1000 widgets in 8 months...

Printer-friendly version

A friend of mine pointed me to this post by Nick Hodges. I wonder if it is a hint about what happened at Embarcadero with Delphi XE. It should have been the Windows, Mac and Linux cross compiler, with a new CLX (the 1000 widgets?), which instead just before release turned out to be just an interim release with some third party products integrated, a primitive SVN integration, and some enhancements here and there.

IMHO it really looks they set a schedule too aggressive, and weren't able to match it, management pushed an irrealistic plan - I too was skeptical about their ability to deliver a Windows+Mac+Linux cross compiler first, than even a Windows+Mac cross compiler in less than a year

It's some years I am telling Delphi (and the whole RAD studio) can no longer sustain a yearly release cycle. Maybe Prism can, but it is only a compiler, not a whole compiler(s) + IDE + framework ecosystem. If you say that, the answer is always "they have to deliver something to the SA purchasers". To me this still means "Delphi is our cash cow", not they care about their customers. It looks nothing has changed since Borland times - no, wait, the new upgrade policies are even worse for those who upgrade seldom (and to clear suspicions, I am not talking about me, the only releases I didn't buy till now are Delphi 8 and Delphi 2009, I am going to skip XE too) because their applications don't need the latest version each year. And if you believe that's wrong, check how much software is out there still compiled with Visual Studio 6.

As I often pointed out, issues were not only in Borland upper management, and weren't left there when the devtools business was sold. I believe that besides Hodges, there are some more people who should leave Delphi, because they failed over and over to deliver an up-to-date development tool.

It was SOX before, now it's SA the finger they are hiding behind. Frankly, I perceive very little value in a yearly release. It often meant poor quality releases, that took two-three releases to get fixed, while new features added show new bugs, and old ones may never gets fixed, even if they heavily impact features like localization. We stopped buying SA long ago. A new toy to play with is always welcome, but when you have to develop and release a large commercial application you can rarely afford to upgrade your main development tool each year, especially if its quality is not adequate, and new features often useless but for simple applications. Getting new versions we are not going to use anyway is really useless. Thereby now we upgrade only to the releases we believe can add actual value to our applications. If that means to wait two-three years no problem - or better, some problems may arise because Delphi can be very late to adopt new features like Unicode or 64 bit code, and the more they try to stick to deliver a release each year, the less breath they have to deliver larger feature changes properly.

I am starting to believe the almoust religious affection many Delphi developers have for their tool is dooming Delphi itself. As long as this revenue generator model works, more or less, BorCodeDero has no reason to change it and move to a model that suits best customers needs. I understand Delphi must pay for its development, but Embarcadero looks to have a better porfolio of produtcs to sustain its revenues, and should seriously think about increasing the time allowed to develop a new release of Delphi. If Embarcadero really needs a steady revenues stream, they could think about different models, for example a maintenance program that really mantains supported releases, instead of releasing just an update pack to fix some bugs and leave many others to the next release or beyond.

Othewise it won't be enough to fire the project manager to deliver an appealing product.

Update: John Kaster has announced he's leaving Embarcadero. I believe this is a good news, as long as people like him are going to be replaced by better ones. There's still one at least who should leave (or retire)....