I also totally agree with you and with @sacha . I don’t think the “FUD” is unjustified in most cases and I don’t really think it’s FUD. He makes very valid points on most accounts.
3 Likes
I don’t think his premise is valid. He is saying Meteor has to move either direction (ddp or graphql) or it will decline and those are mutually exclusive. I think a third option is to focus on core and empower the community to manage the rest. Those are stacks that serve different needs. That hard earned flexibility (3 plus years of coding by Ben) has some drawbacks but it should be embraced as a major strength not a weakness. All the other frameworks will decline when their view layers lose their hype so stepping back and focusing on solid flexible foundation is a sensible long-term strategy.
6 Likes
IMHO, Meteor is the best prototyping solution and also the best “get-shit-done-quickly” framework for startups. This is also the reason why I chose it for my most recent startup, although I pretty well knew it’s in decline (somehow) - but after all: There was just nothing like it out there.
I personally think, Meteor could shine and grow again if it came back to this vision, instead of chasing the enterprise market. I have a strong enterprise background, and my gut-feeling is it won’t ever succeed in this world.
But why should it, if it’s so strong in other use-cases?!
You might argue the biggest challenge is to find a sustainable business model for this. But hey, as @sacha already pointed out in his blog post: WordPress managed to do this, too.
Meteor could become the go-to solution for any entrepreneur who needs “real” business logic that goes beyond a simple solution like WordPress. And then give them a quick and simple hosting solution, just like Heroku, for an affordable price (instead of pricing on enterprise levels, like Galaxy). The average deal size might be lower, but it’s a pretty interesting, scalable market.
4 Likes
Meteor is an incredible platform, and I’m excited for this announcement and what’s next. Thanks to the team at Tiny for believing in what this platform is capable of doing, especially for developers who work with startups and entrepreneurs!
4 Likes
Just to clarify a point I made in my post: the problem with that “Meteor 1.0” stack is not the technologies in themselves – I’ll be the first to say that many of the things I’ve done with Meteor would just not have been possible with another stack. Minimongo is a great experience, far simpler and easier to learn than Apollo/GraphQL.
The issue is that these technologies are not being used anywhere else, and even MDG themselves are not using them internally anymore.
So the goal should not be to hold on for dear life and keep these technologies on life support indefinitely; instead it should be to recreate the awesome developer experience of “classic” Meteor but this time with well-established technologies like React and Apollo.
I mean can anybody seriously argue that if MDG was starting the whole Meteor project today, they would still decide to develop their own view layer instead of simply picking React or Vue?
Not to stretch the political metaphor too far, but I feel like the “conservative” wing of the community is so afraid to lose what they have now that they don’t dare to dream of a better future where not only is Meteor just as easy to use as ever; but it’s also now a first-class citizen in the global JavaScript ecosystem instead of sitting behind a wall…
15 Likes
This is dead on. The value proposition should be about making these technologies as accessible as possible, which is what Meteor did initially.
9 Likes
Is this not a form of the logical fallacy appeal to popularity?
Popularity does not imply correctness or quality.
Popularity is more a function of marketing, PR and evangelism and I don’t think MDG will ever share a pedestal with Saatchi & Saatchi.
MDG have their own business goals and they chose to shift their focus to Apollo/GraphQL.
The only thing that matters is that the Meteor classic technologies are useful and value-adding to businesses, have the ability to be maintained and offer something superior to alternatives.
The bedrock technologies used to implement Meteor classic like Node.js, MongoDB and Websockets are all alive and well and are not deprecated in any way, so what is the maintenance obstacle? There isn’t one.
Why recreate when we already have an excellent existing solution that works well today?
I looked at React and Apollo when they came out and did not see how we would gain anything from switching to them. I doubt React would ever have become as popular as it has if it did not have a global corporate behemoth like Facebook pushing it. I now hear some people saying that a new framework called Svelte is going to kill React. I will keep watching from the spectator stand, eating popcorn.
Blaze was developed because it fulfilled a need that was not provided by the alternatives existing at the time. Today, people are free to use whatever view layer they want with Meteor.
We continue to use Blaze.
My businesses’ and a number of client’s solutions are built using Meteor, and we are making a living from these, so you bet we have an interest in preventing the erosion of support for frivolous reasons.
Meteor has continued to get better and better whilst preserving the functionality that made people choose to adopt it in the first place. A large chunk of the “wall” was demolished long ago in version 1.3 when full NPM ecosystem support was introduced.
I read Hacker News and other sites daily to keep me informed about new programming languages, frameworks and tools and their best use cases.
To date, the Meteor classic stack has been the most productive framework we have ever used and remains our first choice for new projects.
