Sturm und DITA-Drang at tekom
This year’s tekom/tcworld conference reinforced the ongoing doctrinal chasm between North American technical communication and German technical communication.
I am speaking, of course, of the proper role of DITA in technical communication. If any.
Executive summary: There may be a valid use case for German non-DITA CMS systems, but vendors are shooting themselves in the foot with factually inaccurate information about DITA as a starting point for their arguments.
The program this year included several presentations, in both English and German, that provided the German perspective on DITA. They included the following:
- Marcus Kesseler, co-founder of Schema, 5 Reasons not to use DITA from a CCMS Perspective (English)
- Intelligent information panel, Marcus Kesseler, co-founder, Schema; Kris Eberlein, Eberlein Consulting and chair of the DITA Technical Committee; Eliot Kimber, Contrext Solutions and a member of the DITA Technical Committee; Torsten Kuprat, Acolada; moderated by Dr. Wolfgang Ziegler
- Uwe Reissenweber, CEO of Docufy, Was versteht man eigentlich unter DITA? (German, no presentation summary at the link)
- Intelligente Information panel, Stefan Freisler, co-founder, Schema; Sissi Closs of Hochschule Karlsruhe; Sven Leukert, SAP; Andreas Klüter of Empolis; and myself; moderated by Jürgen Muthig (German)
The DERCOM effect
We also heard a great deal from a new organization, DERCOM. Founded in 2013, this organization is an association of German CMS manufacturers (the acronym sort of works in German) and includes Schema, Docufy, Empolis, Acolada, and three other member companies.
DERCOM has released a position paper entitled “Content Management und Strukturiertes Authoring in der Technischen Kommunikation” or (as you might expect) “Content Management and Structured Authoring in Technical Communication.” This document is available both in German and in English translation. Unfortunately, the link seems to be obfuscated. Go to the main DERCOM page and find a clickable link under “News.” DERCOM member Noxum has a direct link for the German version.
Uwe Reissenweber explicitly introduced his presentation as providing the official position of DERCOM.
With refreshing candor, Uwe Reisseweber introduces himself as a lobbyist for DERCOM and will provide their perspective on DITA #tcworld15
— Sarah O’Keefe (@sarahokeefe) November 12, 2015
Note that he used the German word “Lobbyist,” but perhaps “advocate” would be a better English translation than “lobbyist” since the latter term is so loaded with negative political connotations. Marcus Kesseler said that he was not speaking for DERCOM but rather for Schema in his individual presentation. Here is what I observed across the various presentations:
- There was remarkable uniformity in the statements made by the various DERCOM members, even when they said they were speaking for their employer rather than the association.
- There were a number of talking points that were repeated over and over again.
- The descriptions of DITA were so technically inaccurate that they destroyed the credibility of the speakers’ entire argument and made it rather difficult to extract valid information.
For example, Uwe Reissenweber asserted that the DITA specialization mechanism, if used to create new elements (as opposed to removing them), does not allow for interoperability with other environments. That is, once you create new, specialized elements, you can no longer exchange your content with other organizations. This statement is technically inaccurate and betrays a fundamental misunderstanding of specialization. When you create a new element (for example, a warning), you base it on an existing element (for example, note). Because DITA maintains inheritance information, a downstream user would know that the warning element is based on note and can process it as a regular note element via a fallback mechanism. This is a critical—and unique—feature of the DITA architecture. Marcus Kesseler asserted that vendor lock-in with DITA-based content is no different than a system (like his) that uses a proprietary content model because so much of the business logic is tied up in the CMS rather than the content model. This overall accuracy of this statement depends on how tightly business processes and other information are bound into the CMS. But it seems indisputable that it would be easier to move DITA content from CMS A to CMS B (with any attendant business logic issues) than it would be to move XML Flavor A from CMS A to XML Flavor B in CMS B. In the second case, you have to move all of the business logic and worry about possible incompatibilities between XML Flavor A and XML Flavor B. “You can’t learn specialization in an afternoon.” This is a completely true statement from Uwe Reissenweber to which I say, with great professionalism, “SO WHAT??” Surely we are not advocating the idea that anything that takes more than an afternoon to learn cannot be worth the effort. After hearing these statements and others (see my Twitter feed for increasingly agitated coverage), it becomes difficult to take any of the presenters’ arguments seriously. And this is unfortunate, because I do want to understand their position. Kesseler, for example, displayed a chart in which he made the case that business logic is embedded either in the CMS or possibly in the DITA Open Toolkit, but not in the core DITA topics.
