By Mark Gibbs
Network World, 8/31/98

Aren't they messy? I mean, really! Have you ever seen a sloppier bunch than your intranet content providers - that is, with the exception of your kids?

If you're a parent, you probably know all too well how difficult it can be to get kids to clean up their rooms. More often than not, they ignore your pleas for tidiness.

And if your intranet has been running for more than a few months, that means it needs cleaning, just like a kid's room would. These five tips ought to help.

1. Find and fix dead links
When perusing the intranet, users get frustrated following links to nowhere, as happens when authors move or delete documents or Web sites but don't change associated hyperlinks. That's why you need to get at least one of the various tools now available for finding and correcting broken links. (You might need more than one because no single product yet offers a full feature set.)

I recommend Mercury Interactive Astra SiteManager 1.0 on the strength of its content analysis, reporting and site-mapping features. But Web Analyzer 2.0 from EveryWare Development and Coast WebMaster 2.02 from Coast Software are worth considering, too. (For a review of these and other Web content managers, see Network World, Dec. 1, 1997, page 53.)

Most important, IT has got to use these tools regularly. Ideally, that means searching and fixing on a weekly basis or, on a large intranet, daily. If the intranet is overly large, you may have to do daily scans selectively.

What's more, it should be standard practice for intranet content providers to check for problems with their links. A good tactic for getting them to do so is awarding a monthly prize for the best content.

2. Remove redundant content
As your Web authors get more confident, they'll start adding small icons that label page items as "New" or "Revised" or what have you. But novice content providers tend to duplicate each graphical element rather than use the same item they've already made for a previous page. For example, a human resources manager might have identical New icons in her press releases and benefits information subdirectories.

This duplication is common, but most Web authoring packages don't address it. That means you've got to alert content providers to the problem and get them to draw their graphics from a library of commonly used items. The library can be filed in a content provider's subdirectory or a shared directory.

Things are more complicated if content providers are copying graphics from one page to another in an authoring tool and then saving the page with the copied image. The authoring tool usually will ask for a file name to save the graphic under and the users, not understanding that they already have a copy of the image saved, will save the file under a different name.

For example, our HR manager writes a press release, then calls up a benefits information page so she can copy a New icon, stored as binew.gif, and place it and the press release on a new page. When saving the page, she saves the New icon as prnew.gif - same content as binew.gif, different file.

This is harder to fix. It requires you to scan for all image files and then check those of the same size to see if they do, indeed, have the same content.

You'll need a utility such as Cerious Software's ThumbsPlus for this work. ThumbsPlus builds a database of all image files and then lets you sort the images by size and display thumbnails. This makes it easy to inspect a large number of files quickly.

3. Centralize resources
As I indicate in Tip No. 2, resources used by multiple authors are much easier to control if there is only one centrally managed copy of each. To make this work, you need to keep shared authoring components such as graphics and boilerplate text in a standard location on all corporate Web servers and index them through a set of Web pages that also explains their use.

4. Prevent content from growing old
By cross-referencing your intranet server logs with a list of documents ordered by last access date, you'll find out which documents aren't used often or at all.

Documents that aren't used often might be more useful if people knew about them or could find them more easily. For example, you might provide a link to these documents from a common, heavily used page. Or, even easier, bolster the documents themselves with keywords that more accurately describe the content. This would make the documents come up more often in searches.

Documents that don't appear to be used at all are more problematic. It's possible that the documents might not be of interest to anyone, but you can't just throw them away. Even a rarely accessed document might be crucial when needed.

You do want to weed unused, unlinked and infrequently accessed documents from the intranet, but you don't want to destroy them. Moving such documents to long-term storage, such as tape, would give you the flexibility of restoring them to the intranet if need be.

When you move the documents, you'll have to find and fix links to them using a tool described in Tip No. 1. You could point those links to a generic document that says the target has been archived and offers a form for requesting the material.

Another technique for preventing intranet content from growing stale is time-stamping documents. You can do so by examining the modified date set by the operating system for files that haven't changed in, say, six months. Alternatively, make content authors include a last-modified field in each document.

In either case, you can remove documents that don't comply or are older than a certain date.

5. Enforce standards
Your clean-up campaign has to include checking for compliance with publishing standards. If your publishing standards are well-defined, this will be easy. Requiring every document to contain the author's name and e-mail address makes it easy to find those who don't comply.

But once you find offending documents, what should you do? Much depends on how many incorrectly formatted documents you find, your authority and how much work is required to fix the problem. In general, you should check all newly added documents each day for standards compliance.