Sounds familiar:
There’s actually a tradition for GitLab to release a new feature every month on the 22nd. And that’s great because GitLab is the most feature-rich git platform currently available.
However, we’ve met a lot of GitLab users on forums complaining about GitLab’s features being “halfway-done.” Not all features are fine-tuned, which leaves some customers unsatisfied.
Sounds familiar…
All too familiar
You might remember I also said that Apple is making a mistake going from an 18 month release cycle to a 12 month one. Sequoia’s introduction last year was a disaster (Installation, Firewall, Network and Applications), iOS 18 bricked iPad M4 Pros that needed to be exchanged, and now Tahoe is also calamitous (2 serious Audio bugs plus at least 20 other major bugs) and can’t even be installed on a Mac Studio M3 Ultra.
Rapid Release is a recipe for disaster.
Apple’s motto used to be “It just works.” - maybe they should just get back to that.
That one is looking like an edge case, as it seems to mainly involve users who had first updated to 15.7. But yes, it’s not a good look.
They won’t be able to until they fire the entire generation of “everyone gets an award” they’ve hired.
The new generation programmers. I had a few which wanted to work for my company. All not usable at all. Software engineering follows always the stable and reliable running release and not the most and best functional situation with many bugs. They are not acceptable. But that this generation needs to learn within the future.
Entry level developers/engineers need proper mentorship and guidance. Without that you’re asking for trouble. I say this from experience and why I’m no longer an electrical engineer. I had shit mentors when I was just out of college and was just expected to ‘know how things work’. I had wonderful software development mentors (one of them being my wife) to learn from when I switched careers.
My daughter (24) is thriving in her software development job in large part because she was always part of our discussions over dinner on our software development problems and solutions. 12 years of robotics teams (as member than coach) further trained her on how how to do things properly - but with plenty of room for mistakes. Thankfully, she’s had good bosses in her company that value training and mentorship.
When we hired developers we looked for people that enjoyed doing ‘side projects’. Showed, to us, that it wasn’t just about money.
This thread is a great illustration of how Xojo’s rapid release model fails
While not JUST a symptom of rapid release its certainly an outcome
A lack of decent unit & regression testing and a steadfast refusal to implement any
While I agree with
in that its fixed the questions Xojo should be asking themselves is
not I say should as I’m pretty sure they wont
Adopting best software engineering practices just isnt a cultural priority there
And it shows
Did when I worked there despite best efforts to do it
And I dont believe that has changed in any way
I love the Debian way:
Stable: The robust way for everything mission critical with predictable time frames, but in most time older version though security updates are still applied.
sid: For everybody, who wants to live “on the edge” and can handle or revert possible errors.
While all servers and workstations in production are on stable (or old-stable), my daily driver is on sid. So I am the first one, who see possible bugs and or failures.
Xojo would have to seriously overhaul their work habits to support this in any way
The slow beta release model 
https://forum.xojo.com/t/dbkit-beta-19-is-now-available/86626
Sounds like the CEO writing code himself… wait… really?!? REALLY!!!
Heh. Looking back on TOF, beta 6 was announced August of 2023 so in 2 years he’s managed to put out 13 beta versions and is doing a major rewrite. What a joke. At this rate the new version will be out, and usable, in 2027 at the earliest.
WHAT???
And why is this a kit? Why not built in?
If I had to guess this is a vanity project the engineers are encouraging Geoff to do to keep him from micro managing them. The rewrite is typical Xojo though: start coding without thinking through the implications and then rewrite it once they’ve coded themselves into a corner.
It’s been a while since I’ve looked at the code but it was…um…not good. I will never use it because I don’t trust Geoff’s coding abilities.
2 years later it’s still labeled beta and a major rewrite is in the works. How embarrassing.
yes
really
You can even examine the code if you want by downloading the examples and opening Databases
Ion there you will find 2 separate items - one for desktop & one for web
This set up is required because they made it so when you try to bring in items that are NOT compatible with the kind of projects you’re working on they get stripped out 
This defeats the entire purpose of the compatibility flags but hey !
Because they refuse to make controls implement a common interface for data binding there are special subclasses of a huge number of built in controls
And as far as I can see each TABLE requires a connection so you’re going to use up a limited # of db connections very quickly
unlike Active Record which manages with a single db connection
Android Release will be out in 2047…
What a mess!
Another mess!