I am extremely confused but I think that Kesseler is saying that business value is in CCMS and not DITA. #tcworld15 pic.twitter.com/dfiqh5uaDh — Sarah O’Keefe (@sarahokeefe) November 10, 2015
His Schema co-founder, Stefan Freisler, believes that only 5–10% of return on investment realized from a CMS system is in the content model. Instead, the vast majority of the value resides in the workflow layer.
These are interesting points and worthy of further discussion.
DITA support in DERCOM CMSs?
Eliot Kimber, who has a lot more patience than I do (also, I had a scheduled meeting), stayed through a heated post-presentation Q&A with Kesseler. Kimber had this to say in his trip report:
It was an entertaining presentation with some heated discussion but the presentation itself was a pretty transparent attempt to spread fear, uncertainty, and doubt (FUD) about DITA by using false dichotomies and category errors to make DITA look particularly bad. This was unfortunate because Herr Kesseler had a valid point, which came out in the discussion at the end of his talk, which is that consultants were insisting that if his product (Schema, and by extension the other CCMS systems like Schema) could not do DITA to a fairly deep degree internally then they were unacceptable, regardless of any other useful functionality they might provide.
This lack of support is another starting point for discussion. (I would also note that it’s often not the Evil Consultants, but rather the Opinionated Clients, who are insisting on DITA.)
With a proprietary content model, you are putting your trust and a good bit of your ability to execute in the hands of your CMS vendor. Provided that the vendor does a good job of introducing new features that meet your needs, you could have a long and mutually beneficial relationship. But what if your vendor starts to falter? What if they are acquired and change their strategy to something that doesn’t meet your requirements? DERCOM members are asserting first that they are better at adapting information models to the needs of their clients and second, that the content model provides only a small part of the value of the CMS.
Do you throw your lot in with a vendor, their release cycle, their software development/R&D efforts, or do you choose to rely on a standard and therefore rely on the OASIS technical committee, with all of the advantages and disadvantages of the committee-based standards-building process?
If the content model is largely irrelevant to the CMS functionality, why not just have the best of both worlds and support DITA inside the Very Superior DERCOM systems? Some of the vendors are doing just that. Empolis supports DITA both in its core CLS offering and in a new, low-cost SaaS system that is under development.
It remains as an exercise for the reader to understand why the other vendors are not following suit. Eliot says this:
DITA poses a problem for these products to the degree that they are not able to directly support DITA markup internally, for whatever reason, e.g., having been architected around a specific XML model such that supporting other models is difficult.So there is a clear and understandable tension between the vendors and happy users of these products and the adoption of DITA. Evidence of this tension is the creation of the DERCOM association (http://www.dercom.de/en/dercom-home), which is, at least in part, a banding together of the German CCMS vendors against DITA in general, as evidenced by the document “Content Management and Struktured Authoring in Technical Communication – A progress report”, which says a number of incorrect or misleading things about DITA as a technology.
During the German-language Intelligente Information panel, Professor Sissi Closs pointed out the importance of DITA as a multiplier. She mentioned that widespread adoption of DITA would lead to a network effect, in which the standard becomes more valuable because more and more people are using it and therefore training, support, community, and qualified employees are more readily available.
Some statistics
In North America, DITA is the clear leader in XML-based content work. We estimate that at least 80% of structure implementations are using DITA. The equivalent number for Germany is in the 5-10% range, based on research done by tcworld.
This chart was shown in Reissenweber’s presentation and attributed to tcworld as of 2015:
Here is my English translation. In each grouping, the upper bar is for software companies and the lower bar for industrial companies.
Scriptorium’s perspective
For Scriptorium consulting projects, we use a standard methodology with roots in management consulting. In the assessment phase, we develop the following:
- Business goals for the organization
- Content strategy to support the identified business goals
- Needs analysis
- Gap analysis
- Requirements
- Implementation plan, ROI, and budget
The decision whether or not to use DITA is generally made in the requirements phase. Most North American companies, at this point in time, assume that DITA is the path of least resistance because of the large numbers of authoring tools, CMS systems, and supporting systems (like translation management and content delivery platforms) that support it.
