I agree this could've been "solved" by implementing organization-wide labels, instead of creating a new field that is also limited to issues only (not available for PRs) as mentioned by @probitcarwyn.

Thinking about it a bit more, issue types could have been implemented as namespaced labels where types are a namespace (optional at the discretion of the organisation). You often see namespaces emulated via naming conventions like type:bug or type-bug. Existing labels would then just be in the default namespace, and something tidy could be done to make a given namespace repo or org scoped.

There's another parallel to labels too, project fields, that also annotate issues with metadata, but only if the issues are in a project. The main gotcha with these if that if you remove an issue from a project, this metadata is lost. It's causing quite some confusion for us now having these somewhat disjoint metadata sources, each with their own semantics. Arguably project fields are the richest, but lacking a multi-valued option which labels have.

A similar dichotomy exists between milestones, and project based date fields with sub-issues, again with different semantics.

It would be interesting to see the thinking behind the data model and semantics of these features, especially how the semantics relate to each other. Was this considered? I do wonder if a more formal theoretical basis for these features would be beneficial?

Thinking ahead to issue workflows (which has vanished from the roadmap) i do wonder if all this makes that task much more complex than it needs to be. The lack of workflows (which are really allowed metadata state transitions) is the biggest bugbear we have with GH issues at the moment. Workflows and especially enforced workflows being a big requirement for larger organisations that have certifications to consider. For example, there's no way to block issues progressing through a project board without some pre-condition being met. Actions would be great for this, but all you can do to enforce is write a convoluted action using the APIs to move things back after they move when they shouldn't. Bots are a sticking plaster over this stuff.

The thread in the initial issue (during the opt-in phase) https://github.com/orgs/community/discussions/139933#discussioncomment-10885036 touches the same problem. "Issue types" in its current shape seems harmful (less is more).

Org level labels (that appear in all repos) combined with namespaced/nested labels

This would be great. In our case, we have have 3 kind of issue labels: a. Product scope labels (e.g. billing, onboarding) b. Prioritization type labels (e.g. regression, bug, new features, priority: high) c. Skill / kind of work labels (e.g. design, Safari). The notion of issues type feels absurd in our context (too opinionated to help).

Off-topic: Going to amazing for us would be to be able to subscribe to notifications that are only about new issues with specific labels so product owners can be aware of new (a.) Product scope labels.

https://github.com/orgs/community/discussions/16682 scoped labels solve this entirely without added noise. It's the only reason my org sticks to Gitlab at this point.

I agree 100% with using scoped labels + organization labels. Would be a much more flexible solution. Our org has lots of labels already that are in the <category>:<entry> format, and we've been annoyed for years at the lack of ability to sync common tags between all our repos. This solution for types is a step in the right direction at least, but the two solutions described above would have solved all our labeling issues at once. As it stands we still can't sync the rest of our labels across our repos without external tools.

Thank you for this feedback! We are aware of this and have plans to bring issue types to this page as well. We will keep you posted on when you can expect this as we know it is an important piece of information to see as you work with milestones!

We do not have a concrete timeline quite yet as we make plans to migrate the milestones, labels, and /issues dashboard pages, but as soon as we have more details we'll be sure to share them with you all!

Appreciate all of the feedback on this! We will keep you posted for any updates.

@rileybroughten Labels are a workaround. Making it an explicit field is the right way to go, what you're observing here is just the usual migration pains + change aversion.

I disagree @LeaVerou , if you think about the actual data model underneath this, and especially how you might want certain labels to only apply to certain types, then scoping/nesting is what you end up in that case too. Once you get to there you're back to questionable value of a disjoint metadata field, especially when you think ahead to how you might implement workflows (state machines across label states). Having been on the private preview for a few months now with Issue Types, when to use which (label vs type) and the lack of types on PRs has been a consistent topic of confusion.

I like having enforcement of a singular "label". I don't want people to label things as feature and bug. Having a field for singular classification is nice on top of the labels to further describe the issue.

We too need an option to disable this field. We deleted the bug, feature and task labels, and we don't want this field. The previous labels and this field provide no value to our workflows for managing issue, but they do create cognitive load for our users filing issues.

Good suggestion! We're working on getting the public documentation updated to include this field.

@rileybroughten projects are not supported in md files issue templates, as they are supported in yml file of issue templates, any ETA?

We don't have an ETA at the moment, I've raised this again with the team and will keep you posted with any updates 👍

Docs have been updated to include type 👍

EDIT: There is a bug in the mobile app: When I define the issue type in my form, the option will not show up in the mobile selection menu!

Btw feedback in general: Despite many people not liking that it was not implemented as org-wide labels, I love the new issue type. I assume that one difference is that one issue can have multiple labels but only one type, so this differentiation gives more emphasis to the type. ❤️

As a reference: https://github.com/trueberryless-org/.github/blob/main/.github%2FISSUE_TEMPLATE%2F0_bug_report.yaml#L5

Try creating an issue in this repo... (bug only exists on mobile devices)

Thank you for this feedback, we'll work on getting the documentation updated!

When making edits to your issue forms, you'll see a note on the right hand side for configuration options.

