In a talkback about an unrelated post, a reader (Nikola) suggests abandoning Delphi for an open source tool like FreePascal/Lazarus claiming it is impossible to sell an IDE if you have to compete with a monopolist (Microsoft) and many other existing free offerings. In short, I disagree, but the issue deserves a longer answer.

Let's first examine why open source softare exists. There are many reasons. Many OS projects were started by groups of developers with a common need (for example, GNU/Linux). Other OS projects are totally abandoned proprietary tools the users community keeps working on (like InstantObjects in the Delphi components world). In other cases, the owner open sources a software to get it out of a small niche and plans to live on consulting, training services, and possibly a high-end version of the tool (there are many examples!). Finally, another category is that of software aimed to clone an existing paid offering that is monopolistic, too expensive, not responding to user needs (MS Office vs. Open Office come to mind). Far from being exhaustive or complete, this gives you a starting point for discussing.

Will Delphi fall in any of these categories? Right now, no. Could the NewCo open source it and live from services? I really doubt? Can we all help building a Delphi clone because Borland/NewCO is monopolistic not driven by user needs? I disagree, but opinions can vary. Certainly compiling Object Pascal to new platforms (like the Mac) is something only FreePascal supports right now. Is sofware development a Microsoft monopoly? Absolutely not, with tons of other languages and tools widely used (from Java to scripting languages to the only Win32 development tool left, Delphi!).

Of course, to draw any conclusion we'll have to see who buys Delphi or if a new company is formed (something "detractors" seems to downplay but I see as the best future for Delphi). We'll judge on facts, but for now I'll keep supporting Borland Delphi with my books, articles, and work. My clients use it and need help on it. Period.

There is an interesting scenario I've mentioned in my "Delphi dream" post. A community driven IDE. Delphi 2005 and 2006 have benefitted from contributions from the community in areas like routines optimization in which Borland has decided not to invest, but the FastCode project has decided to put a lot of effort. This could become more and more relevant in the future, but is already now. There are tons of third part components, libraries, IDE plugin, you can use to customize Delphi they way you need it. The community is working today and has worked for years to make Delphi a better tool. It has supported niche usages Borland has not reason to go after. Let's Borland/NewCo build the compiler (and move it to other platofmrs) and the core of the IDE, and we'll keep providing added value on top of this base infrastructure. The infrastructure should be properly priced (and a free GPL version would be nice), use a subscription model, revived, etc.

I think we still need a bunch of top-notch engineers paid for their great work, like those sitting on their desks in Scott's Valley working on 2006 patches and Highlander while we blog and chat about the product they build and we use. Long live Delphi!