Conference notes from: Content Management Strategies/DITA North America 2010

Shawn Benham and Janet Ikemiya, IBM

#cmsdita2010 S.Benham: Use projects to better understand the products/solutions customers are using & the key biz goals they have

#cmsdita2010 J.Ikemiya: f2f--user groups, conferences, classes customers attend, all good places to find customer feedback participants

#cmsdita2010 J.Ikemiya: 3 ways to get feedback participants: use contacts who have estab. rel'ships w/ customers; face2face meetings; email

#cmsdita2010 J.Ikemiya: let everyone know you are looking to estab 1+ rel'ships w/ customers--they can keep an eye out for good candidates

#cmsdita2010 J.Ikemiya/S.Benham/IBM: define & formalize your customer feedback program

Pushpinder Toor, PTC

#cmsdita2010 P.Toor: komatsu case study is on target with Aberdeen study. ///DM us! We can share those details, and intro you to komatsu too

#cmsdita2010 P.Toor: Ingersoll-Rand/Club Car reduced delivery lifecycle 80% - huge impact for customers. "only wait a week not a quarter"

#cmsdita2010 P.Toor: keep content all in one place. Automated, associated.

#cmsdita2010 P.Toor: want to reduce localization? Animations for tasks; illustrations replacing content.

#cmsdita2010 P.Toor: want to reduce localization costs? A picture is worth 1000 words! Take out 1000 words!

#cmsdita2010 P.Toor: reuse eliminates redundant translation and authoring tasks

#cmsdita2010 P.Toor: DTP can be up to 40% of localization cost (Aberdeen)

#cmsdita2010 P.Toor: monolithic documents also require a DTP formatting pass. (more expense!)

#cmsdita2010 P.Toor: monolithic documents require complete translation (expensive!) Wed Apr 21 11:15:29

#cmsdita2010 P.Toor: localization = translation of words + adapting to a culture, local look and feel

#cmsdita2010 P.Toor: more drivers for localization? Cost of translation (38%); increasing vol of docs (30%); regulatory reqs (24%)

#cmsdita2010 P.Toor: #1 pressure leading to localization? Products launched in increasing number of markets or regions (44%)

Elliot Nedas

#cmsdita2010 E.Nedas: to drive down translation vendor cost? Reduce your costs by reducing your vendor's overhead by automating

#cmsdita2010 E.Nedas: cost breakdown of translation vendors: 50% overhead/costs; 35% translator; 15% vendor profit.

Don Bridges, DCL

#cmsdita2010 D.Bridges: fix these before conversion: multiple procedures in a single task topic, multiple levels of steps...

#cmsdita2010 D.Bridges: fix these before conversion: notes in tables; procedure in a table; page references...

#cmsdita2010 D.Bridges: <1k pages ? Convert by copy/paste. >2.5k ? Are you kidding?

#cmsdita2010 D.Bridges: book-sourced conversion is good when 80% of output topics don't need reauthoring (or not right away)

#cmsdita2010 D.Bridges: topic-sourced conversion is best, but comes at a cost--look at all the incarnations up front-- it maximizes reuse

#cmsdita2010 D.Bridges: it is cheaper to convert than to reauthor

#cmsdita2010 D.Bridges: few writers have the clairvoyance to author content thinking it will be converted in the future

#cmsdita2010 D.Bridges: reuse - people are reluctant to do it; they think it's hard

#cmsdita2010 D.Bridges: reuse is what affects the quality; consistency is saying the same thing the same way every time=>Reuse!

#cmsdita2010 D.Bridges: main DITA ROI drivers: multiple output reqs; similar product lines; translation reqs

#cmsdita2010 D.Bridges: DITA reuse is "single source authoring"

#cmsdita2010 D.Bridges: Cost in this economy? Pfft. Cost is always an issue.

#cmsdita2010 D.Bridges: DITA really does improve quality b/c of consistency.

#cmsdita2010 D.Bridges: your yeild (legacy data conversion) depends on consistency of input data; how big a paradigm shift; yr flexibility

#cmsdita2010 D.Bridges: there is no one right level if reusable size

#cmsdita2010 D.Bridges: #1 requested conversion ? Quark to XML

#cmsdita2010 D.Bridges: your customer will tell you which legacy documents to covert. Always the ones most used; longest life; & most value

S.Frederiksen

#cmsdita2010 S.Frederiksen: qa maps for missing pieces, too many pieces, and structure

#cmsdita2010 S.Frederiksen: qa topics as the stand alone items they are, even if you view them in context.