No other framework has allowed us to develop reliable SaaS solutions quickly, with the fewest bugs, without requiring us to seek out hotshot programmers.
I eagerly look forward to the imminent release of Meteor 1.9 which will allow us to enjoy the performance benefits of Node.js 12.
12 Likes
Congratulation! Hope Meteor became more and more powerful!
中国开发者发来贺电！
3 Likes
Popularity does not imply correctness or quality.
This was exactly the reason why I have chosen Vue over react when was afraid of Blaze not being maintained. And still have very warm feelings to Blaze.
I read Hacker News and other sites daily to keep me informed about new programming languages, frameworks and tools and their best use cases.
Exactly. Lots of fuss around, no substance on how to solve business tasks. Sometimes I feel most of people in broader JS community are just tech junkies and not entrepreneurs, which to my feeling contradicts to Meteor community. Man, MDG could have bet on small entrepreneurs and natural growth with all the participants interested in growing together, on the contrary to betting on enterprise (but that I believe is Silicon Valley and Investors relations things). Yet, not being able to make money with such a great product and such resourceful and deep thinking community is a shame of MDG (with great respect to technical wisdom).
11 Likes
Yes, I think this was the core issue. MDG wanted to use VC money to develop core web infrastructure. And this does not work. This is a problem of today’s investment methods available (non-profit, VC, research grants). They are bad for core infrastructure. VCs want return on investment and developing core infrastructure sadly cannot really provide that. So I think they were simply pressured to go into Galaxy first, enterprise focused, and then to reinvent themselves with Apollo.
4 Likes
I’ve the exact same sentiment about the NodeJS ecosystem, in theory nodejs should be the best choice for entrepreneurs since you’ve to use only one language but it seems to be dominated by cooperate tech/hype. It is different than php and RoR where you feel it has a strong and focused entrepreneurial community. Meteor has that entrepreneurial vibe within its community which I really like, but the community just need a bit more attention and nurturing. MDG managed to create the nodejs entrepreneurial nest which is a great achievement but then they pivoted to enterprise tech perhaps under VCs pressure as @mitar suggested.
4 Likes
Exactly. This makes a lot of sense. I would love to see an out-of-the-box Apollo integration supporting subscriptions, as simple to setup as Meteor.publish . 
5 Likes
This has been done. 2 years ago. https://github.com/cult-of-coders/apollo-live-server including the out of the box integration. “meteor add cultofcoders:apollo”
3 Likes
So the goal should not be to hold on for dear life and keep these technologies on life support indefinitely; instead it should be to recreate the awesome developer experience of “classic” Meteor but this time with well-established technologies like React and Apollo.
Clearly stated, that’s completely the truth.
And that won’t prevent people using Meteor 1.x to keep their app living on it for another long time.
3 Likes
Borland chased the enterprise market and look at what happened to them!
1 Like
I would be careful calling Apollo well-established. It’s a 3 year old project. Agreed, it seems to be faring better because of some uptake by big players (airbnb etc). But it’s still very unknown outside of the typical silicon valley company. And the typical silicon valley company is also known for continuously changing technology stacks to the hype of the day. So Apollo might be as fast out as it got in. Only time will tell if Apollo is here to stay.
I think the major reason Apollo is getting more traction is that you can gradually introduce it into an existing project, while they can keep their view layer, database, server platform largely untouched. Whereas Meteor is basically an all-or-nothing solution. That makes Meteor only a good choice for rewrites-from-scratch and new projects. But those are much more rare and it’ll be much harder to convince people to switch their entire stack (A lot of people don’t even know Meteor is just a layer on top of Node and you can do all your regular node stuff).
7 Likes
Hm, it’s far more known and used already but by just some silicon valley startups. Many of my clients both here in France and elsewhere, startups, medium size companies and some big corporates already use it, and not only partially. It won’t take over REST in so few years, if ever, but it’s not jsut Apollo, it’s Relay, GQL and Facebook teams behind, it’s gained a huge momentum already and represent more than a smal technical improvment. To me, GQL is to back-end dev the same paradigm shift that component oriented programing has been to front end development.
4 Likes
would be perfect for tiny to integrate that into core! (seriously, mdg should have done that 2 years ago)
2 Likes
This is one of the things about modern tech entrepreneurship that I find a bit weird. So many companies settle on 1 product to sell. Imagine if IBM only had one product, or if Nintendo only sold one video game. I always found it weird that silicon valley companies rarely work on more than one product.
3 Likes
Agreed. I’m already receiving emails from clients and stakeholders who are getting worried by his article. I wrote an article in response.
17 Likes