Julio,

What a very interesting concept. You are correct about marketing, they do reuse information quite a bit. If, as a writer you need information from marketing, it can be all over the place.

One of my issues with DITA is not necessarily using the tools because I can figure out the tools, I have trouble finding training I can afford or training the company will pay for. When it comes to markup, style sheets, and DTD/schemas, I firmly believe that I need to understand the parts you can't see. As an HTML coder, I understand markup and I understand the concepts behind the DTDs and Schemas, I want to understand better how to read them and even write them.

As a new technical writer back in the 1990s, I learned early on that you need to have an understanding of how to make things work outside of the tools. Robohelp in the early 1990s made me an MS Word expert FAST.

Hi Ginger,

Thanks for stopping by.

Fortunately, some of the learning curve, in terms of style sheets may not be too bad if your output is HTML (which is what marketing leans on these days). That's just a matter of applying what you know about CSS to the output created from DITA source and modifying them to meet the needs. (Then using the new CSS when processing the next generation of content.)

Reading and writing the DTDs and Schemas are a bit of a challenge and the documentation is not necessarily the best, but there is some documentation that can help, especially if you look at some of the things Eliot has written about specializing DITA which are readily available. (I think some of them are even on dita.xml.org.)

As for affordable training, I guess it depends on what you're looking to find. I think that one of our partners, Suite Solutions, does have a bunch of web-based offerings that may fit the pocket book well. Take a look.

Julio,

It is always tempting to take a tool that has proved useful in one area and say, this is great, let's use it for everything! But when you scale a system up to encompas more uses, you must either make it more complex or make it more generic.

If you make it more complex, it becomes a less attractive option for every individual user who does not need all of its complexity.

If you make it more generic, it becomes less efficient for every individual user because less structure means less automation.

There is a good reason that pubs people do not like using Word for tech docs, and there is a good reason why other office workers would not even consider using FrameMaker for memos and reports. The best tool for any job is the one that best addresses the specifics of the job without any overhead or extraneous features. To achieve that, tools need to be specialized to their tasks.

I'm no big fan of DITA, but I think it would be well advised to stick to the sphere to which it belongs rather than trying to engulf all content. The most likely result of the attempt to encompass everything is that it's reputation will suffer as it disappoints people who are forced to use if for tasks for which it is not optimal.

Hi Mark,

There is a third option: you can build an ecosystem that meets the needs of all users. There are many environments now that help those who are not XML savvy to create DITA topics that are valid and look like Word to the non-tech pub individual. This allows the authoring and reuse of content (maybe with the help of tech comm folks) to create the information products needed. The benefit is that the casual user can search for the sort of content they need and just reuse it if it works for them. As they get more comfortable with how the system works and get more aids to help the produce what they need, they can see the long-term benefit when it comes to consistency and accuracy.

The status quo doesn't necessarily mean that it's the best methodology.

Hi Julio,

There is a third way, but that isn't it. What you describe is simply the second way: becoming more general. Which editor they use is not the point. The point is, does the object they create have structure specific to their task which makes their task more efficient? If someone creates something in Word, or any other WYSIWYG editor, they will be guided by its appearance alone. (After 25 years of trying, we still have not persuaded people to use styles!)

They will create a generic object without specific boundaries. The thing they create will not always be findable, and not always reusable if found, and the whole finding and reusing process will often be more cumbersome than simply writing what you need. The result will be a vast and unmanageable collection of misshaped topics that is full of redundancy and duplication. This is not speculation; it is observation. People will take the shortest route to getting their own task accomplished. A structured system must be specific to their task and make their life easier or they will either circumvent it or use it in a non-structured way.

There is a third way, and it the one that David Weinberger describes in Too Big to Know: let each group implement the structures that make their own work more efficient and then use domain to domain mapping at the enterprise level to make the work of other groups accessible without forcing them to use a single common format that is not optimal for their specific function.

Hi Julio,

You may have seen this already, but JoAnn Hackos recently posted an interesting case study on the DITA Awareness Group on LinkedIn, in which she concluded by saying, "...With these tools now available, we are moving DITA out of tech pubs to the enterprise where there are lots of buisness cases for single-sourcing."

I thought this had some relevance to your post. That is, do you think DITA might naturally start to outgrow tech comm departments in industries and companies where a strong business case can be made for its use outside of traditional tech pub boundaries? (And, I guess to echo some of Mark's concerns, is DITA well suited to that kind of expansion?)

I'm with Mark & Jason on this one: I've seen too many "misshaped topics ... full of redundancy and duplication" after people had a structure and a tool foisted on them, when they just wanted to put some information down. And the tool really doesn't matter, I've seen individually helpful topics coagulate into useless gobs of knowledge in corporate wikis, in FAQ databases, in corporate user forums, etc.

In my experience, you can more easily make the case and get the budget to create another knowledge repository than getting contributors comply consistently with the structure that makes it useful. Even colleagues from one domain, even writers have a hard time with that. And I've yet to see a working, cross-departmental example.