#cmsdita2010 S.Frederiksen: if you copy/paste content from legacy: you are going to XML, not DITA.

#cmsdita2010 S.Frederiksen: never write, create, or edit topics inside a large document: it will ruin reusability

#cmsdita2010 Technical track was frequently vendor showcase. Always disappointing when a learning opp turns out to be a dog-and-pony show.

#cmsdita2010 I am always amazed when demos show features of XML publishing as if it's a product/tool feature: Look content update automatic!

#cmsdita2010 R.Eason: stats from Acrolinx customers: editing time down 3-4x; word count down 4-15%; translation costs down 30%.

#cmsdita2010 R.Eason: includes reporting on whole repository & readiness-shape of content

#cmsdita2010 R.Eason: add intelligence to avoid long,complex copy edits down stream. Editors start with analyzed files.

#cmsdita2010 R.Eason: linguistic matching engine to catch phrases that would otherwise trigger translation cycles

#cmsdita2010 R.Eason: can catch ambiguity-- do you mean "to the left" or "on the left" (tightening screws)

#cmsdita2010 R.Eason: can tell difference between "check" verb and "check" noun

#cmsdita2010 R.Eason: get color coded report--domain-specific spell check; style issues (future tense issues--100 rules OOB

#cmsdita2010 R.Eason: plug-in to word, frame, Arbortext, xmetal

#cmsdita2010 R.Eason: natural language processor that acts like a copy editor. Drive down litigation risk

#cmsdita2010 R.Eason: Acrolinx is an info quality tool that adds a quality control layer at the time of content creation

Pushpinder Toor, PTC

#cmsdita2010 P.Toor: working on way to publish to SCORM as just another output format from Arbortext.

#cmsdita2010 P.Toor: questions are objects that are pulled into maps. Question/response structure.

#cmsdita2010 P.Toor: all PTC University courses are done in DITA (specialization b/c dita spec module was in development)

#cmsdita2010 P.Toor: audio file, student materials, graphics in thumb/big (one copy, LMS manages)

#cmsdita2010 P.Toor: publish course as WBT from single source. PTC courses done this way today

#cmsdita2010 P.Toor: check-in/check-out OOB in Arbortext Editor

#cmsdita2010 P.Toor: can check out top level only or also all dependent files. Easy to move around while still following rules

#cmsdita2010 P.Toor: structure tab gives view into modules "in use"; can drill down, w/o duplicate copies in repository

#cmsdita2010 P.Toor: can publish very small guide of exercises and still publish PDF of full student guide w/o printing 300pgs

#cmsdita2010 P.Toor: same single source can produce the slide set (not pptx, PDF!) get benefit of slides always in sync w/student materials

#cmsdita2010 P.Toor: published renditions can go to desktop or directly to CMS so available to everyone

#cmsdita2010 P.Toor: composition launcher builds all output formats in one-click

#cmsdita2010 P.Toor: can build taxonomy as a course developer and that's all pulled in

#cmsdita2010 P.Toor: PTC LMS includes multimedia in their delivered training content

#cmsdita2010 P.Toor: Arbortext Editor DITA resource mgr allows browse right into content repository and manage multi-map content

#cmsdita2010 P.Toor: case study Arbortext DITA + windchill to deliver PTC University training

Joe Gelb, Suite Solutions and John Allwein, Emerson Process Management

#cmsdita2010 J.Gelb: to do review and discover validation problems the easiest way to do this in our system is to publish (ouch)

#cmsdita2010 J.Gelb: there are still a lit of features we want...But that's for future (including building custom integration for TM vendor)

#cmsdita2010 J.Gelb: cost of sw doesn't change as we grow, (but cost of maintaining custom code isn't cheap)

#cmsdita2010 J.Gelb: we are looking toward integrating with PLM systems and CRM in the future

#cmsdita2010 J.Gelb: to associate TechPubs Docs to other Docs in enterprise, we built app&ui to do this (a manual process) got triggers!

#cmsdita2010 J.Gelb: pub ui will require development to change going forward. Good for corp branding guidelines, but ...

