I was going through my spam email, and came across a typical spam from Embarcadero that made me throw up in my mouth a little.
Instead of releasing 10.4.1 hot on the heels of the horribly failed Rad Studio 10.4, Embarcadero is trying to “sell” it to customers who don’t have a support/upgrade agreement in place.
This is just sick and wrong.
We have a customer support agreement in place at my company, so I think we pay a few thousand a year to get free upgrades, in a very expensive subscription-like model, so my outrage is purely empathetic to those who are lacking such agreements…
But let me just say that Delphi 10.4’s primary advertised features didn’t work…. like at all. Time is money, and my company probably spent $4000-$5000 just paying programmers to evaluate moving towards Delphi 10.4… and the consensus we came to was… “don’t”
The new Code Insight engine feature failed to parse my code more often than it succeeded leading me, and every one I work with, to turn it off completely in favor of the old engine, which was broken 10 years ago, but less so.
The new record initialization feature was implemented while the team seems to have completely ignored the huge bugs that were released in the 10.3 code. That code, combined with the lack of quality fixes in the previous release scared us away from 10.4 entirely… we stuck with 10.3.
What was worse was that the Embarcadero blog squad was showing off how you could use record initializers with inline var declarations to make your code more readable and reduce repetitive coding tasks, yet if you actually bothered to follow the blog examples in your own code, you’d quickly find out that bad compiler code generation would result in your code trashing memory, acting wild and unpredictably, as the memory for those records would not be properly initialized… causing the system to potentially chase wild string/dynamic array pointers… and costing you thousands, or potentially even millions of dollars in damages if it resulted in behavior that was hard to track down.
Embarcadero, we demand an apology.
We also demand a fundamental culture shift in the way your company interacts with customers and sells its products. I gave up on writing QC reports long ago because every QC report I ever wrote was ultimately marked “duplicate” or “cannot reproduce” or at best “fixed”, however, no fix was ever found. And any interaction I had with customer service was generally rude, almost criminally rude (like hey, do you want my money or not?). And interactions with tech support simply demanded that I fill out a bug report, shut up, and wait decades for fixes.
You still haven’t fixed the features you released in Delphi 2006 (you still can’t do VCL glass effects to any degree of usability). Delphi 2006’s entire feature set could literally be accomplished with 12 lines of code… (I know this because I found a code snippet to do what they did before it 2006 was released)
The VCL has gone with virtually no updates since circa 2003 yet you boldly claim that Delphi is the “best” way to write Win32 Apps, “period”. This is all a lie. You are relying heavily on the idea that your few remaining customers are either stupid or over-invested, or stupid and over-invested.
The time to get a lifeboat off this sinking ship passed a long time ago. Now, us long time supporters of your company are just playing the violin.
Figure out how to do business in a dignified manner, Embarcadero… and do it yesterday. Apologize for your garbage code, do right by your customers, and put change in place that ensures future quality of your products.
I haven’t tried 10.4.1 yet, but this is far from the first time Embarcadero/CodeGear/Borland have claimed to have made quality improvements only to release garbage that doesn’t work. So history tells me that nobody over there has done jack nor shit to actually test this software let alone respond to the issues that arose during testing. It is probably just another money grab.
Prove me wrong. Change my mind.