Docs have been updated!

Thanks!

P.S. On the two-week-old comment — I almost never start these via web UI. I'm pretty sure the primary audience of that web view is the first time users. Like with many similar things, I treat these forms as shareable configs. So when I suggest changes in other repos, I look into what I've got in my typical configs and copy from there. Rarely, I'd go check the docs. But it never occurs to me to use the “create file” web page. Once I have a config that I like the shape of, I just copy it into many/all my projects and go from there. There's just no mental model for “has anything changed in the docs?”

@rileybroughten Can you share the reasoning behind not expanding this to user accounts? I have many private and public repos that aren't part of an org that would greatly benefit from having issue types but don't need to be an org. Yes, in theory, I could create an org for myself and move those repos into it, but that seems counterintuitive and a lot of hassle for something that should simply be available.

Thank you for sharing this feedback @scottdorman. We wanted to provide these to help organizations standardize and classify issues across their repositories, especially in the initial release. We're discussing this internally and will provide updates on any plans to expand them to user-owned repositories in this discussion.

If they aren't available for user accounts, can we disable them in the UI? I have a private personal repo that still shows it as a filter, as way to categorize existing ones, and on issue creation. Since it can't be used, its a bit confusing to have there (I spent a few minutes just now trying to see how to configure it for my repo and/or user).

I would like to advocate for making issue types available for personal repositories as well, given that personal repos can also be large and require extensive issue management.

For example, take my project. Although this is a personal repository, it has over a thousand issues. Currently, I have to implement issue types using labels like type:Bug. Having Jira-like features such as issue types and priority would greatly help in organizing and tracking these issues. I've been looking forward to these features for a long time.

As for why I don’t just create an organization, this repository would be the only one in that org, and I’m the only active maintainer. That setup doesn’t make sense for me. It would be great if I could simply use issue types in my personal repository.

Why do you discriminate against personal repositories?
I am using this feature in my EMU account and dying for this feature in my repositories.

Thanks for this feedback! You can also choose to show or hide the Type field on items in your board view.

oh, I realize that, but it displays it mixed in with the labels which kind of defeat the purpose of having them separate (and honestly render them useless if I understand the design purpose of quickly being able to identify a toast type).
The other option might be to display the type tag between the repo and assignee .

Thank you for this suggestion, we have it noted 👍

+1 for a colour wheel or field for html hex code

+1

I agree. We need the ability to choose the color just like with labels. But a color wheel and vertical saturation bar would be excellent to make it easy to quickly select colors so we do not have to deal with the hex codes

As far as bulk converting there are a few options:

• Utilize the bulk editing capabilities in a project using copy/paste or dragging table cells (see documentation)
• Use the filter bar and quick filter dropdowns to make bulk edits to existing issues.
  
• Use the API to do this programmatically (see "Automating issue types" section above, and note REST API support is coming soon!)

@rileybroughten this is what we did too, but it would be nice to have a way to override the pagination so that folks don't need to do it once per page (and large repos can have A LOT of pages!). Especially since it takes some time to apply bulk actions, and one cannot trigger another bulk action while one is already running.

but it would be nice to have a way to override the pagination so that folks don't need to do it once per page

This feedback has been open since the opt-in. Still not working. We have some 10k issues- having everybody write custom tooling to do this at scale is a no-go.

Hi @rileybroughten, unlike in your screenshot, the Issue Types option is no longer available for me. I saw it like two weeks ago still... 🤔

How to categorize issues in bulk and tag them correctly when bug or feature labels were used before?

We are investigating this, see https://github.com/orgs/community/discussions/148715#discussioncomment-12901668

Development is ongoing for managing issue types with the REST API - we hope to release this to you all in the coming weeks 😁

We're starting to roll out support for the REST API, if you'd like to test this early please provide the name of your organization(s) and we can get you access this week. Thank you!

Check out the changelog which details support for issue types in the REST API. Let us know if you have questions or feedback! 🚀

Check out the changelog which details support for issue types in the REST API. Let us know if you have questions or feedback! 🚀

Thank you!

Works like a charm.

I already did my jira migration to github using the graphql api, but thank you for doing this nonetheless!

Thank you for this feedback! They are scoped to only being present on issues to help you categorize and classify your work, break it down if needed using sub-issues, and report on it so you can understand the breakdown and progress of work across repositories. Then creating or linking a pull request helps you track the code changes and fixes to resolve and close out the bug, task, etc.

Hello,
I agree with @Seldaek, it would be great if the types of issues could be communicated to the PRs, as this would allow the changelog to be generated with PRs sorted by type directly.

Thanks !

This has been called scoped labels or nested labels in other comments on this thread. I do think this is what Types should be and it should be unified with labels. See also:

• https://github.com/orgs/community/discussions/16682

I agree it would be nice for the list to be customizable.

Could you elaborate how you would like this be done?

One bug we noticed: it seems that if an issue is set to an issue type of "Bug" it doesn't show up in the Bugs section, as that is keyed on "bug" labels only. Ideally, it would be an OR.

What is the Bugs section in this case? If you could provide more details or screenshots on this setup that would be great!

