Searching using a famous search engine

[skiwi]

I was using a famous search engine to search for the phrase "sparx database engineering".
The first result returned was
http://www.sparxsystems.com/enterprise_architect_user_guide/9.3/database_engineering/database_engineering.html


Now here's the thing.
It would be preferable if the documentation for the latest version was returned, viz
http://www.sparxsystems.com/enterprise_architect_user_guide/13.5/model_domains/database_engineering.html



My recommendation

• search engines should be restricted from seeing anything other than the latest version of a page
• on the page itself links to previous versions of pages should be provided, perhaps in a  margin

On a related note, I often link from our documentation to Sparx documentation,
to reduce ongoing maintenance it would be useful have an option such as http://www.sparxsystems.com/enterprise_architect_user_guide/latest/model_domains/database_engineering.html

[Glassboy]

My recommendation

• search engines should be restricted from seeing anything other than the latest version of a page
• on the page itself links to previous versions of pages should be provided, perhaps in a  margin

Not the first time this has been mentioned :-)

[Eve]

On a related note, I often link from our documentation to Sparx documentation,
to reduce ongoing maintenance it would be useful have an option such as http://www.sparxsystems.com/enterprise_architect_user_guide/latest/model_domains/database_engineering.html

This could actually cause more ongoing maintenance. Sometimes it is necessary to add or remove pages. We can keep such structural changes limited to version increases, but couldn't guarantee that a link to latest would still be valid (or notify you if it changed).

[MaXyM]

Database of redirects should help.
However I disagree with hiding old pages from indexing robots. These should remain available otherwise it would harm SEO of the domain. There are methods how feed robots with recent pages. There are are robot index files and so on.

with regards

[Uffe]

Hey - a ring! Let's throw a hat in there!

The problem here is that the famous search engine doesn't return the most relevant result first. This tends to be self-perpetuating, since people tend to follow early links first, and so the famous search engine marks those as more popular -- even though realizing that a part of a URL constitutes a version number and giving the more recent one priority isn't hard to do. It is, in fact, urine simple. But as I say, this is a problem with a famous search engine, not with the indexed web site.

I don't agree that telling asking politely the famous web crawler not to index documentation for older versions is a good idea. Larger organizations especially are typically a ways behind the upgrade curve, and you want to be able to find the documentation for your version using the famous search engine rather than the unmitigated horror show local site search the site owner provides.

But some sort of alert on older-version pages seems like a great idea. There should be a link to the most recent version of the same page, if available, or to the lower-most matching node in the documentation tree if not.

Ideally, each page should also contain information on which versions it's relevant for. That way you'd know, if you were looking at an old page, whether there'd be any point in looking at a more recent one.


/Uffe

[Sunshine]

You could just add the version to the search parameters