Tagging

Last night on Twitter I asked “Looking for a way to tag topics in an authoring platform. Not part of Author-it, does any app do it?” and a few helpful people gave me suggestions.

I later clarified and thought I’d expand on that here. My clarification was that “Re tagging: we want to tag topics in the authoring tool so whatever the output, the user can filter on tags eg, v1.0/email or v2.3/document”, it was at this point I realised I needed more than 140 characters to explain what I was looking for…

Our product is in the throes of moving to a modular/component delivery. We will no longer have one installer to install one big product with everything in it, instead you will likely (it’s still being figured out) get a core application which you can then enhance with additional modules.

The challenge to our team will be to document this product in a way that is useful to both the people who might use the application and to those people who customise the application for specific customer needs (which may or may not involve customising a module or creating a bespoke module for a customer).

One idea that may help us in this would be the ability to tag all of our content with module/component/functionality tags, and a version tag. There may be more tags required but we can figure those out later.

However, the authoring product we currently have, Author-it, doesn’t offer this functionality. We have looked into a way of manually doing it by hacking XML files but it’s less than ideal.

So what I was asking for, badly, on Twitter was a native application that includes tagging of topics on the authoring side, with the tags then available in the output. Ultimately we want to be able to build dynamic sets of information AND allow the users to change the view of the information based on the tag(s) they’ve chosen as well.

MadCap Flare comes closest as it at least has a notion of ‘tags’ in the product but not sure if that allows a dynamically built output, and Robohelp allows tagging of returned search results but no way of producing that content dynamically.

Does such a product exist? It  feels like it should, that it’s not a million miles away from what a lot of technical writing teams would want… but I don’t think it does. Prove me wrong!


 
 
 

10 Responses to “Tagging”

  1. Peter Grainge
    21 February 2012 at 13:58

    Gordon
    Two options that I see.
    Merged Help where you have a shell parent and various child projects. One child for the core content that all users have and then further child projects for each module. The shell and the parent get installed at all locations and the child projects get added as required. When the user opens the help they get whatever children are present and it runs as one help project as far as the user is concerned.
    The other is Dynamic User-Centric Content. All the help is present but a dropdown allows the user to select the area they want and searches etc are limited to that content.
    Both are described in detail on my site.
    I am not sure what you mean by RoboHelp allowing users to tag search results.

  2. Noz Urbina
    22 February 2012 at 09:57

    Ok… you know what I’m going to say.

    DITA was made for exactly this, and includes mechanisms for delivering filtering selectors to the front end so that different audiences can be provided different selection options on the same content set. Also, this scenario (products being broken up and re-bundled) is what pushed Adobe to move to DITA, among many others.

    Why not use a technology expressly designed for the task at hand rather than pushing Author-it or Flare to do DITA-like?

  3. Noz Urbina
    22 February 2012 at 10:00

    Afterthought – this is a classic case of ‘Content Agility’. when I harp on about content that has agility, or Ann Rockley et al say ‘Intelligent content’, it’s about adding the intelligence to the content rather than the tool. It then makes the content able to inherently do these funky things rather than moving our content from vendor to vendor hoping for they’ve got the right features, and investing in their proprietary methods of doing something that can be done in a standard way across many tools.

  4. Gordon McLean
    22 February 2012 at 12:00

    Hi Peter,

    can’t see anything on your website about dynamic user-centric content… can you provide a link?

    Noz – yeah I know DITA will help with this, but, again, vendor support isn’t great. I can’t afford to custom build a system, so looking at vendors. Author-it don’t seem to do it, Robohelp doesn’t offer it on the authoring side, but I’ve used web CMS that have all of this!

  5. Michele Imperiali
    26 February 2012 at 21:07

    Hi Gordon,
    I work at Author-it and if I understood what your requirements are, I’d say that the solution for you is using variants in Author-it Aspect.

    Here there is a short introductory video that should give you an idea of what you can achieve using Aspect: http://www.youtube.com/watch?v=HJ3PDvb_mT8

    You can read more about it here: http://www.author-it.com/index.php?page=dap

    I’m usually answering to questions on Twitter but I probably missed your one, my apologies for that.

    Let me know what you think about Aspect and if it’s what you were looking for.

  6. Ellis Pratt
    7 March 2012 at 10:45

    Maybe take a look at the labels feature in Confluence. You can create lists based on these labels and filter content on labels as well.

  7. bgg
    11 March 2012 at 19:40

    Labeling in Confluence is completely flat and a really weak feature as label changes are not tracked in the history of Wiki pages. It’s impossible to find out who did change what in labels, and the label lists offer only a limited range of filtering. PDF output of Confluence is hardly customizable, and pages cannot nicely be exported or imported, so it’s useless for online help outside of Web usage.

    I have quite some experience with Madcap Flare as well as with CMS like Drupal and WordPress, and Confluence has the weaknesses of Wikis combined with the limits of payware :/ Working with it in a large project now and it’s a PITA.

  8. George Lewis
    28 March 2012 at 09:14

    Gordon,

    I’ll go with Noz on this one.
    A client of ours has a web portal with infinite different combinations of capabilities that can be assigned to an one user.
    We deliver the UA as a Support Center using our assist-ua platform. We author in DITA which allows us to assign tags to each topic – as you suggest – on a number of dimensions, and import this into the assist-ua platform. The platform is aware of the capabilities the user has, and displays content accordingly.

    Although I have not come across the term before, this does seem to be what Noz refers to as “Content Agility” – the platform dynamically displays the content based on the user profile.

  9. Mike McGuire
    14 April 2012 at 02:47

    I was about to suggest a wiki solution — specifically Confluence, which we are evaluating — but bgg took the wind out of my sails. We were hoping that a mixture of API coding on our part and free/commercial plug-ins could allow Confluence to handle a publishing we are moving towards that mirrors the one you describe, Gordon.

    I know that vendor support for DITA is weak, but its content tagging is robust on the authoring side, and, even if the built-in filtering doesn’t meet your needs on the output side, you could hack together some XSLT to spit out the customized XML content, then use something like dita2go or DITA-OT to create the output.

    Which brings me to another question: what is the format(s) of your deliverables? Before we went goo-goo over Confluence (and wiki-culture in general), we had divided up our efforts into 1) authoring of content, 2) storage of content in meta-data tagged XML (probably DITA), and 3) deliverable production using another suite of tools (html wrapped in FAR-frameset and PDF). It seems to me that DITA would handle the first two areas for you — especially the meta-data. The trick is the third area of functionality.

  10. Dmitri Popov
    20 April 2012 at 13:15

    I once had a similar task and it in part then led to creation of my own HAT :) But frankly speaking I use far more simple approach – each topic has a “condition” based on project variables and if this condition turns false – the topic is skipped from output. I can group variables into “presets” of values and then switch them before clicking “compile” thus making a help file with proper set of topics targeted at specific audience.
    Not sure that this will solve your problem, but a lot of tech writing tools have similar functionality, e.g. there may be another way, without using “tags”.