Meta Data: What do you use | April 30, 2004

I was thinking about meta data earlier today. I see lot's of sites who's head tags are stuffed full of meta tags and link tags, yet I rarely make use of such data on the sites I build. So I'm interest In what meta data you put into your sites and why.

Posted at April 30, 2004 12:10 PM

Comments

David Barrett said on April 30, 2004 12:30 PM

One form of metadata that I’ll be using on the redesign of my site is traceback pings. I thought of a way they can be used to provide “related articles” links.

I’ll lock off the Movabletype traceback script to accept connections only from the webserver itself. Because a “related article” link should really be bidirectional, I won’t be using this directly. What I will be using is another script, that I’ll be writing myself, that will ping both the destination AND source blog entry.

The “related entries” will end up just being a list of TB pings.

Because my site will have four MT blogs (each blog having a seperate scope) with categories consistent between them, I’ll have the power to do even cooler stuff, but I’ll keep that under my hat :)

Richard Rutter said on April 30, 2004 12:50 PM

My blog is stuffed to the gills with meta information, just because really. At work, however, I’m about to release a few changes to Multimap.com which will contain these meta tags:

In case Microsoft introduce SmartTags sometime in the future.

To indicate the position and country of the map being shown. Will also include the city when that is available. These tags can be picked up by crawlers such as GeoURL and maybe others as location-based technologies become more prevalent.

Specific intruction to crawlers to index.

I don’t bother with descriptions and keywords as they are pretty ineffective nowadays, particularly with respect to SEO.

I also add a bunch of link tags:

For an address bar icon.

To provide an additional method for navigating the site. These are used by Mozilla’s ‘Site Navigation’ toolbar and also exposed by text browsers such as Lynx.

Richard Rutter said on April 30, 2004 12:53 PM

My blog is stuffed to the gills with meta information, just because really. At work, however, I’m about to release a few changes to Multimap.com which will contain these meta tags:

In case Microsoft introduce SmartTags sometime in the future.

Specific intruction to crawlers to index.

I don’t bother with descriptions and keywords as they are pretty ineffective nowadays, particularly with respect to SEO.

I also add a bunch of link tags:

For an address bar icon.

To provide an additional method for navigating the site. These are used by Mozilla’s ‘Site Navigation’ toolbar and also exposed by text browsers such as Lynx.

Nathan Pitman said on April 30, 2004 1:44 PM

I use the ‘description’, ‘keywords’, ‘language’ and ‘author’ tags. Not sure why, just habit I guess… :)

Minz Meyer said on April 30, 2004 2:02 PM

Like Richard I also use the link tags heavily. Besides the one he introduced, I often find myself using
<link rel=”section” title=”[Name of MT Category]” />
and link to my MT category overview pages.
I am not sure if this is 100% correct from a semantic point of view, but well, it adds structure to the overall site.

David Barrett said on April 30, 2004 2:31 PM

Ah, you were talking about metadata tags, and not metadata in general :)

Milan Negovan said on April 30, 2004 5:04 PM

Lurking over at the SEO newsgroups it appears that most search engines (SE) don’t care about keyword metatags. It’s way too easy to spam SEs with irrelevant keywords so they ignore them these days.

Most of them ignore robots.txt, imho. Btw, check out Sun’s humorous robots.txt file.

Personally, I chose to stick with the description tag, content-language, and content-type. I also use the author tag, but it has no bearing on anything really. On certain keywords Google puts me right at the top which proves my point that keyword metatags are pointless. Content is what matters the most.

Jonathan Snook said on April 30, 2004 6:56 PM

Metadata for external purposes (external to the organization that runs the site) is often unnecessary. Things like keywords and description have become so abused as to not have any meaning at all.

For internal purposes, however, metadata can be the key to a more effective site. As mentioned in a previous comment, being able to link related information, or improving search capabilities could be done using accurate metadata. The Government of Canada, for example, uses the Dublin Core with the intention of tying in all departments and agencies into a central search tool.