DERCOM companies will have difficulty making inroads into this market without an affirmation that they can provide DITA support. Any advantages that they might have in workflow or editorial management are irrelevant because they will be ruled out as a prospective vendor by the DITA requirement. Additionally, most of these vendors do not have much presence in North America, so support, training, and maintenance are a risk.
Classic case of disruption
In Germany, the DERCOM vendors are clearly dominant at this time. However, their continued insistence that their technology is superior and the upstart DITA-based options should be ignored follows the classic pattern seen with disruptive technologies. When a disruptive technology offers a clear advantage that is different from the main feature set of the incumbent approach, the incumbents have much difficulty responding to the new challenge.
In the case of DITA, the disruptions are in the following areas:
- A wide variety of options at each point in the tool chain (authoring tool, content management system, translation management system, content delivery portal, and others)
- Incremental implementation. Because DITA can work both on a file system and on a CCMS, organizations can opt for an incremental implementation, where pilot projects are built and validated on the file system before CCMS selection begins.
- Open standard. Interested parties can participate in the standards-development process through OASIS. Vendor assessment is based on conformance to the standard, which makes evaluation across vendors easier for the buyer. Content exchange is easier to implement for the buyer.
- The ecosystem of DITA architecture, software vendors, consultants, open-source community, and more. (Take a look at the DITA projects available just on GitHub.)
NOTE: Translations from German are mine. If the original authors would like to provide a better or alternate translation, please leave a comment. My tweets of German presentations are real-time translations by me. I apologize in advance for any errors.
Additional reading (if by some miracle you made it this far):
- DITA & Deutschland (German, redakteuse.de)
- DITA or no DITA: Answers from 20 Years of Business Experience or the German original (SCHEMA blog)
- DITA—friend or foe? (Axel Regnet)
Marijana
Woohoo, I read it all 😉 And I especially like the disruptions listing – these are the things that are really unique and in my opinion great about DITA. Maybe DITA advocates need to emphasize more on this so that one can’t just simply say “Yes, but we’ve been doing this since 20 years now. It’s not new.”.
Sebastian Göttel
Thank you for your summary and your differentiated point of view. It seems blog posts are better suited for this kind of discussion than twitter tweets.
It is interesting to see that Marcus Kesseler’s arguments are dismissed because of FUD. But your counter-argument – starting at “With a proprietary content model, …” – is from my point of view just FUD.
I would like to point out that most Dercom members have DITA experience and do offer DITA-based solutions. But we see that non-technical persons often times confuse the DITA information model with CCMS functions. If the peaceful consultants were to compare CCMS functions, they may discover that advanced features in most Dercom CCMS, not seen in DITA CCMS, are more valuable to potential customers than scoped keyrefs or other DITA particularities (which are still not standardized).
DITA is not a CCMS Uber. It is not disruptive but rear-facing by its emphasis on the technical side of an information model while there are already solutions on the market which are much more flexible and are hiding a lot of technical complexity from the end user.
Like you have said in your own blog – https://scriptorium.com/2011/11/orthodoxy-versus-nuance/ – you need to have a strong business case for implementing DITA. Just because there is some free software and an open standard, is not a business case. SGML or Docbook are also open standards and there is a lot of software to be found at Github. If you were to spend real money on a DITA project than you should weigh your options. A commercial off-the-shelf CCMS with an easy to grasp, flexible information model might better suit your business case (if you’re not trapped in FUD ;-).
My feeling is that this is a political debate because who wins the argument, wins the business. This is true for both sides. So do not expect an objective discussion. And we must leave the decision who’s right or wrong to the markets.
Larry Kunz
Sarah, thank you (danke!) for this. I saw the battlefield reports as they came in, in real time, from the DITA wars at tekom/tcworld. But I needed this article to elucidate all of it. There are still a couple of things I don’t understand. I realize that each of these could be a whole ‘nother blog post, but here goes:
1. What are the DERCOM vendors’ selling points, other than being dismissive toward DITA? Are they touting improved workflow? Greater versatility in terms of data formats? Something else? “We don’t support X, so we’re going to convince you that X is insignificant” is an odd basis for a marketing strategy. So I’m assuming their message must go beyond that. (Perhaps some of the answers are in the DERCOM position paper, which I haven’t yet read.)