Good discussion all.

The beauty of DITA is that it can be shaped to meet the authoring needs of an organization. If the tool in use is as flexible as DITA is, you can provide templates for a DITA specialization that guides the user (including comments) extremely well as to the information needed at any point in the structure. Where that fails is when an organization doesn't take the time, energy, or modicum of expense to provide that guidance. That's when the situation Kai has seen occurs.

There is nothing that can prevent information overload other than governance of the content publication stream. That, unfortunately, cannot be totally controlled by automation (at least yet). Those companies that realize they have the problem of more duplication and inconsistencies than is useful to their customers finally wind up having someone who doesn't have a political stake in the company come in and do an analysis of the problem and put forth recommendations. Then they either implement recommendations or continue on their way, much to the consternation of all involved.

What amazes me is the amount of money a company will spend to continue their current mess than to put processes in place that will, in the long run, cost them less and finally get a handle on it all.

Yes, DITA will outgrow technical communications and become part of all customer facing information publishing processes. It will happen because, at some point, companies cannot afford to continue on their current paths.

Julio,

I repeat, after 25 years of trying, we still haven't persuaded people to use styles. After far longer than that, we have not succeeded in getting people to consistently follow company style guides. Consultants can be very free with their recommendations, but it is one thing to say, "If everyone behaves like this, all will be well." It is quite another to create the conditions in which it is actually possible to have people behave in that way.

People working in corporations are under time pressure to create their own short term deliverables. They will always take the shortest route to creating those deliverables, regardless of corporate reuse or content management policies, because that is what they will always be judged on by their immediate manager, and also because that is the only meaningful measuring stick they have to assess their own work.

People optimize their own work environment to achieve their own short term goals. For any kind of structured content collection to work, it has to present an environment to each user that is optimized for them to achieve their own immediate goals.

You are not going to get that out of DITA. It has far too much overhead and abstraction to be optimal for most corporate purposes, and no structured writing system is ever going to be optimal for creating the kind of high-design layouts the marketing needs, for instance. You are not going to get optimization for every individual task out of any one tool, system, or architecture.

It is worth noting too that the notion that one need to have a common structured format in a common repository in order to achieve corporate information sharing is very much outdated. In the 90s we all believed that intelligent content was the key, and that without it, there could be no findability and no sharing. I was singing that song as loud as anyone. But we were all wrong.

Today content intelligence (the ability of software to find and interpret content in diverse formats) has proven that 90s notion completely wrong. You don't need commonality of format or of location to enable corporate information sharing. You just need a good search engine equipped with entity extraction and relevance ranking algorithms.

Structured content creation still has a place, I believe, but it is in optimizing local processes, not in ruling the enterprise.

I guess we'll have to agree to disagree, Mark. I think that, if you create the right ecosystem, you can definitely have folks work within a structure. Once that's in place, creating the required outputs is achievable and becomes more automated.

Hi Julio,

Actually, I agree with you that if you create the right ecosystem you can have folks work within a structure. I just think, based on 20 years of watching people try and fail at creating such ecosystems, that it is very difficult to do.

Now, I am the last one to say that just because something has failed in the past, you should not try again. If Apple had given up following the failure of the Newton, there would be no iPhone. But, after countless examples of the attempt being made and failing, across many industries, the person suggesting that we try again has to face the question: what is different this time, and why would that make a difference?

In the case of the Newton vs. the iPhone, the question is easy to answer: what's different is the Internet, and more specifically the wireless Internet. Networking changes the whole PDA picture fundamentally.

So if you are asking companies to spend large amounts of time and money attempting an enterprise structured content strategy, after so many failures in the past, you have to answer the same question. What's different now, and why would it make a difference?

I can't see anything particularly new of different about DITA that would make the kind of difference that is needed here. Topic orientation is not exactly new, and for most corporate purposes it seems to create overhead without producing a commensurate short-term benefit.

On the other hand, I see that the rise of content intelligence has provided a working solution to many of the problems that an enterprise structured content strategy was supposed to address, thus reducing the urgency of the problem.

So I put the question to you this way: What is different now that would give us reason to think this will work this time round when it has consistently failed in the past? I ask in all seriousness, because I have been a structured content guy for the last 20 years, and I would love to see a convincing case made that the 90s vision of enterprise structured content is now achieveable.

I think that what's changing is the technology and there is more supporting DITA than supported SGML. There is nothing new or much different in DITA other than it's architecture that strongly supports reuse and specialization. That is it's key strengths. Add to that the metadata architecture (with support for arbitrary metadata and taxonomy) and you have a strong underpinning for a flexible ecosystem.

Also, take a look at what's happening today and I can only see more adoption in the future. http://www.ditawriter.com/?p=846&utm_source=rss&utm_medium=rss&utm_campaign=so-whos-using-dita&utm_medium=twitter&utm_source=twitterfeed