Download it from CodePlex.

Some of my gripes include:

• The theme system being tied to the HTML split between Master Pages and WebParts. It would nice to be able to easily customise the HTML that is generated.
• SharePoint Designer marks files as dirty when they are opened. What is being modified? Occurs most frequently when I’m working with Master Pages.
• Superfluous XSLT generation when a ListViewWebPart is turned into an XSL Data View. Why include templates that are not used? Why not offer to store the XSLT in separate files for easier management. There are cases where the XSLT code generated in invalid too (bad code in conditional tests).
• Moving XSL code to separate files introduces another problem with dirty files. Saving Master Pages that reference external XSLT files can result in a dialog that asks whether to overwrite. I’ve found that I should always accept the offer to overwrite

I’m pretty sure these issues are easy to fix, but Microsoft and other SharePoint developers recommend that development is done with the ISV-oriented tools. This is very inconvenient since the product is sold as a rapid development platform and packaging/deployment requires quite a bit of extra effort to get right. Hopefully the SharePoint experience will improve for front end developers in the next release.

Going into the conference I was looking forward to the Sunday session with Jim Hugunin but there turned out to be some more treats for the IronPython developer. Feihong Hsu ran a session on Python.NET and how you can bridge from CPython to the .NET platform, taking advantage of rich Windows APIs. Michael Foord spoke on Silverlight as well as his company’s spreadsheet which embeds IronPython.

Feihong organised an open space session for Saturday evening after the PyWin32 gathering to talk about Python.NET and we were joined by the IronPython developers and management (Dino Viehland, Harry Pierson, Jim Hugunin and others). We discussed a number of aspects of IronPython and progress towards the 2.0 release. It looks like this may be complete in October given that they released the first beta last week. Again Michael Foord had something interesting to say on what Resolver Systems are doing.

Michael presented an open source project called IronClad. This is quite an insane assortment of code from C# to Python to assembler all in the name of accessing Python modules written in C. To date they have the bzip2 module running but are working on support for modules like NumPy which are important to their customers.

After the open space session we headed into Chicago for dinner at India House. This gave us a chance to find out some more stuff about the IronPython implementation, and other factoids. Dino hinted that he was working on getting Django up and running. Little did we know he was going to be demoing this to the crowd on Sunday.

Sunday saw Jim’s big talk and I managed to get a few photos. It wasn’t easy, but I think these turned out a bit better than earlier shots at the conference. Dino showed off the fairly minimal changes needed to get Django running on IronPython and Jim demoed the IronPython interpreter running under Dynamic Silverlight.

After the keynote, Dino gave a me a quick run through of the IronPython and DLR source code. This was very interesting and it gave me a real step up in understanding what goes on under the covers. Thanks Dino!

The first keynote from White Oak Technologies was the story around their use of Python. As a consulting organisation they are presented with challenge when deploying Python solutions for clients. They outlined various ways that clients would push back on the use of Python and how they turned these into opportunities. Nothing really special here but it’s interesting to hear that people are having success when challenging other developers and management on the use of Python.

Up next was Guido van Rossum, the creator of Python. His primary goal was to give us an update on the progress of Py3k and Python 2.6. Apparently both are scheduled for release in August. I would assume this is only tentative as they are only at the alpha stage at present. I knew the Python language had been in existence for some time, but I didn’t realise it actually started as an academic project in 1989.

It’s age appears to be part of the problem for Python in 2008. Design decisions around character encoding (no unicode by default) and locking in the interpreter are a contentious issue for a lot of people. The latter issue is not going to be fixed according to Guido but unicode is going to be the standard since there is less of a performance hit today.

Guido is still recommending the 2.x branch of Python for the next few years. Obviously there are performance and reliability concerns for early migration to 3.0. This is good to know since I know that IronPython will be able to key up with the pace of the CPython schedule.

When I made my tutorial selections I was hoping to attend a Python for Java Developers session. This would have been useful given my experience with C# but it seems that I was in the minority and it was cancelled. I switched to the Django session but I think I may have been better attending the session on performance optimisation.

Registration wasn’t too busy today since the main conference crew won’t arrive till Friday. I got a PyCon bag and some flyers but the T-Shirts weren’t ready. Apparently they’ll be available tomorrow, I’d hate to miss out on one!