The Refined GitHub extension adds a Bugs section, maybe that's what is meant here?

I expect "PR Type" would add confusion in my org so I'll make the case against it. A PR can have many commits and address multiple issues. For example, both a bug and an enhancement.

defaut issue type examples of Bug and Enhancement overlap with the default lables of Bug and Enhancement

This feature is in preview. Once it lands the defaults will change.

because in our repository, both PRs and issues use the same classification labels

Why is that necessary? If you want to know whether a PR is fixing a bug, isn't that implied by what issue the PR is closing?

Why is that necessary? If you want to know whether a PR is fixing a bug, isn't that implied by what issue the PR is closing?

But often, a PR that fixes a bug doesn't close any issues at all. Then we can only use classification labels to classify every kind of PRs.

I don't know if the issue types confusion is a development workflow I've never heard of or lack of information on why issues have types. It has popped up a bit throughout this thread. In the case of the latter.

• Issues are work that needs to be done

• PR are work that is done

With that in mind, labels serve two purpose: to classify for reporting and for an aid in filtering/finding.

Issues are a little more complex. Consider the entire development workflow. Not all stakeholders are developers. So how does a non-technical person communicate a feature to be added to the software? this is where issue types come into play. The common issue types are as follows (and in the hierarchical order):

• Epic = Epic User story. language is non-technical in nature
  An epic user story is generally for an addition that has a lot of moving parts, features/components etc.
• Story = User story. language is non-technical in nature
  A user story describes a small area.
• Feature = Something to be added that more likely than not has many moving parts. language may or may not be non-technical in nature
  A feature, generally, would be used to describe a component/module.
• Task = A single item to be added to the software. language is technical in nature
  A single item to be added to the software

The remaining common issue type is:

• Bug = A problem with the software. language can be both non-technical and technical in nature
  A bug describes something wrong with the software

An Epic, Story and depending on your policies, possibly Feature are all issue types that are generally used by non-developers to communicate a part (or large part) of the software for development. A Bug also falls into the same area as well.

Once you have those issue types the dev workflow moves towards a developer creating a Feature or Task issue that contains the technical details of the work to be done. Outside of planning nearly all of the developers tickets they work on will be Task and Bug issues. Those are the issue that will have a PR directly linked to it. Completion of these issues which would generally have a parent issue type of Feature, story or Epic would then enable the parent to be closed.

This workflow of development opens up the possibility of reporting how long it took to create a feature (duration the parent was open for). In addition and particularly when your software becomes more complex, you can't work on a single feature per release. You can however complete some tasks as part of a release.

I'm personally of the opinion that issue types are an "about time" feature, with the only missing feature being issue dependencies (blocked, blocked by and related). If you would like to see a repo of how issue types can be used, please feel free to look at one of my repos. At the end of the day, issue types are your decision and I hope I've been able to assist with the general gist of what they generally are.

I like your reply @jon-nfc , the split in thinking as to which role (issure creator vs issue curator) makes use of labels vs types is useful. This is especially the case if you then base your issue template on the type selected.

Perhaps the overlaps in default values for types vs lablels (both have bug and enhancement for example) needs consideration, with some docs on example use cases and flows.

I like your reply @jon-nfc , the split in thinking as to which role (issue creator vs issue curator) makes use of labels vs types is useful. This is especially the case if you then base your issue template on the type selected.

I can't take the credit for what I've explained as it's part of the common development workflow. It does have a name, although it escapes me.

Perhaps the overlaps in default values for types vs lablels (both have bug and enhancement for example) needs consideration, with some docs on example use cases and flows.

Although the labels do overlap with types, they are sensible defaults. If the labels that overlap issue types were removed, with the exception of using conventional commits, how do you determine/filter if a PR fixed a bug? was an enhancement? etc.

I don't know if documentation will help or if it's the answer. Github is a platform for software developers. With that in mind, there is no requirement to document issue types and label usage (outside of their existence) as developers have been professionally trained. However this far from reality, as Github is also used by self taught "software developers." This groups knowledge will range from the "I know boats" through to "very knowledgeable" More recently the IT world has moved into using Github, due to the latest buzzword "GitOPS" I believe this is the group that will possibly struggle with the software DevOPS concepts. Although again, I don't know if documentation will fix the underlying issue of a lack of knowledge.

We have it on our backlog to be able to reorder this list, I'll keep you posted with updates as we start working on it!

Issue types are currently available for organization-owned repositories only. We'll provide updates in this discussion for any plans to expand them to user-owned repositories.

Let personal repositories use this feature as soon as possible.
I am dying for this feature in my repositories.

Thank you for this feedback @MattMcAdams - we will provide updates on this discussion with plans and timing of when you can expect to see issue types supported in user-owned repositories as well

I agree with @MattMcAdams on this. I, for one, have a personal organization and manage many different projects, as does a client with a similar setup. Not all projects are handled the same, and are not managed the same. Certain issue types do not apply to one project, but they do to another project. It would be a great feature to have this at the repo level as well.

We don't have immediate plans to address this, but I've passed this feedback along to the team. Thank you!

+1 for this feature!