#cmsdita2010 J.Gelb: built custom publishing queue and UI to get DITA publishing in sharepoint

#cmsdita2010 J.Gelb: wanted to be able to create translation pkgs on demand--needed to build it on top of Sharepoint too

#cmsdita2010 J.Gelb: built reporting dashboard on top of sharepoint to be able to have bird's eye view of where everything is (big list)

#cmsdita2010 J.Gelb: sharepoint can't manage translation component management or workflow. We had to build all of that capability too.

#cmsdita2010 Best paper on "when to branch and when not to" is by Perforce.

#cmsdita2010 J.Gelb: also had to build branching capability into sharepoint to get capability to have configurable maps (minor revisions)

#cmsdita2010 J.Gelb: built hooks to initiate publishing tasks

#cmsdita2010 J.Gelb: sharepoint thinks of files as URLs; had to build whole mechanism around that capability just to get link information

#cmsdita2010 J.Gelb: we are also scraping info out of the files to put metadata into sharepoint.

#cmsdita2010 J.Gelb: custom software dev for expansions to basic check-in/check-out to what we needed

#cmsdita2010 J.Gelb: custom code begets custom error msgs that require audience help to resolve 

#cmsdita2010 J.Gelb: oob except for developing component management info layer & adding custom touch points to all tools, so play together

#cmsdita2010 J.Gelb: had to build in capability to get "where used" & this was a BIG challenge to getting SharePoint to work as CCMS

#cmsdita2010 J.Gelb: to get uncompromised, high-end DITA authoring & publishing, had to build the CCM infrastructure

#cmsdita2010 J.Gelb: sharepoint is not good at: supporting XML Docs, managing components & relationships to each other, aggregating into pub

#cmsdita2010 J.Gelb: sharepoint good at collab, basic ver. control on document level, metadata, Dev platform for customization

#cmsdita2010 J.Gelb: to use Sharepoint we had to build a CCMS layer to get the whole thing working

#cmsdita2010 J.Allwein: SharePoint sitting on top of MarkLogic server is front end to answering user questions.

Frank Shipley

#cmsdita2010 F.Shipley: use RDF to model links in resources so you can track & manage relationships between them

#cmsdita2010 Next up-- lunch then "We wanted a CMS. We were given Sharepoint. Turning disappointment into opportunity"

#cmsdita2010 F.Shipley: stats about content--what's used/reused where, impact, &c is all in that RDF link graph

#cmsdita2010 F.Shipley: profiling links using RDF triples in DITA: extract link metadata & build RDF link graph--can't break documents

#cmsdita2010 F.Shipley: using XML catalogs to resolve resources and do term management

#cmsdita2010 F.Shipley: if you want to reuse topics, ID must be unique across organization. Best practice: ID generated/match file name

#cmsdita2010 F.Shipley: became long product demo, and a cool one, but not what I expected

#cmsdita2010 F.Shipley: faceted search (ecommerce) -- filtered search + navigation: "my size","color"--provides guide to what exists

#cmsdita2010 F.Shipley: why classify? So you can find them!

#cmsdita2010 F.Shipley: why identify resources? So you can reference them for use & reuse

#cmsdita2010 F.Shipley: how to name (identify) & how to organize (classify) all those resources

#cmsdita2010 F.Shipley: before DITA, hundreds of files; after DITA, thousands with lots of dependencies

#cmsdita2010 F.Shipley: RDF can be an answer to your DITA nightmares

Rebekka Andersen

#cmsdita2010 R.Anderson: model of engagement and a focus on shared artifact that are developed collaboratively

#cmsdita2010 R.Anderson: understand that technology is but 1 component of this complex organizational change & communication that is CMS

#cmsdita2010 R.Anderson: focus on tools should be <10% of project impl. 50% is change management 40% is process management #cmsdita2010 R.Anderson: Lead people not projects. #cmsdita2010 R.Anderson: think about your larger organization before you start going through this process. #cmsdita2010 R.Anderson: need metrics--business plan, test plans, weekly project management plans, document the unwritten rules #cmsdita2010 R.Anderson: don't just play with a system--have daily goals, defined things to evaluate & a plan to get the info you need. #cmsdita2010 R.Anderson: watch for one-way communication channels. Need collaboration and cooperation organizationally for success #cmsdita2010 R.Anderson pres tells me we're on right track. We are highly collaborative and integrated as a company->also how we work w/cust

#cmsdita2010 R.Anderson: address culture of org separately from how technology addresses or supports the org

