Compsoft Flexible Specialists

Compsoft plc

Compsoft Weblog Compsoft Website News Archive Privacy Policy Contact Us  

Monday, November 10, 2008

Richard Thomason: Day 1

Keynote session

Jason Zander made a brave pre-alpha presentation using Windows 7 and Visual Studio 2010, demonstrating improved testing, debugging and delivery in .net, some excellent tools for analysing and importing VS 6 applications, which made me think hard about doing something along those lines with Equinox in VS2008, something I have been considering for some time but have thus far rejected. In particular, Microsoft are making efforts in future releases to improve and simplify support for multi-threaded architectures, as this will be the primary way to deliver acceptable performance in applications with big graphics requirements.

1 Introduction to F# - Luke Hogan

This sesion was something of a longshot as my previous experience of functional programming is that it has a lot of brackets and is only really useful for academics and writing programs to shut down nuclear reactors. However the overview emphasised links with the CLR and implementations of parallel and asynchronous techniques, so I thought I'd give it a shot. The session was full.

Luke introduced the programming environment, which included a Basic like interactive pane, and showed how to write more declarative code. This boils down to syntactic changes which achieve similar results in other languages. The languages has elements of Basic and especially the functionality hiding features in Javascript, but also has interfaces to UI controls and the CLR generally. The parallel functionality was a direct call to the .net functions.

The session was very interesting. It would be interesting to compare the facilities in Javascript with those in F#. I'm wondering if the sandbox nature of Javascript would probably restrict the things you can do, although the language itself is so powerful. On the basis of this session, I might have a little go with F#.

2 Evangelising Sharepoint - Gianpaolo Vittorelli

Not much choice this time. Neil is doing Virtual Earth. There's one on customising Sharepoint with an awful presenter who I remember from last year. The others are doing Ajax and data retrieval. The best available is a high level requirements seminar on how Sharepoint was used in a number of projects.

This was an excellent non-technical session led by a developer/manager with considerable customer facing experience in SP.

Sharepoint developers are a highly sought after, scarce resource, however Sharepoint development has an arguably bad rep. Why is this? SP can be used in a variety of ways - from a configurable internet all the way to a framework for a bespoke .Net application. The main issues are the need to learn how SP works, plus it requires a slightly different mentality - the need to co-operate with the SP framework, rather than program using .Net as a development environment. There is a loss of control involved. Basic requirements for staff are to be a good .Net developer - you can't use inexperienced people unless they are very very bright. Workflow in SP is not great except for simple applications; there will be improvements in 2010.

In many ways, there are parallels with Equinox - it's simple programmable and extensible, but SP is a less controlled environment, where simple configuration changes and uncontrolled end user development can make a big mess. Therefore it's necessary to severely limit these options.

"Long tail" applications are ideal for SP, where you start simple and the users continue to add functionality forever. User self-development is the key to its success - something like Equinox queries and reports - however SP adds forms also.

Possible SP sponsors:
  • HR etc want an intranet, and have made a start but want help.
  • IT want to get rid of file shares, plus better document management and don't have time/skills to implement themselves.
  • Some other business unit has a process which needs automation.


SP in his opinion is not a cheaper/more profitable platform. However, if SP is already there, or if the application is a small natural SP app, it would be cheaper/more profitable. The benefits of SP are probably more important than the profitability.

0 Comments:

Post a Comment

<< Home