I like the concept of link tags and the link toolbar for Firefox. But found that a lot of the sites I visit don’t use them, and those that do, don’t use them in a way that makes it more effective in the way I use their site. Where the link toolbar comes in handy is in creating next/previous/up/down navigation in a multi-page, multi-hierarchal document. These documents, however, are often created with their own on-page navigation so that the document is easier for all users to navigate.

Mark said on April 30, 2004 7:06 PM

I use the standard keyword, description, and author on my personal site and client sites.

I normally don’t have a need to use the robots tag because I usually want the default settings anyway, which don’t need a tag. If I need modifications I’ll throw them in .htaccess.

I have the geo/ICBM positioning stuff thrown in on my personal one as well, just because I think it’s neat. :)

James Craig said on April 30, 2004 10:58 PM

Besides the standard HTML metadata, my blog has GeoURL info and a few Dublin Core tags. I try to limit the page weight to data that I think is practically useful today. I can always add more tomorrow if I find a use for it. Until then, I don’t want to waste my readers bandwidth.

Michael Wilcox said on April 30, 2004 11:08 PM

I (sometimes) use a meta tag for author info (never keywords and description) and I use the link tags for site navigation, if applicable. I think that’s just plain useful.

Georg said on May 1, 2004 2:00 AM

This one: meta name=”generator” content=”HTML Tidy for Windows (vers 1st January 2002), see www.w3.org” is there in case someone wants to know what’s holding my pages together.

My: meta name=”description” seems to give results in google…

The ‘link rel’ for Opera, Mozilla, Lynx is also useful.

That’s it.

Mark said on May 1, 2004 2:45 AM

meta name=”author” and meta name=”copyright” for the sheer hell of it.

meta name=”description” contains the ‘excerpt’ data from any given blog entry, while meta name=”keywords” contains the ‘keywords’ data for the same blog entry. I’m not exactly sure how useful this is for visitin bots and what have you, but I guess there’s no harm done.

meta name=”generator” is Movabletype.

Then there’s a few extra bits and pieces to manage bots, image toolbar logging and to prevent MS Smart Tags hijacking my page.

Mark said on May 1, 2004 2:47 AM

I forgot to add, I don’t know whether or not it’s got anything to do with my meta tags, but I get a massive amount of referrals from Google search results, so something’s working.

MaThIbUs said on May 1, 2004 4:03 PM

Same as Richard mentioned.

Jeena Paradies said on May 2, 2004 2:06 AM

I use description with a page summary wich ist on my rss and for some time on my homepage because some search engines use it too.

I use

but I don’t know if it is ok to use mailto: there, but Mozilla and Lynx opens the mailclient so you can send me a email if you want.

I use the robots meta because the content on my homepage changes sometimes ervery day and it is unnecessary to Index the archive summary of the year, its better to use noindex, follow so robots can index only pages with real content.

I use some DC too but it is not rearly necessary.

Jeena

Jeena Paradies said on May 2, 2004 2:09 AM

Oh some content disappeared?

I use

but I don’t know if it is ok to use mailto: there, but Mozilla and Lynx opens the mailclient so you can send me a email if you want.

thomas said on May 2, 2004 3:25 PM

I don’t use any. Because I don’t really know which ones to use and how to use them. Does anyone have any good links about which ones are reccommended and how to use them and why, etc.?

Tim Parkin said on May 2, 2004 6:22 PM

Just q quick one, I know this was about other meta data, but you should know that yahoo is now using the meta description and keyword information in it’s search engine algorithm

Dan August said on May 3, 2004 11:38 AM

Like Thomas I think this is an underserved area, and if anyone knows any good articles on the subject for further reading, I would be very grateful.Right now I’m more og less just basing my meta-tags on “pure instinct”.

Richard Martin said on May 5, 2004 3:30 PM

A lot of work has been done in this area in the public sector in the UK. The e-Government Metadata Standard builds on the Dublin Core and is a useful starting point: http://www.govtalk.gov.uk/schemasstandards/metadata.asp.