HCS Knowledgebase

When I worked in the HCS group at HP, I found a simple and efficient way to organize the paper copy of our knowledgebase (KB).1 It required much less work and paper to update than the alternative.

So that the techs could find the printed-out articles in the binders, I planned to make tables of contents and indices by client and article title. The trouble was the page numbers those needed. If I used book-style page numbers, I would have to renumber at least one 2″ binder full of pages, and then correct the index and contents by hand every single time an article changed length.

Could I get the computer to do the numbering, tabling, and indexing for me, by copying the articles from Internet Explorer (IE) into one massive MS Word document? Yes, but when I thought about it a little more I realized it wasn’t worth it. It would be a huge, complex document full of tables and images. Documents like that take forever to open and work with, crash more often, and go corrupt more often. I would have to reformat the article titles as Word headings to have them recognized by the table of contents, and thereby risk deleting the wrong paragraph mark and spending ten minutes fixing the problem.2 The procedure to add or change an article would then be:

  1. Copy article into Word document.
  2. Add heading so that it will be included in the TOC.
  3. Refresh table of contents and index.
  4. Reprint all changed material, even if only change is to page number. (Up to approx. 500 obsolete pages per week.)
  5. Hole-punch all changed material.
  6. Add changed material to binder.
  7. Remove old material.

This seemed like more work than my manager or I would like, so I thought some more about what IE could and couldn’t do for me. It couldn’t make tables, indices or page numbers, but it did paginate within the article (e.g. “page 2 of 5”). So I didn’t need to number pages. I needed to number articles.

At this point my classical education came to my rescue. Every ancient text has an internal numbering system used in the various printed editions, so that no matter which version you have, you can pinpoint the desired passage. The Bible’s book, chapter, and verse system is the best known of these. So I borrowed that handy wheel lying over there in the library aisle for my own project. It can no longer be said that a bachelor degree in Humanities isn’t useful in the real world.

I assigned each article a client-and-serial-number ID code such as ERC-42, and made a spreadsheet of codes, clients, and article titles, to be the source for my tables of contents and indices. I labelled every article with its ID code by hand. The initial compilation still took a day or two, but to add articles after that was simple:

  1. Print article from IE.
  2. Label its pages with its ID code.
  3. Update and reprint table of contents and index. (Look ma, we only killed seven pages worth of trees instead of a whole ream!)
  4. Hole-punch handful of pages.
  5. Remove old versions from binder and install new ones.
  1. The master copy was on the intranet, but after a power outage disrupted customer service, management decided they needed a copy on paper for emergencies.
  2. I don’t know if you’ve ever seen it happen, but sometimes when you delete a paragraph mark in Word, the paragraph it belongs to completely changes style. What’s happening is that the styling is stored in the paragraph mark, so when you delete it the paragraph switches styles to whatever is stored in the next paragraph mark. This is often the Wrong Thing To Do.