I was recently building apps on the Slack platform and couldn’t help but think that Atlassian has totally dropped the ball on this boondoggle, Forge.
Forge suffers from poor architecture and is not even half baked. There are some deep deep issues with Forge and it may be a good idea to go back to the drawing board.
I encourage the Atlassian team to take a look at the Slack platform and notice how they have thought of all the gaps in Forge
Atlassian has been working 3 years on Forge, and they knew from the get go that it was an awesome idea at many levels, capable of bringing IT to a whole new level: GDPR, ease of deployment, private clouds for the Department of Defense, ability for the dev to focus on the added value, but also a whole new kind of apps that could be built… It’s a trillion-dollars idea.
But as some people say, ideas are free. Implementation is king, and Forge isn’t even good today for basic apps. It became hard before becoming mature. It started on the wrong basis.
Worse, Atlassian managers are proud of saying it’s extremely limited on purpose. I just **** wish Forge could connect in Oauth to our Connect apps, to allow us to transition to Forge smoothly. I also wish we could use SQL in Forge, and our own SpringBoot. Cue the Atlassian leader who will tell us all of that will be possible in Q4 next year, forgetting to say that it will be implemented like a bag or knots.
But in all this debacle, I’ve also looked at how to build Google Docs apps or O365 apps. Concerning Google Docs, they have a built-in IDE, no git (!!!), and they have no built-in monetization. Google Docs didn’t even notice that Atlassian is 10 years more advanced.
A lot of platforms have defects. I keep my hopes up for Atlassian. They can still turn around!
I just wish we were part of their TEAM and they didn’t gloss over our needs.
- The Custom UI is an awesome idea to keep proof/versions of the code that customers used,
- The Forge UI is shit, because it’s impossible to motivate a career developer to make a career in Forge UI, and we told it to Atlassian and they didn’t care, it’s a huge recruitment issue,
- The Storage is shit, even their future ideas of what the storage will become,
- The JS lambdas are meh, in theory they’re a good idea but they’re slow, incontrollable, hard to debug. We need to be able to run whatever we want and not having to estimate whether they will take more than 10s (it’s like being given 640KB of RAM and you have to program everything accordingly). The lack of background threads, the … well this implementation is not the right one.
Please go back to the drawing board. You’re on a trillion-dollars idea, Atlassianers, and you’re dwelling on the wrong implementation.
But in the end, Forge became difficult before becoming mature.