Drawing to a new conclusion...

As those of you who read my various other web sites (and the earlier incarnation of this blog) may be aware, I started a new job at the end of April this year which resulted in me moving from a Java (and occasionally SQL) developer, on to C# (via Microsoft Visual Studio 2008) and back to (what has become over the years) the nightmare of Delphi 5.

Delphi, specifically version 5 was way ahead of the crowd in it's time, unfortunately that time was 1999. I used to love Delphi, it was so advanced compared to similar offerings such as Visual Basic that anyone who had used it seriously didn't want to go back to other languages unless they were offered a seriously large pay-cheque to do so.

Then a series of events happened which saw Delphi get left behind, the two main (from my perspective) being:

1. Java arrived and offered the opportunity to program in a new object oriented language that allowed code to run across different platforms. Many Delphi developers moved over to Java, and by using JBuilder which also originated at Borland, we felt at home in a familiar IDE too (versions 1 to 3 of JBuilder were written in Delphi, JBuilder 3.5 was the first native Java IDE for JBuilder).
2. Anders Hejlsberg was poached from Borland by Microsoft and as a result the lead architect had gone. Don't get me wrong, there were a lot of very talented people at Borland, but despite this there was always the perception (accurate or otherwise) that after Anders jumpred ship, the product was left in disarray.

Delphi was never as popular again.

This is all a long time before the .Net platform, and more importantly for the remaining Delphi developers, C# started shipping. In fact, Anders Hejlsberg was the chief architecht of C# after his J# language, understandably, upset Sun, so morphed into Cool, which was eventually released as C# (at least that's how I remember it).

Despite remaining a solid language, Delphi started to stagnate and fall behind Java and C# in terms of features and the integration of new technologies at that time. Although Delphi version 7 was a particularly good release (as far as I remember, was it version 4 that was always rubbish, even after two service packs?) it just wasn't enough. The job ads for Delphi were starting to decline, and its hey-day was now behind it. As a develper, that raises some concerns, and for me it was time to move on.

Java became my language of choice. Not only was the JDK free, but I'd been playing around with it since late 1996 in the version 1.0 builds. So it was that I moved to a job offering cross training to Java from Delphi, which after some complaining eventually happened.

And so Java started to pay the bills as well, if not better, than Delphi had in previous years. However apart from a few notable times (such as when James Talbot introduced me to the Spring Framework, thanks James) Java development was never as much fun as Delphi in those early days.

Don't get me wrong, I like Java a lot, but it rarely had the comparatively instant gratification moments that Delphi often provided. I think a lot of this can be down to two points: My Java programming involved little Swing (or indeed any visual) work, so whilst it was immensley satisfying to get JSP being constructed from a web server, or process huge amounts of financial data, it didn't grab my imagination in quite the same way. The second problem is that Java does so much I could never hope to be comfortable with it all. It is true that this is the case with Delphi (DCOM required me to grab a book and ISAPI extensions have never been produced by this particular mortal) and also C#, but with Java you always felt like you should know it all, it's not enough to just be aware it exists until you need to use it.

So now I'm spending a proprtion of my work time developing in C#, and to my great delight I find it has some of the same highlights as Delphi used to, only more applicable to the world of now. I don't wish to be unfair to the latest versions of Delphi, which I believe are very good, but the whole community around C# is massive, and it's not the dog of a language that Visual Basic was (is?)

I'm now re-learning how to implement things like threading in yet another language. I don't regret this for a moment though as not only may the approaches to some things have changed over the years, but I also believe that in doing this you always get a little deeper understanding, and hopefully wash away a few of the bad habits you may have picked up over the years.

One of the new areas I'm interested in investigating further is the Windows Presentation Framework (WPF or Avalon to some). This looks similar in many basic ways to Swing in Java, only it feels like a later revision. As a result hopefully I'll be able to take back some of what I learn to future Swing applications.

WPF is still quite new (having been released alongside Windows Vista) and the plan is eventually to make this and Silverlight one and the same thing (or so I believe, at the moment Silverlight uses a subset of WPF).

The down side on this is that a lot of the announcements about Silverlight 4 seem to be moving away from cross platform solutions towards Windows only technologies such as COM. I guess time will tell on that front.

WPF is also a core component of Windows 7, so there's significant opportunity there. Also, I believe, the Windows Mobile 7 UI will be rendered in WPF... that will be awesome if/when it happens. And if they start to add the tools into Visual Studio Express from installation then the uptake of WM7 will only improve in my opinion (you can code Windows Mobile Apps without Visual Studio, but it's a lot more complicated without Visual Studio Professional of above. See my later Windows 7 follow-up post in a few days for more of what I think about this).

There's lots of new things being offered by .Net and C# that I need to look into, if not become an expert in, but just to be aware and not completely ignorant of them. The way I'm approaching this is far more relaxed than Java, so hopefully it will be more fun and I'll achieve even more, then take some of the positives back to Java some time in the future.

At last I am once again excited about the possibilities. I just need to shed some (or all) of this Delphi work to give it more of the time it deserves. Oh well, needs must.