videos from Lang.Net symposium

Thanks to Paul Vick for pointing out the Lang.Net symposium talks are now on the web.  I’ve only watched a few so far, but here’s my thoughts/review of them


Paul’s talk on VB.Next

A quick overview of one area VB will probably be heading is making the language more scriptable.  This isn’t anything revolutionary as we all recall VBScript, and we all also know that died for a number of reasons mainly lack of support in other browsers and lack of a public standard that wasn’t proprietary Microsoft control. Today as far as scripting goes, JScript rules the nest.  But for client side applications there is still a void from what VBA and the Scripting control use to provide. Admittedly it’s not that hard to roll your own with reflection, but not as easy as it use to be. The kind of scripting we are talking here is more like running code in the immediate window of VS which you can do today, and probably extends through to possibilities like recording macros, something Office still lacks for VB.NET (although Visual Studio does have VB.NET macro recording today)

How important this will be to the VB community I think is questionable.  Indirectly it will be important as there should be some good spin-offs, especially inside Microsoft.  One advantage that shouldn’t be under estimated is that the VB team will be writing the compiler and editor helper bits in VB.NET.  This kind of level of dog fooding will for the first time probably drive them to really consider power features inside the language a lot more ๐Ÿ˜‰

Concerns will be how much this takes time away form other things the language and users really need. And the scariest thought is this will be yet another attempt at dumbing down VB, with the typical patronising “Mort” persona rubbish, where the VB team no longer asks real users what they want, instead they ask silly questions like “assume you’re Mort ….”.  I really hope we|they have matured beyond that by now.


Modelling and Languages

Jay and Silent Bob Don Box and Chris Anderson give what was probably an entertaining presentation. I love Don Box’s presentation style, always fun/entertaining.  But if you are looking for content, to actually watch an informative presentation then skip this one.  They really don’t talk about what Microsoft is doing, more they talk about what has already been done, e.g XAML.  You should get the message that Microsoft thinks modelling is important and folks there are looking at making that happen, and hopefully doing so in a way that congregates other tracts they have worked on.

Sadly this presentation doesn’t really present anything concrete, only a foofy world of maybe.  It’s disappointing to me as I view modelling as incredibly important and the interaction with languages is crucial. Don and Chris have been working on it for some time so it was sad to see their presentation was nothing more than a rehash of one they did twelve months ago.


remotion Mixins

A good bit of European humour opens this session. Not really in depth, but does provide a good overview of how remotion is using Mixins today.  Their implementation is a runtime object factory that looks for attributes and does the mixin magic accordingly.  I think I’d prefer a more AOP compile time injection, and perhaps more design time feedback on possible issues.  As the guys “hint” at, it would be great to see language support for mixins.

In the VB world, Amanda Silver, reportedly mentioned “Extends” at TechEd EMEA 2007, but to date the VB team hasn’t made those slides generally available or opened up any discussion as to what mixins would mean.  Hopefully that will change soon, if we are to take them seriously over the claim to “democratise” VB ๐Ÿ˜‰



Erik Meijer provides an interesting and honest presentation of Volta. So far this has been the best session I’ve watched.  It had a good level of technical content combined with the view of the overall goals.

Of course, I don’t agree with some of the way Volta is targeted/implemented at present, but I’ll apply the Moore’s law factor for now ๐Ÿ˜‰ I would like to see Erik and the Volta team starting to look more at the modelling side of things rather than just attributes in code, maybe even start adding some substance to Don & Chris’s work <g>.

Lots of good take away points, especially the parts about making things easy to adopt yet giving the developer the ability to make explicit choices.  Someone should make that one a poster and put it up around certain team’s offices… “if you’re going to do magic, give them the choice”