I'm remembering the Redmine slide from Zach Holman's talk, where he makes fun of the overly complicated Redmine issue creation screen, compared to what the GitHub screen looked like (back in 2011).
That's not how it works. It is going the way of Jira. What you mean is if it keeps going that way it will be Jira.
Until we get to the point we can decide things are finished and move on to other problems, everything will turn into Jira eventually. It's like entropy. We have the power to stop it, but it wouldn't guarantee those sweet, sweet growth targets.
The problem is Microsoft has two products in this space but everyone hates Azure DevOps which is supposed to be a JIRA competitor and GitHub is where all the momentum is. They'd love to ditch having to maintain ADO and GitHub long term but that means crapifying GitHub so they can migrate their ADO customers. At the same time they're hoping to pull Atlassian customers that use JIRA and GitHub to ditch the latter and just entirely be on GitHub.
What well end up with is a service that sucks to use for all cases.
ADO almost made me want to use Jira instead.
Turned out that,
while the product it not great,
the pain I experienced using it had more to do with corporate processes than the software itself.
I really dislike the use of former and latter. It’s confusing, people get it wrong, and it can be interpreted at least two different ways to make it ambiguous.
Overall, the claim above, as written, is a rather generalized prediction, not an inevitability.
Enterprise buying power and expectations create various pressures, sure. But there are other pressures and factors that need to be accounted for such as demands from other types of customers and the company’s expertise with what has worked well so far (simpler is better, compared to Jira).
Entropy is a law of physics, sure, but the ways and degrees to which it manifests in software is far from obvious or inevitable.
We live in a world of possible future scenarios, not of narrative-based certainties.
I predict GitHub Issues will remain considerably simpler than Jira for the next five years at least. As code analysis tools improve (traditional static analysis as well as LLM-based), I think we will see a lot of experimentation in this space.
GitHub is free for most people, so they are only beholden to large organizations and their feature requests. This is how it seems now - a feature creep by committee. It's getting more and more bloated and unwieldy, sort of like what happened to AWS.
It is less simple, but nowhere Jira-level, and yet still more useful (for my team and I).
We've been using GH Projects at my current org and program for two years. The one feature I wish it had was nested issues.
In Jira, you had Epic > Task > Sub-task and that's it. With GH, you can just have issue > issue > issue (insert "it's turtles/issues all the day down meme"). So it's more powerful, but can be ignored by folks who don't want to do this.
This is fiction considering Microsoft is extensively using Azure DevOps internally and is still developing it. Moving projects away from it and to GitHub is impossible because they're incredibly far from having feature parity.
Feature parity is probably not required as long as the different teams are able to adapt their workflow to GitHub's approach. Anecdotally, every employee from Microsoft I've talk to about this point during the last two years keep telling me that ADO is over.
Feature parity is absolutely required. We are ADO customer because A) Inertia and B) GH Actions is nowhere close to features of ADO Pipelines.
Every conversation we have with Microsoft about our ADO -> GH migration is either get GH to feature party or if you force us to migrate, we will evaluate ALL our options.
People have been saying this for half a decade, and fearmongering everyone into moving elsewhere (sadly my team fell for it).
Azure Devops is such an underrated tool, it's a shame that it's being ignored by Microsoft. Not only that, but they're enshittifying it by turning into Github. I kid you not, it actually went backwards in terms of features. E.g. Instead of nifty UIs that was implemented for their pipelines, we now instead have to write shitty yaml files with absolutely no validation and guidance. This is the same company that (re)wrote Word, Excel and Powerpoint in the browser! The mental whiplash from witnessing this is very jarring.