#cmsdita2010 R.Anderson: technology can't solve the problem or save the day. You still need to do the work and understand yr org/culture/&c

#cmsdita2010 R.Anderson: be clear with self & vendor so you're operating on the same assumptions. Articulate your plans and document them

#cmsdita2010 R.Anderson: it all comes down to perceptions of value of the technology

#cmsdita2010 R.Anderson: how innovation is communicated (channels,over time) among members of social system (Evertt Rogers)

#cmsdita2010 R.Anderson: used "Diffusion of Innovation Theory" (map stages of process) & "Activity Theory" (map context of activity)

#cmsdita2010 R.Anderson: calling knowledge transfer the "diffusion process"

#cmsdita2010 R.Anderson: I wonder how she didn't find us? We transfer knowledge to customers (anyone who will listen) constantly.

#cmsdita2010 R.Anderson: prevailing tool-focused approach to implementation. Don't let tool define you: should be the other way around

#cmsdita2010 R.Anderson: a lot more info now than 5 yrs ago, but implementations are plagued with problems

Liz Fraley, Single-Sourcing Solutions

#cmsdita2010 L.Fraley: Ingersoll-Rand Club Car "Translation savings expected to be 92K 1st year and 500K in the first 3" Cycle-time down 80%

#cmsdita2010 L.Fraley: Canadian Telecom, 'companies see what we've done [started in 1999] and they're very, very envious'

#cmsdita2010 L.Fraley: Canadian Telecom, "This is all doable b/c we went to XML and Arbortext and Documentum in 2004"

#cmsdita2010 L.Fraley: sometimes the business changes, and you don't have the opportunity to get where you really want to go

#cmsdita2010 L.Fraley: The most successful businesses for the next century will...have been based on a combination of competence & passion

#cmsdita2010 L.Fraley: Understand what specific product or service you offer, and then do it as well and expertly as possible.

#cmsdita2010 L.Fraley: Will your content stand up like Robinson Crusoe has? 18th century novel w/100s of new formats,versions,tech since

#cmsdita2010 L.Fraley: Canadian Telecom hasn't really touched stylesheets in 5 years. 1 staff supporting 250 writers. "Runs like a Cadillac"

#cmsdita2010 L.Fraley: LifeSciences company says XML publishing is "bottled gold" remains unnamed b/c HUGE advantage over their competition

#cmsdita2010 L.Fraley: Medtronic brought their translation down 10X because they went to XML publishing.

Gershon Joseph, Cisco

#cmsdita2010 G.Joseph: rule of thumb--if 50% of a topic is conditional, it's time to break it out and focus on the differences.

#cmsdita2010 G.Joseph: profiling on release leaves history in content no one cares about; also, requires you to re-analyze w/every release

#cmsdita2010 G.Joseph: customer-specific docs become a different slicing & dicing of features.

#cmsdita2010 G.Joseph: treating inclusion/exclusion of features and bugs like a bill of materials for doc production.

#cmsdita2010 G.Joseph: user community was confused by assumption to include all unless specifically excluded. Went to reverse model.

#cmsdita2010 G.Joseph: profile each sentence, the smallest segment generally used by translation memories.

#cmsdita2010 G.Joseph: profiling using DITA kwd tag + feature-id attribute; can then use suppress-feature-id to drop content. Many examples

#cmsdita2010 Gershon Joseph "Advanced Feature-Based Publishing": cms can help work around including/excluding content

Catherine Lyman and Martha Morgan, NetApp

#cmsdita2010 M.Morgan: our HW writers made the shift best. They were all organized together and could work together.

#cmsdita2010 M.Morgan: identify where you would have content redundancies if you didn't have a single source.

#cmsdita2010 M.Morgan: people are hesitant to go to others and ask for changes to "someone else's" content.

#cmsdita2010 C.Lyman: DITA as Lingua Franca--you write in DITA and we'll take care of it; we'll write in DITA and you take care of it.

#cmsdita2010 M.Morgan/C.Lyman: starting now you have an advantage -- mature tools, people to lean on. Call any of them, they'll talk to you

#cmsdita2010 C.Lyman: benchmark now. Reward team members who help expose those #s. Look at pressure points. Typical assignment reqs, #words

#cmsdita2010 C.Lyman: don't over-estimate time to DITA productivity - to writers, to business leaders. Ultimately pubs is a cost center.