2. This state of affairs, in which DITA adoption for structured authoring is 80 percent in North America and 5-10 percent in Germany, didn’t just happen last week. It’s been building up for a long time — and many observers, including you, have remarked on it. But why? Is it because the German CCMS vendors are so highly influential in their home country? Or are other factors at work?
Sarah O'Keefe
> 1. What are the DERCOM vendors’ selling points, other than being dismissive toward DITA?
You’ll have to ask them. What I heard was “we are better and DITA is nothing new.” I assume there is more.
> 2. …why are DITA adoption rates different
I think that is the key question. I’ve seen some discussion to the effect that German CMSs are purpose-built for heavy machinery docs whereas the North American market is focused on software docs. I’m not too sure that explains the huge divergence. I do wonder whether DITA, having more or less conquered in North America, will now start to expand more into Europe. But the business case for switching from already-established custom structure is going to be tough for many organizations. What they have is probably good enough, even if you subscribe to the argument that DITA is potentially better.
Ellis Pratt
Were the German speakers making a distinction between “baked” and “fried” content – baked being where the rules are coded into the CMS and the writers basically fill in forms, fried being the rules are embedded into the content itself? The latter requiring writing discipline from the content creators who understand the semantic rules.
The hype from some in the DITA camp can affect its credibility. Some of the claims on the uptake of DITA are based on some fairly shaky evidence.
Scriptorium’s methodology is a sensible basis for making a decision 🙂
Sarah O'Keefe
I did not hear any discussion of forms-based authoring.
Franz-Josef Knelangen
Thanks, Sarah, for pointing that out. I, for my part, tried in my presentation “Creating Every-Page-is-Page-One Topics with DITA and ePublisher” (in German) to convince the auditorium to see the core DITA topics as building blocks for “presentation topics” – creating eppo composite topics or eppo submaps using the DITA “chunk” attribute.
The benefits:
1. Your building blocks are defined in the DITA spec (they can save as a minimalist’s style guide – only use building blocks defined there!), they can be validated and reused in other presentation topics
2. Mark Baker’s EPPO “theory” helps to avoid the fragmentation of your documentation in tiny information bits often seen in plain vanilla just converted legacy doc sets …
There was no general objection against that or against DITA in general in the following discussion. All participants were on the “How can we leverage the benefits of DITA and structured authoring in our projects” level.
So, maybe I was lucky or it was the time (1st presentation on last tekom day) that kept people calm 🙂
Sarah O'Keefe
I’m sorry that I missed it. I was sitting in another session, trying to absorb XML/DITA/structure-related vocabulary so that I could participate usefully in the panel discussion later that day.
It seemed that the DERCOM-related sessions were uniformly anti-DITA; there were definitely other sessions, in German, that had a different perspective.
Fer O'Neil
There was a strange vibe emanating from the DITA presentations at Tekom this year–I also heard from an attendee that shall remain nameless that the German presentations were very vendor-centric this year, which could explain the more volatile positions being taken on the issues. I won’t enter into the primary debate here, but I will add that I find the argument regarding CMS savings coming from workflow a but spurious, only in that it is impossible to ever have a true A/B comparison to know what actual comparative savings could be with a competing system–I wish it were possible however, as it would make these discussions and decisions much easier. From someone who works with the CMS model (me, as an employee tech writer), I can say that the unstructured “forms based authroing” is just making the most of the situation versus an equivalent content model. Overall, from Tekom, I heard some strange and contradictory stats (don’t they say that 74.3% of all stats are made up…) to support this or that content strategy and I didnt leave the conference thinking anyone had found the “future of techcomm” white whale.
Guy Haas
Did Sven Leukert of SAP have anything to say about SAP’s transition from its home-grown Knowledge Warehouse XML system over to DITA?
Sarah O'Keefe
He mentioned that SAP moved to DITA, but I don’t recall anything that specific…
Frank Ralf
Touché. Thanks for this great write-up and reply!
