20 Sep 2007
Gartner Portals, Content, and Collaboration Conference: Day 2 Notes
Portals, SOA, and Mashups
Portals have worked well in some enterprises, but there are problems:
- Heavyweight technology requires specialized skills
- Over-engineered platforms and protocols are complex and expensive
Portals will likely be an entry point for enterprise mash-ups
- This site is an interesting index of mashups that have been done, categorized by type of mashup: http://www.programmableweb.com
- This site has a ton of downloadable widgets for incorporating into web site pages: http://www.widgetbox.com
Interesting tidbit: Systems reflect the structure of the organizations that build them. If you have a dysfunctional organization, it is likely your systems will be too. Building the “perfect” system doesn’t help. Over time, the system will degrade to match the structure of the organization. (Melvin Conway, 1968).
I’ve definitely seen evidence of this at clients. A really obvious example is in workflow-centric solutions. Companies rarely take the time to streamline and cleanup processes the way they need to before formalizing the processes in automated workflows (“Implementing the new system is enough change. Let’s automate the way we do it now to encourage user adoption. We’ll come back later and tweak the business processes.”). So what you end up with is a system that works exactly like the current business process, warts and all.
More info on Conway’s Law at http://en.wikipedia.org/wiki/Conway’s_Law
Portal of the Future
More discussion about the differences between Digital Natives (the generation of workers entering the workplace today) and Digital Immigrants (the generation of workers who’ve had to adapt to digital technology). Where this relates to the “Portal of the Future” is primarily about expectations regarding how we interact with relevant data. The nature of this interaction has to be client-independent, device-independent, context-aware, and truly-mobile.
- Portal vendors will decompose products into services
- Portal as entry point for enterprise mash-ups
- Portals sharing data (due to most enterprises having multiple portals deployed)
Sites to check out for good examples of alternatives models of content aggregation:
These are basically the next generation of the old “My” portals.
Portal interoperability touchpoints:
- Portlets (Standards exist here, but not for the rest)
- User Profiles
- http://www.popfly.com/Overview/ (Microsoft)
Seven Key Strategies for Combining Portal & CM
Probably a better title for this presentation would have been, “Seven high-level ideas that are related to portals, content, and collaboration”. It didn’t really discuss the combination of the two at any level of detail.
1. Build a foundation for a high-performance workplace
- Provide a flexible environment to support the non-routine tasks people do every day. Key technologies: Collaboration and Content Management (Basic Content Services).
- Focus on non-routine tasks and decision-making enablement
- Content and collaboration are “birthright” technologies for your employees
2. Use the right tool for the right task
- The problem is that email is being used to create/share content
- Use BCS as a way to migrate people away from email
3. Take a process view
- If you want to really get a handle on collaboration and content you need to start by looking at the processes in which collaboration plays a part or which involve content.
- Gartner talked about two types of content/collaboration apps using two acronyms: CEVA & WEBA. Hopefully, these new acronyms won’t out-live the conference, but just in case they do…
- CEVA stands for Content-Enabled Verticalized Applications. This is an application that is focused on a core business process. For these apps you should be able to provide templates, workflows, and automate some or all of the business process. (A lot of the workflow apps I’ve built over the years have been CEVA’s. Who knew?)
- WEBA stands for Workplace-Enhanced Business Applications. These are apps that our really outside of your core business process. For these you need to define policies, communicate best practices, provide support, and examples.
4. Intranet & repository consolidation
- The path of least resistance is to freeze your legacy content and create all new content in the new repository. The legacy repositories will dry up and go away over time.
- The other alternative is migration, which can be costly, but at least gives you the chance to clean up.
- The route you choose is really about cost.
5. B2C web presence
- Most enterprises will need both Portal and Content Management
- Each has specific areas of functionality they are responsible for, but there is some overlap in the gray area between the two.
- For some reason, the speaker went deep into eforms at this point which I found strange as the bullet is about B2C web presence. I don’t downplay eforms as important, particularly as a low-hanging fruit for most enterprises, but it seemed like a non-sequitur.
6. Customer-centric communication
- Nothing significant to note for this one.
7. SOA framework for the assembly of services
- Again, kind of fizzling out at this point but the speaker did point out that your metadata approach can impact your ability to provide meaningful integration between or federation of your repositories. Enterprise-wide efforts to define and update an enterprise-wide data model help but can be costly.
Usability & Productivity: Are We There Yet? (Jakob Nielsen)
This was my first time hearing Nielsen speak. I found him to be insightful, engaging, and entertaining. Catch him when you can.
Usability is not just about visual design–it must also include information architecture and interaction design. You cannot just come in after the coding is done and slap on a pretty interface.
Everyone can do (and should be doing) usability testing
- Doesn’t require a lot of fancy equipment
- Closed door
- Web cam
- Screen capture
- Plan on roughly five users for each type of user in the system. If overlap exists across user types, you can get by with less than five.
- Keep it simple
- Make it iterative. Don’t make it a big production. Maybe test one user type on one iteration and another on the next.
Of award-winning intranets, the following are the usability methods employed by those teams:
- User testing (#1 method)
- Card sorting
- User test of prototype or wireframe
- Heuristic evaluation
- Field studies
- Accessibility testing
The Web 2.0 sites/technologies getting covered the most by the media are not necessarily what you should be doing in your projects
- The ugly truth about user-contributed content: 90% of the users contribute very little, 9% contribute a moderate amount, and only 1% are the hardcore, heavy users
- Get back to basics
- Task-oriented navigation
- Learn how to write for the web
- People scan when they read online. Writers have to adapt to that style.
- Use HTML pages instead of PDF. PDF versions of online manuals on a corporate intranet could be costing you “hundreds of thousands of dollars” in productivity loss
- Tiny text with low contrast is really hard for people over 45 years old to read
“Banner blindness” is a strong phenomenon
- Showed some eye tracking test graphics in which users were not giving a single glance to any of the ads
- Non-ad content that simply *looks* like an ad gets skipped
- Environments where a user should know there are no ads (like an intranet) are not immune to this effect