🚀 go-pugleaf

Hi all

For about the last 8 months, we've been using conditional text to manage a several-thousand-page software reference for (now) 7 product variants with (now) 10 writers. Like Sean, we implemented it on a tight deadline. That meant we didn't think it out well enough. The number of products and features keeps growing, so the number of tags keeps growing, and we're now up to (gulp!) 40. It isn't working very well at all, so one of our projects this year is to simplify or abandon it - either to reduce the number of manual variants, reduce the number of conditions, or use some other way of managing the chunks of content. We don't have the budget for a fancy management system, so that'll be a bit limiting. Lots of text insets is a possibility we'll be looking at.

I don't really think the number of conditions is the problem by itself. One of the biggest problems for us is that we have 2 fundamentally different "types" of conditions: product-based ones and functionality-based ones. They overlap. In some cases we can only use product-based conditions, but in a lot of cases either works, and we haven't applied them consistently enough.

So from our experience, I'd say:

* plan your conditions well and very thoroughly. Use only one scheme. And don't mix include and exclude conditions.
* train everyone thoroughly in how to apply them. Create exercises, using the trickiest examples you can find.
* after you've applied the conditions, test each variant to make sure the text displays correctly
* apply the conditions mainly to paragraphs. The smaller the unit, the harder it is to read. If you have a source paragraph with 6 different tags in it, applied to a word here and a sentence there, it's practically impossible to get them right in the first place or to ever edit that paragraph.
* if you'll have to make more than about 1/3 of a chapter conditional, don't single-source that chapter. Make multiple versions instead.

If anyone has suggestions on how to manage something like our system better, I'd be very grateful!

Cheers, Rebecca