Long Live the RGA
Although there are a rather a lot of things to do right now, what with trying to get VS Whidbey out the door, and settling into my new life in the pacific northwest, I did get over to the VSTO 2005 DevLab as much as I could this week. Coming from MCS into the product groups in Redmond, I do have some idea of how people actually use the stuff we put out, but we should all take the time to top-up on customer experience as often as we can.
We have a notion of what would be a good product or a feature-set within a product, and there is clearly some rigor in the way we validate our design choices against real and projected customer requirements. There seems to be a policy here of keeping us ever-so-slightly under-resourced. The idea being, that this will force us to focus and to prioritize more precisely when we triage bugs, when we work out which of the range of new features must get released, and which can be deferred to a later release, and when we design the UI to our features (especially the project systems within Visual Studio). Unfortunately, one of the side-effects of this policy is that its so easy to cut the customer perspective from our busy schedules. It's an obvious target, partly because it's so difficult to get it right anyway.
Events like the Office DevCon a couple of weeks ago, and the DevLab last week are therefore doubly important. They give us a chance to catch up on all the customer perspective considerations that sometimes take second place. A golden opportunity to bring the correct balance back to what we do.
This is also a large part of why we put out beta versions of our software - unless we get lots of solid feedback from people who use our stuff, how can we sign off on it? Public beta programs are especially useful in unearthing the Random Geek in Ashtabula who is doing something weird with our stuff that we just couldn't possibly anticipate. We obviously do a lot of directed testing, but we can only do so much scenario-based testing. For instance, the Trinity team knows how to test the functionality of VSTO (you would hope so, as that's our product), we also know something about Visual Studio, and about Word and Excel. However, we can't predict all the wacky solutions bubbling up in the head of the RGA. So we absolutely need him or her to bash on our product - and, crucially, to tell us where it doesn't fit (or where it breaks).
Now, of course, some of the things the RGA wants to do just aren't right - VSTO is essentially a tool, and therefore reasonably open-ended, but you can only bend it so far. If it were truly a universal panacea, we'd never need any more tools, and that's not realistic. Nevertheless, in trying to bend the tool to his/her devious requirements, the RGA may well uncover some unexpected weakness in the overall design or some detail in the implementation of the tool that we need to know about. Long live the RGA.
There's also a selfish personal angle to this. These events give me the opportunity to renew my friendship with familiar faces such as Charles "Ice-Cream" Maxson, Stefan Zimmerman and Julie Kremer. That said, Julie (deliberately? who knows) continued to fox me with references to foodstuffs that only a native of the US would understand. After Marcie Robillard (The Datagrid Girl) bamboozled me at TechEd last year with talk of 'little debbies', Julie is now talking about Push-Ups <sigh>.