I'm finding myself in a peculiar swirl of convergence where everything seems to be about small chunks of thought, not necessarily complete or consistent. These chunks need tracking, comparing, versions controlling and melding into presentations of varying narrative richness.

I'm sure this doesn't represent any major shift in the industry, more a case of the when you're buying a VW, every second car you see is a VW syndrome.

Traceability and Agile Requirements

I am a member of a Software Engineering liason group between industry and academia. Over the last three months we've been discussing Agile Requirements and the perennial problem of Traceability through decision-making, requirements definition and the refinement of understanding over time.

Something I was disappointed to hear was that there isn't any killer app out there for either traceability or testing of requirements. The group did agree that whilst requirements testing is an open field at present, in the end it all comes down to keeping your stakeholders happy and convinced that the project is worth it.

One interesting job category discussed, new to me, was the Requirements Engineer about which there was an interesting debate as to how universities can produce them. The combined experience in the room resulted in a consensus that all universities can do is provide some background - you can't shortcut the industry experience needed to be able to develop requirements.

Roadmap Development

In the Roadmap Development thread, I discussed some ideas for documenting a roadmap and consequentially system architecture as a series of small assets. In particular, I explored the idea that you may want to combine these assets in many different sets and sequences, depending on your current viewpoint into the project.

Low-effort Documentation

We recently had a thread on low-effort documentation which explored some of the issues about documentating a system in an agile and collaborative way, involving of-course lots of comments on wikis. http://www.wikimatrix.org/ makes it relatively easy to compare the merits of various wiki engines, for anyone interested in exploring further. 

I have had very positive experiences with wikis for collaboration, where threads of discussion are collected, and for casually editing documentation. However, they don't allow the threading and management of chunks that really covers the needs I'm talking about at present.

GroupEdit

As originally discussed in Wanted: Better Document Collaboration System and with a call for Developers and RFP in public review, this is an interesting view into the workflow of a publisher as being  about ways to track, discuss and authorize or revert changes.  

MS Word's Track Changes, weblog editors and other collaboration tools don't quite cut-it so they have decided something needs developing. They are currently using a combination of Subversion and a text editor with some disciplined inline markup. The links above include an interesting list of suggestions of various collaborative software tools and version control philosophies.

In particular, their needs are differentiated by Editing for Content versus Editing for Style.

ScribbleDEE

ScribbleDEE is another project of mine that has been languishing in the design stage for about two years, admittedly at a low priority, whilst I tried to work out the issues in how you compare variations of UML diagrams and especially allow for tracing through fragments of design. 

It turns out to have very similar complicating issues to the GroupEdit project with regards to the tracking of changes and the desire to authorise some variations or changes. However, the presentation side, as being fragments of UML, is of course very different from editing in the context of a largely textual document.