The sad state of modern word processors

March 21, 2013
Serge Huber

This post was originally published on AIIM's Expert Blogs by Serge Huber, CTO at Jahia Solutions

********

Anyone working with medium to long documents will probably agree with me on this: the currently available word processors are a long cry from being usable. If, like me, you need to write technical documentation, you’ve probably tried a few different options and none really fit your needs. And if you have to comply with a look and feel, or need to create complex documents across the enterprise, you might be forced to use Microsoft Office that really doesn’t handle these types of workloads that well. Microsoft Word for example is more useful for writing letters or resumes than it is good at writing long technical, legal or scientific documents.

When I was still a student at the Swiss Federal Institute of Technology (EPFL), I had to write some long technical reports, and at the time the recommended weapon of choice was FrameMaker (before Adobe bought it :)). I fought it initially, trying to stick with the software I knew, until I ran into a big problem with a report that was due the next morning. I had used a picture format that for some reason was confusing the word processor and it would no longer let me save my document ! I had to convert all the illustrations in my 100 page document before I could finally hit “save”, and fortunately I didn’t loose anything. So for my next project I used FrameMaker. It had a few quirks, like a steeper learning curve than Microsoft Word, and the lack of a spell and grammar checker, but it was a great tool to write structured and good looking documents. It was really great at making it easy to reuse structure and formatting across documents. As I used this software more and more, I was convinced that eventually everyone would see the light and that all desktop word processors would include the same structuring features.

One of the main issues with proper document authoring is formatting freedom. Lots of users like the almost absolute formatting liberty that Microsoft Word offers, but it often comes at a cost: documents may become impossible to format properly (once the formatting has become complex and highly customized) and keeping the same formatting across documents becomes difficult. Also, some users might confuse formatting with structure, thinking that visually differentiating elements does replace proper structure definitions, and this causes problems if the actual content has variations on formatting for the same semantic elements. This is why formatting styles were introduced in early versions of the product, but even these may introduce problems since they are too flexible. When writing content, it is often better to be able to concentrate on the content rather than the look, and leave the latter to more experienced designers. In large companies, the design is usually imposed, and this is usually a good thing since it makes it easy to maintain a branding across all documents, as well as adding any legal or security notices that should be systematic. It also makes it a lot easier to modify the design should any change in branding be required by some marketing decision, or if the global document structure is updated.

On the opposite end of the spectrum, the home user that just needs a word processor to write a letter will not want a lot of complexity, and often will not even understand the need for styles. It is of course only my opinion, but I think that spending the time to properly design and reuse styles is a good thing, even for personal use, but I’m also realistic as to the fact that some users will always want to modify things manually, and therefore a word processor such as FrameMaker will not be a good fit.

So in effect the word processor should have a way to let users use style customizations or not, and make it really easy to reuse styles across documents. Also, styles should apply on properly defined content structures (not just paragraphs or characters, but for example on sections, chapters, tables, images, table of contents or indexes), not on any unstructured content. Ideally it should also be possible to create dependencies between styles, so that if a certain style is applied, sub-styles will be applied on sub-structures. All this is not just a figment of my imagination, this is all part of the functionalities offered by FrameMaker a long time ago, and some of them have been introduced in later versions of Microsoft Word. But as FrameMaker is no longer widely used, it is now mostly in maintenance mode, and therefore only receives minor updates (last I checked it still didn't include a grammar checker nor was it ported to the Mac platform). So most users will not consider it since it is now an obsolete product.

What I’m hoping for is that someday, word processors will get smart again about formatting and structure, and make it real easy for users to do the right thing, instead of allowing them to do just anything. Nowadays, a lot of the generated content will probably be reused in some form or another on the web and having properly structured documents greatly eases the integration of the content with tools such as CMS’ or other online publication services. Maybe the only way this will happen is in one of the open source projects such as Apache OpenOffice or LibreOffice, or maybe Microsoft will finally cater to the needs of long document writers ? When you look at the way modern CMS can style structured content and can dynamically assemble and reuse content, as while staying relatively simple-to-use, you really wonder why word processors can’t do the same ? I’m sure much progress can be done to properly help users generate hundreds of properly formatted and structured documents.

When you think about it, it is mostly a question of the user interface. Adding structure to a document should be as simple as clicking a button to mark the content, or in some cases the word processor could even try to be smart about it and automatically propose structure elements as the user type is typing (such as title, sub-title, …). Also, since a lot of users are still interested in getting a good looking result, it should by default look good, so that users will actually be encouraged to use such markings. I don’t think any of this is rocket science, but it would go a long way towards improving the handling and generation of text-based content, which is still the majority out there.

Serge Huber
Author : Serge Huber
Back