Python 101 Tutorial (Steve Holden)

It turns out that Steve was another British ex-pat living here in the US. It gets weirder in that he lectured at Manchester University for a number of years. Given Steve’s position in the community I expected a sharp introduction to Python. It didn’t disappoint and I picked up a fair bit. The ’slice’ mechanism looks really useful, I wonder if it can be implemented with any of the new C# features?

Getting Started with Django (Jacob Kaplan-Moss)

I was expecting this session to be a little more exciting. Jacob has some fine ideas about how Python frameworks should be built but his presentation style is not as striking as DHH. Since this was an introductory session I can’t complain too much but I really want to hear some more about Django deployment and debugging over the next few days.

Internet Programming with Python (Wesley Chun)

At this point I was pretty exhausted. This tutorial seemed to be geared toward newbies to network programming, rather than a best practice session on leveraging Python for internet programming.

The second half of the year is open for conferences and the like so maybe OSCON would be fun. Microsoft are running their PDC, but I’m not sure I’d get as much from that as I have done previously. I’ll have to have a think about it…

This is a good job since I’m doing more work with these open source technologies at the moment since Microsoft have done little to improve the ASP.NET platform. For .NET developers, MonoRail is the only usable framework for development. Additions such as ASP.NET AJAX do little to make the platform better for the applications I am working on. The problem with AJAX is that it’s not an end in itself. You need to have a reason to use it and a design to match, or you end up on the road to poor performance and usability.

The crowd at RailsConf was very different from that at Microsoft events which I have regularly attended in recent years. This was refreshing but also a challenge as I’m not a UNIX guru by any means. It’s clear that there is little interest in the community to actively support Windows and provide a bridge for .NET developers to migrate. This schism is something that the veterans like Dave Thomas are looking to cross as there is a lot of time and effort invested in other technologies. DHH’s keynote address railed against support of XML Web Services and legacy database schemas. It will be interesting to see if ThoughtWorks, or another vendor, will step up and provide more support for working with enterprise technologies.

Deployment seemed to be a hot topic for all users at the conference and it was interesting to see the capabilities with Rails. Mike Clark’s talk on Capistrano was wonderfully presented. The simplicity of being able to pull together a build system which integrates source control and remote deployment was amazing. This is something that is sorely missed in the .NET community and could probably be built on top of NAnt in a less elegant way than Capistrano.

Jan Kneschke provided an interesting insight into building a Web server to handle a massive number of connections. His Web server, LightTPD, was popular with the Rails community due to its FastCGI implementation but now has competition from Mongrel. He’s planning on improving the mod_proxy support so that “Lighty” will work with Mongrel in a similar way to Apache.

On the final day of the conference, Stuart Holloway of Relevance, and former Java junkie at DevelopMentor, provides an insight into some of the Ruby techniques used by the Rails team. Stuart has spent a lot of time going through the Rails internals to figure out what the Rails core team has done to provide such cool functionality. Since I haven’t spent too much time with Ruby a lot of this was very deep, but the explanation of analogues in the Java world helped a lot. This talk provided me with a lot of food for thought and is the reason I signed up for RailsConf.

If you are interested in other sessions I attended, I’ve posted my incomplete notes for download.

Regards the venue, facilities, and catering I think that the RailsConf organisers did a good job. Providing wireless internet access for 550 people is not an easy task, and the fact I could check mail from time to time was good enough for me.

Overall RailsConf was an excellent event, and I’m lucky to have attended the first conference. RailsConf II is likely to be a lot bigger next May, and the overall feeling will be different. It will be really interesting if Mike, Oliver and Dei come along and we can catch up on a year of Rails development.

Day one of the real conference kicked off today with a keynote from Dave Thomas on what he feels are areas where Rails can improve. He didn’t spend a long time covering it, but support for other data sources behind a model is an important addition. Having improved database key support would also be nice to have for those working with legacy systems.

The best sessions of the day were on Capistrano, a Rails deployment engine, and Asterisk. The integration of VoIP with Web applications is surprisingly simple with this open source platform. I felt the Open ID session was a miss because the information had been presented in a similar way to Dick Hardt’s identity presentation, and there was little on the Rails specifics.

A very useful notion from the world of Rails are migrations. These help you update your database based on changes to your application, and are very much automated by the platform. Unfortunately these are not fully transactional since MySql and friends do not have support for transactions around DDL. I wonder when they are going to catch up with SQL Server in this regard?