It is late and I am exhausted. And I have no pictures from today (/me looks at Richard Miller).
We started off on a good foot, having already taken a good stab at objectifying the tasks. We further hashed out all more objects, added more attributes, added verbs and relationships. Some objects were much easier than others; for instance, Product really only consists of a product name and version number in the scope of an SR. Yes, it is a "child" member of other objects.
In retrospect, we got bogged down in many areas, and sometimes it seemed like those areas were really minor and we were spinning our wheels. But the "spinning wheel" did demonstrate that even in our small group there is ambiguity and misunderstanding of core elements. For example, is a Primary Contact a subset of a "Contact Type"? Or is it different enough from other contacts (ie, Secondary, Manager) that it deserves it own types. One member argued quite vociferously that it is its own object because it is hanlded differently, like populated from a User Profile, while the others are not. After ripping up and redoing Contacts in various permutations, we finally decided on a single Contact object with various conditional properties and verbs.
There were other examples of the same thing, I just do not remember them off the top of my head. These little excursions took up a bit of time. On top of that, we also delved heavily into the Knowledge Base and Search, since we had decided to expand our scope the previous day. While much of our journey through this topic is quite useful in the context of filing and resolving an SR, it consumed time as well. So even though we had covered a bit of ground, Richard Miller declared that we were several hours behind in the late afternoon. :) I am not sure what that means for tomorrow.
Some very newsworthy things that came out of our session. I have not signed any non-disclosure agreements, but I do think the Managers want to keep a lot of new developments under wraps. So I'll go about it indirectly. We chatted up some more "Blue Sky" features as we discussed things we did not like about the current implementation. One of the key features to our new approach is using Search heavily at the outset of a possible SR creation process. I know, you are thinking this is a HORRID idea. But if Search were actually much improved (in terms of performance and relevance), we see this as being a huge boon. We could be wrong, time will tell. Basically, you have a form that provides an opportunity for the user to provide a ton of information. Much of the form is optional, but the idea is that the more you provide, the better the search results. Using ideas like Google's word-completion and instant results, and eBay's and Amazon's left-hand pane of refining and drilling-down, we explain how these kinds of features would significantly enhance the user's perception of Search by providing fast, dynamic feedback on the criteria entered. On top of that, the user may have a chance to save the search filters/results and shove all the pertinent information entered into an SR, or maybe even a Community forum post. Some of the above ideas have already been developed and we saw some simple demos. Like using quickfill and/or word completion in various areas. Very nice to see that they are already make in-roads in that direction.
I am particularly torn about the latest prototype GUI mock-up that our group has achieve so far. I claim my role, so I am not blaming everyone else. I say I am torn because the pages/screens that we "developed" today still look very busy and crammed full of things to fill out. It almost looks like we have merely re-arranged the existing SR fields that one normally fills out. I think the key importance in our approach today is that we are aiming for two things:
- allow as many optional fields as possible
- thus giving the user a choice between providing less detail and possibly a more vague search, or more detail and possibly a more accurate search
We are both assuming that search will be improved significantly, and providing ideas on how exactly to do that. The dynamic feedback mechanism is crucial I think, since it gives the user a good idea as to how many docusments are returned and how to refine it. I think. It looks good on paper right now. :)
Ok, that's it for me. I hope to procure more pictures tomorrow.