RH5 Search - Eliminating Obsolete Topics

[Mediadocs]

:confused; Does anyone know how and at what point RoboHelp 5 generates the data it uses to support searches--does it refresh this every time you do a build/publish? When does it clear out the old data? I'm maintaining a large existing webhelp modular help system that may have been improperly maintained in the last few years.

At which point or points I need to do a clean-up in order to avoid users retrieving obsolete topics that were never deleted.

For example, if I have obsolete data in the Search results, will clearing the publish folder and republishing help me or do I have to go through the project folders as well?

Regards,
Aileen

[Captiv8r]

Welcome to our community, Aileen

When you generate WebHelp, all the files that are used to support Searching, Indexing and whatnot are created fresh. When you publish, normally only newer files are pushed out to the server. So unless your user has saved a favorite pointing to the old information, they should not be able to find it.

However, this does introduce a question that folks sometimes have. What about that old data that no longer applies? How does one get rid of it?

You may wish to do this from time to time. Unfortunately it's best accomplished during "off hours" when the system isn't likely to be used. You delete the whole help system from the server. Then publish fresh. This action will purge the old data and ensure that only what should be there is actually there.

Cheers... Rick

[Mediadocs]

Thank you Rick!

[HKabaker]

I don't quarrel with Rick's suggestion, but -- depending on the size of the help package and how many changes/orphans there are -- you may want to compare files first.

Here's another plug for Beyond Compare, a nifty utility for comparing batches of files of the same name and directory structure.

If you can access the server in a Windows directory / filename context, Beyond Compare will flag the differences. Comparison criteria are flexible (timestamps to checksums). You can copy everything over, or just the revised files -- all or selectively -- or just delete the orphans.

Works best when you can do it yourself; not so easily if you have to go through an intermediary.

Harvey