Hew, I feel your pain, man. I have also been bitten by some of these problems.
But, seeing MDGs efforts and this amazing community still makes me believe in that meteor-vision I saw, when I first came in touch with Meteor.
Personally I’d like to see these things from meteor:
Introduction, Communication and Learning Experience
- A noob-friendly tutorial + app:
Not there yet? From what I hear the current tutorial todo app is almost too much for beginners to take. And I must say, I can understand it.
EDIT: There is the simple todo app and the tutorial for that.
- A production-ready tutorial + app:
Check. I guess the current tutorial todo app meets that requirement.
EDIT: There is the full todo app, which meets that requirement. But no tutorial on it yet.
- A solid guide that is always at my side:
Check! Thanks to the amazing work of @sashko and everyone else involved.
- An strong, friendly and active community to learn from:
Check! Love this forum, the people and their work. Just awesome!
- A clear exchange between MDG and the community:
Check? Yes, there have been problems with communication. But things like the transmission show (huge props @benstr) where a perfect step in the right direction, if you ask me. I enjoy it every time.
Developer Experience
- An overall ease-of-use vision:
This is where I’m a bit worried about.
First: The new import/export approach makes things more difficult for beginners, but the load order problem needed solving. And renaming your files to do that was probably never a very good option.
Second: Having npm AND atmosphere packages also make things more difficult to grasp for beginners. But then again the wrapper packages on atmosphere where no real solution as well.
So both necessary steps I guess.
But now add Apollo to the mix + three different view layers + the choice between the official structure or mantra + meteors special folders + the deploy command gone and Galaxy not really being a “one command deploy” …
I don’t know.
Isn’t this becoming something very “fragmented” and complex?
But to be fair, I also don’t have THE solution for all of this.
I just wish that the ease-of-use vision stays one of the main priorities.
It was what set meteor miles apart from the rest.
- Rapid prototyping:
Check. This still works and I love it.
Regarding the future I really like @awatson1978 's idea of a prototyping package (which might include insecure, autopublish and autoimport). Really nice I think.
Regarding the removed deploy free tier (RIP): I think this should come back in some form. I know that’s just SO much easier said than done, but it was a real killer feature. One command. And it’s live. Ready to present it to clients. If they want it, remove the prototyping package, finish the app and seamlessly transition to Galaxy. Done. That would be insanely cool.
- Different Databases:
Apollo is coming. Yaaay!
###Deployment
- One command deploy for galaxy:
From a beginners perspective it’s just discouraging and tough to understand why you suddenly need two different services (galaxy & mlab) to get you app online.
Locally your app is running out of the box with meteor. Nice. But when you want to deploy, this suddenly changes? I wish this would be different.
Especially when the guys at scalingo seem to be able to offer exactly that, with a small team and a fairly low price. So it’s possible I guess?
Anyway, all personal opinion and apart from all the flaws I still believe in Meteor and MDG and will continue to recommend it. Looking forward to the future