| 9:13 pm on Aug 4, 2004 (gmt 0)|
Useful meta tags
1. Content type
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
Required?: Maybe. It is obligatory to declare your charset for every document you serve. However it is often better to declare the charset in the HTTP headers before the page is served. If you are using XHTML, you can also use an XML declaration above the doctype to replace the meta tag. Both of the latter methods take precedence over the meta tag.
<meta name="description" content="A short description of your page here.">
Required?: No, but it helps in the SERPs as your description appears with the link to your site. Don't use the same description for all your pages - personalize it to reflect the page's true content.
<meta name="keywords" content="widgets redwidgets foobars">
Required?: Probably not - the keywords tag has been so heavily abused that it is ignored by almost all major search bots - although apparently Yahoo Slurp gives it a very minor place in it's algorithm.
<meta name="robots" content="noindex,follow">
Required?: Better replaced by a robots.txt file, but useful on occasion to mark specific files which you don't want to be indexed. Note that specifying
content="all" is unneccessary - the bots will index anything not specifically excluded, so the "all" is implicit.
5. Image toolbar
<meta http-equiv="imagetoolbar" content="false">
Required?: It's useful to switch off that silly mouseover toolbar in IE, but only needed on pages with large graphics.
6. Author, Revisit-after, fifteen million other things
<meta name="favorite-fuzzy-animal" content="koala">
Required?: Utter waste of bandwidth - you're just pushing your content further down the page. They have no use whatsoever.
Conclusion - there should be a maximum of five meta tags on your site, and most of the time, you can get away with one or two - or maybe none at all.
| 10:32 pm on Aug 4, 2004 (gmt 0)|
Spot on, encyclo - exactly the way I see it.
I never dropped the keywords tag when the search engines stopped using it, because I see it as a very handy place to keep notes about what words my page is targeting. I often have as few as 2 or 3 kws in there, and rarely more than 8.
| 8:34 am on Aug 5, 2004 (gmt 0)|
I'd agree with that as well, although have a look at the Dublin Core Metadata Initiative [dublincore.org] - it is an interesting approach.
| 11:23 am on Aug 5, 2004 (gmt 0)|
Thinking about it, I should have mentioned PICS labels as well - which are required for certain kinds of sites to get past content filters - so they don't fall into the favorite-fuzzy-animal category.
Having said that, PICS labels can successfully be sent as an HTTP header too, possibly avoiding the need for another meta tag.
| 2:32 pm on Aug 5, 2004 (gmt 0)|
To reply to BlobFisk's suggestion about the Dublin Core meta tags, they are a good idea in theory, but there is always one major problem when it comes to document classification by the page author: people lie. The experience with keyword-stuffing of the keywords meta tag shows the extent of the problem - you can't rely on the publisher of the document to accurately portray it's contents within the metadata - because that information may be abused for getting a better SERPs placement (usually) or for other reasons.
Of course, if you want to classify a document and you can't rely on the metadata, you have to use another method: for example, a third-party classification dependent on on-page factors, inbound links, etc. all arranged according to a complex algorithm - the best-known of which is Google.
Otherwise, we could all try putting something like
<meta name="DC.Googlerank.position" content="1">
You never know, it might work... ;)
See also Metacrap [well.com] which covers the problem in more excruciating detail than you'd ever need.
(Sorry about the double post)
| 8:02 pm on Aug 5, 2004 (gmt 0)|
I agree that currently few meta terms on internet pages actually do anything active, most act as a "comments" section for the site maintainer (or as a liars section for the immature egoist).
However, on an intranet many can be invaluable. The "closed" environment makes "truth" in meta-data more likely than out in the world wild web and the author/date originated/date updated/version/contributors/etc. elements can be used to narrow searches.
As usual with anything computers, if you see a need, use it; if not, do not.
| 8:58 pm on Aug 5, 2004 (gmt 0)|
|What about the title meta tag? (As opposed to the title tag.) Does anyone think that serves a useful purpose, perhaps in SEO? I use it as a kind of subheading, so that the "title/meta title" combination are similar to the "subject/meta description" used when posting a new topic on this forum.|
And what about the author meta tag? Its contents are short, but is it nevertheless a waste of space?
| 10:44 pm on Aug 5, 2004 (gmt 0)|
I've seen no evidence that any major search engine considers the title or author meta tags.
| 11:36 pm on Aug 5, 2004 (gmt 0)|
search engines goes for keyword and headlines
| 4:34 am on Aug 6, 2004 (gmt 0)|
There are some circumstances where use of other metatags can be very helpful.
For example, I work in a government organisation which has 40,000+ web pages and also has a large corporate intranet. We have our own search engine (we use htdig) which we use for searching our own pages (both internet and intranet). We have a corporate policy of using the Dublin core standard metatags plus half a dozen others. We configure the search engine to give special weighting to different metatags to give us the most useful ranking possible. So for us, the kewords metatag is very useful, so is date.modified, and so are others that are not often used by other sites.
In conclusion: How you choose to use metatags to record information about your pages depends on your particular circumstances.