Sunday was the first of two pre-conference days before the conference starts for real on Tuesday. Sunday proved to be very worthwhile. Firstly it has been a chance to catch up with many faces from the past - colleagues and customers (hi again all!). It also was an opportunity for a lie in - with sessions not starting until 10am - yippeeeeeeee!

I think Microsoft has done a great job of the "The Big Room" - it contains commnet (works a treat - just remember to sign out of passport and to not associate your login details with any website as machines stay logged in!), a great shop full of excellent books, t-shirts that fit and gifts for the kids, an area full of Windows Vista boxes (using one right now) and a great chill out area with thick carpet and comfy sofas - superb :) The hand on labs are also situated here - but they start on the Tuesday.

Now - down to the session I attended. I went for the VSTS session as I have done nothing with it so far - but have in the past dabbled (and more!) with team development software, agile methodologies etc. I learnt a lot during the day - but for my taste it was all a bit slow and a bit dry. The presenters clearly knew their stuff - but something was missing for me to make it into a compelling day. A few bits and bobs which caught my eye:

  • This looks IMHO a verson 1 product in many, many ways. It certaily looks valuable - but I can see many "compromises" in what is being delivered. The user interface is far from "ground breaking", presentation of info is often simple, integration with other tools is basic, it seemed horribly slow at times....I could go on. BUT the goal they are shooting for is spot on and what they already deliver will offer significant gains to ISVs I work with.
  • There is integration with Microsft Project but it appears to very simple - which is a real shame
  • When asked about Visio and UML the presenters agreed that there was absolutely a huge place for UML and Visio. VSTS was delivering at a  "lower level" allowing architect and developers to converse but not (for instance) analyst to architect or analyst to customer. It is therefore a shame that there appears to be no integration between VSTS and Visio UML.
  • Shelving was a very welcome "find" for me. I had not realised we had an implementation of Shelving in VSTS. Shelving is a great feature that I have had to implement in the past on Unix systems. It is a way to move specific "work in progress" out of the way to get something else done (e.g. an urgent security fix) or to another user (e.g. when going on leave or when it just became too darned hard!). Check out http://blogs.msdn.com/crathjen/archive/2005/04/06/405909.aspx for a brief discussion on Shelving.
  • Unit Testing support in VSTS looks rather nifty to me. Folks who enjoy Nunit should be very happy with what the VSTS folks have delivered. A couple of items stuck out for me. VSTS supports "test last" allowing test code to be auto generated from the functional code. Ok - we all know this goes against "test first" development but lets face it, many companies have many functions that are crying out for a spot of testing. This one feature will resonate well with them. It also allows testing of private methods without being invasive in any way to the code. I think this is a great feature for applications that have deep object chains where it can prove difficult to test all private methods by defining suitable tests to the higher level public methods.  Take a quick read of this Q&A on unit testing.

There was lots of other cool stuff (I mean useful stuff - sorry, been in the US nearly 3 days) as VSTS is a huge product. A conclusion I reached was that Microsoft needs to work on making sure VSTS does not look "too big" for smaller development shops to pick it up. I need to spend a little time thinking on this.