intended to ultimately contain rich content (HTML, images, etc.), and also because the
titleelement lives inside the document, so web clients have to load the much larger document to get access to the much smaller summary inside it (which is fine in some cases, but not in others). Myk Melez
- RSS/Atom, on the other hand, would be a great delivery mechanism for microsummaries. Support for feed-based microsummaries (using profiles of RSS/Atom with heuristics to determine summaries even for web sites that don't provide a specialized summary feed) was part of the original plan for Firefox 2. After some discussion, however, we decided that it would make the most sense for Firefox's Live Bookmarks feature to handle all feed-based bookmark updates and to focus the microsummaries feature specifically on client-side and server-side generated summaries (i.e. those extracted by XSLT-based microsummary generators) as well as simple server-side content summaries. Myk Melez
- In retrospect, I think not supporting feed-based microsummaries was a mistake, and microsummaries should indeed extract summaries from feeds, where available. I plan to make this happen for Firefox 3. Nevertheless, I do think there's a place for the kind of simple server-side content summaries currently available in Firefox 2, despite their limitations, in some cases of authors building sites with content that is a better candidate for summarization than syndication. And there's also a place for microsummary generators, which address a problem solved by neither feeds nor server-side content summaries. Myk Melez