MetaTag

MeatballWiki | RecentChanges | Random Page | Indices | Categories

The meta tag is an HTML tag that resides in the <HEAD></HEAD>. It provides random "meta" information for non-human agents (er, software).

There are lots of meta tag types, and they are defined loosely, outside the standard. That makes them really hard to pin down, much like XML tags. Consequently, I always lose track of them. So, I'm going to list some here.

<META NAME="KEYWORDS" CONTENT="comma-separated list of keywords for SearchEngines">

<META NAME="DESCRIPTION" CONTENT="Description of page goes here">

<META NAME="AUTHOR" CONTENT="Author of page">

<META NAME="CATEGORY" CONTENT="home page">

: The category of the page.

<META NAME="GENERATOR" CONTENT="Program that created the page (e.g. Microsoft FrontPage)">

<META NAME="ROBOTS" CONTENT="NOINDEX,NOFOLLOW,NOCACHE,NOARCHIVE">

: And there are a lot more.

These are not exactly meta tags, they only share the tag name:

<META HTTP-EQUIV="Expires" CONTENT="Tue, 04 Dec 1993 21:29:02 GMT">

: The date should be in RFC 822 format. If it is 0, the page expires now.

<META HTTP-EQUIV="Refresh" CONTENT="number of seconds; url=URL to jump to">

<META HTTP-EQUIV="Pragma" CONTENT="no-cache">

: Tells the browser not to cache the page; doesn't work in MSIE4. And this is useless, since many older Proxies don't understand this HTTP header if only embedded into .html pages.

<META HTTP-EQUIV="PICS-LABEL" CONTENT="PICS rating definition">

: See http://www.rsac.org.

<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=iso8859-1">

: Set the MIME content of the page.

<META HTTP-EQUIV="Content-Disposition" CONTENT="inline; filename=newnameofthatpage.html">

: Gives pages meaningful "SaveAs?"-filenames, see ErfurtWiki. It is suddenly not used by many so called "CMS" because of some sort of pseudo-commercial content protectionism.

Note that tags using the HTTP-EQUIV attribute behave exactly as they would in HTTP headers. In fact, sometimes front end processors will just generate the corresponding HTTP header. HTTP headers are defined in RFC 1945 (HTTP/1.0) and RFC 2068 (HTTP/1.1).

The problem with them being, that they're typically not understood by intermediate proxies and thus should be avoided unless you cannot emit real HTTP headers. In a perfect world, Apache would pre-scan .html pages and emit the headers on behalf of that static page file (other Web servers do that). That is, all HTTP clients should better be able to understand some basic HTML features - the protocol and the page file format were thought to be used in conjuction.

There is a complete listing at http://vancouver-webpages.com/META. And by complete, I mean complete. Many of the tags will not be recognized by common browsers.

My favourite so far is from http://www.cyberpunkproject.org

<META NAME="cyberspace generator" CONTENT="OnoSendai?/2020 [en] (VROS; I) [cyberdeck]">

-- SunirShah

Almost all META tags are used only by browsers and search engines - proxy caches simply ignore the META tags and only look at HTTP headers. So the cache control tags are fairly useless. See TWiki:Codev/BrowserAndProxyCacheControl for some useful articles and a cacheability tester (see how cacheable your Wiki site really is...). -- RichardDonkin

CategoryWebTechnology

MetaTag

Discussion