Ben
Wait … you had that much content of the conflict and you didn’t make a parody DITA specialisation for it?! 😉
Jeanine Shepstone
I was at Marcus Kesseler’s 5 reasons not to use DITA talk at tekom/tcworld. In the presentation, he seemed to measure DITA without a CCMS against other systems with a CCMS. This would of course be naughty, since it is not comparing like with like. At his standing-room only talk, this comparison raised the ire of several otherwise mild people. Not to mention the emotional level of a rather cerebral conference. I started to wonder whether, if a fist fight erupted, it would be a first for a tekom comference, and whether such footage could go viral and make millions ….
Naughty comparisons aside, IMHO Kesseler had several good points. His contribution to the conference proceedings PDF contained the best summary I’ve ever seen of a corporate content department’s tasks. The table in his slides analysing where requirements were met (XML, DITA and/or CCMS) is serious food for thought for anyone thinking of investing in a DITA-CCMS system. One can always argue about whether his scoring was fair. However, even if it isn’t, the list is a serious contribution to evaluating DITA-CCMS systems. For me, the central question for a technical documentation system is where it adds value. The answers include: easing translation and translation memory (XML), publication without reformatting after translation (XML and a style sheet), only translating changes (CCMS), version handling (CCMS), revision handling (CCMS), reuse (content model and CCMS), and managing images (CCMS). There is some business value in the content model, but it seems to me that most of the business value lies in the CCMS. For that reason, as one of Kesseler’s slides said, ”Even with a DITA-based CCMS, you would incur a significant amount of vendor lock-in”.
On Kesseler’s point ”How DITA deals with the number of files explosion”, I won’t claim to understand all the intricacies of managing files in DITA. I know enough to know that it can be a challenge. We’re currently looking for solutions to image translation, since putting numbers in flowchart boxes is not a nice thing to do to our users. In both cases, I think that Kesseler makes a contribution by understanding the business problems.
I’m no database expert, but Kesseler seemed to be making some good points on the unnecessary increase in incidental complexity caused by DITA’s XML-first approach (rather than a database-first approach).
The guys who developed the DITA standard scaled several massive mountains (just to mention a few: reuse, extensibility, and they inspired the DITA Open Toolkit). I can understand their frustration and disappointment when they’re told that they’re only at base camp, and that Everest still hasn’t been scaled. However, for real world tech doc creators, what counts is being at the top of Everest, and there’s similar disappointment when we realise that we need a component content management system to get there. My hope for DITA 2.0 is that there is a willingness to take Mr. Kesseler’s points seriously. Like Tenzing Norgay and Edmund Hillary, I hope that DITA 2.0 embraces the DITA-CCMS design challenges, and empowers thousands of tech doc creators to plant the DITA flag on top of the highest mountains every day (without the associated altitude sickness, avalanches, freezing and death).
Stefan Freisler
The slides of Marcus Kesseler can be found here: http://de.slideshare.net/MarcusKesseler/5-reasons-not-to-use-dita-from-a-ccms-perspective.
@Sarah: You wrote: “His Schema co-founder, Stefan Freisler, believes that only 5–10% of return on investment realized from a CMS system is in the content model. Instead, the vast majority of the value resides in the workflow layer.”
I believe that the majority of value of a CCMS resides in the functions listed on slide 13-14. ‘Workflow’ is only one area of these.
Sarah O'Keefe
Thanks for the clarification/correction!
Markus Abt
Thanks for your informative summary.
As you mentioned, not all German CCMS vendors refuse to support DITA. A further example would be TIM from Fischer Computertechnik FCT AG, also with decades of history in the German market, like some of the DERCOM members’ systems. TIM provides native DITA support, side by side with ongoing support for its proprietary XML structure, leaving the customer the choice.
In the statistics section, the comparison of >80% DITA coverage in structure implementations in North America and 5-10% in Germany according to the Tekom chart does not hold: The Tekom chart states that 77-86% are currently not using or planning to use DITA. This number includes companies not using structure implementations at all, but instead use unstructured FrameMaker, InDesign, Word, Excel, Powerpoint, whatever.
On the other hand, the Tekom chart shows an important fact: While about 7-10% (of all companies in the mentioned sectors) are currently using DITA, another 7-13% are planning to introduce or switch to DITA. This clearly indicates a high level of interest in DITA in Germany, and promises a rapidly growing market for CCMSes which support DITA.