#cmsdita2010 M.Morgan: big corp libraries get harder & expensive to maintain. Don't overvalue consistency & simplicity over ROI.

#cmsdita2010 C.Lyman: are writers of questionable skill or subversive inclination talking? Fix that 1st

#cmsdita2010 C.Lyman: go to business and engineering groups early (1-2 yrs) and be clear on corp drivers. Put customer-facing improv 1st!

#cmsdita2010 C.Lyman: sell to whole org. Answer the question for them: "I see how this can benefit us generally, but what's in it for me?"

#cmsdita2010 C.Lyman: need to sell value prop to other BUs, whole organization, not just your management.

#cmsdita2010 M.Morgan: the biggest challenge isn't DITA tagging (conspicuous!=biggest). Invest in classes in Structured Writing (any sort)

#cmsdita2010 M.Morgan: good DITA is more than valid DITA. It's smart, standards-compliant DITA tagging

#cmsdita2010 C.Lyman: 16k topics 1.4k publications at NetApp. Huge content reuse, flexibility, consistency. DITA is totally worth it

#cmsdita2010 C.Lyman: what we did wrong "if only we had known: the snares and pitfalls of managing in a DITA environment"

Nola Hague and Tom Bondur, Actuate

#cmsdita2010 T.Bondur: working in an agile/scrum situation takes all the professional and inter-personal skills you can muster

#cmsdita2010 T.Bondur: many writers feel insecure abt having their work exposed, dissected, & rewritten by a team in an open setting

#cmsdita2010 T.Bondur: scrum sessions informed by his exp. teaching freshman composition.

#cmsdita2010 T.Bondur: Docs req constant restructuring/rewriting to keep content relevant/focused /flowing. It's hard! Need to spt writers

#cmsdita2010 T.Bondur: you really need to understand it &scrub your Docs. Consultants can't solve those issues. You have to do it eventually

#cmsdita2010 T.Bondur: we collaborate all the time so nothing is a surprise.

#cmsdita2010 T.Bondur: demeanor and tact v.important here. Need to do this in a sensitive way and pay attn to group

#cmsdita2010 T.Bondur: editing scrum -- conference room + remote folks. Mgr moderating and revise the content right there, driving/guiding

#cmsdita2010 T.Bondur: our content is changing. No more books--we are delivering content to match how it's being consumed.

#cmsdita2010 T.Bondur: needed to move to a robust,enterprise dynamic publishing system with a DITA-aware editor built to do it from the ground up

#cmsdita2010 T.Bondur: we overused Frame's tools to the point it wouldn't compile correctly & we'd get spurious content.

#cmsdita2010 T.Bondur: traditional release cycles are a thing of the past..publishing is rapid iterations now

#cmsdita2010 T.Bondur: we are doing this with teams around the world (his copresenter is presenting from Brazil!)

#cmsdita2010 T.Bondur: we have to keep our Docs ready to go--shipment ready every day. Automated/public process to ensure it.

#cmsdita2010 T.Bondur: our scrum editing sessions are particularly good to work with legacy content.

#cmsdita2010 T.Bondur: vigorous editing sessions -- it's a scrum session including anyone required to get it done (Eng, Pubs, anyone)

#cmsdita2010 T.Bondur: the definition of a writing group is currently shifting dramatically and we need to be aware of that to avoid a mess

Presented by: Charlotte Robidoux, HP, and Suzanne Mescan, Vasont

#cmsdita2010 C.Robidoux: collaborative review saves time for reviewers too.

#cmsdita2010 C.Robidoux: Reuse + successful CMS is not linear

#cmsdita2010 C.Robidoux: Define reuse. Reuse is a cultural issue: build trust, structure, measure,& track

#cmsdita2010 C.Robidoux: cultural shift to watch for: Can't get to reused content if you don't put your content in the system (cms)

#cmsdita2010 C.Robidoux: Understand your own processes, your own work habits, so you know what you want you CMS to support.

#cmsdita2010 C.Robidoux: mentioned she was just discussing review phase tools with our own Liz Fraley

#cmsdita2010 C.Robidoux: collaboration in the review phase, where you need it most: automated synthesis of review feedback is key

#cmsdita2010 C.Robidoux: Use the right collaborative technology for the right purpose. Sometimes the phone is better than chat

#cmsdita2010 Conference wrapping up; tutorials this afternoon. Like summer camp, see you all next year!