It’s one thing for a startup to do this while small.
A whole other level of evolution to do it while you’re large. Imagine Apple suddenly publishing their roadmap. This type of organizational neurogenesis done as an adult is impressive.
EDIT: I am in admiration of the change, regardless of opinion on the value of public roadmaps. It’s just rare to see big companies make big changes. Sign of youthful vitality and health, in my opinion.
While it's only for enterprise customers, AWS has a weekly email containing upcoming launches which are under NDA. It also has what amounts to recaps of recent launches, blog posts etc. Only the launches section is not publically available.
The upcoming launches section is literally just an HTML table with two columns: service name on the left and a brief summary on the right like "Add support for X Y.Y". It's almost funny how so much effort goes into fancy blog design when the "real" news is just distributed through a basic plain text email :)
---
To a certain extent, roadmaps become redundant for enterprise customers too in that you start to influence feature request priorities via your dedicated account manager(s)
Vendors may often prioritise a feature requested by medium enterprise to keep them content since if they don't address their features in a timely manager, they may go elsewhere when their contract comes up for renew.
---
I suppose that's why UserVoice type of forums are always a wasteland because the big players can get all the attention?
It's also behind the scenes so perhaps naturally, the result you're left with is acres of "9999 votes for feature X, submitted 14 years ago" and a Product Owner saying "Sorry, no news to report on this just yet"
Apple is a poor choice, because they consider keeping their roadmap under wraps an not only a competitive advantage, but a core part of their company ethos.
Google provides rather detailed roadmaps for GCP, they are really useful. They are under NDA, but customers should be able to get access. I think this recent "C2C" announcement has link for signing up[1]. I was invited, and your account manager should be able to invite you.
Thank you! We work really hard on the GCP roadmap program, keeping it fully accountable, and pushing against interesting transparency limits.
A public GCP roadmap is also in longer plans but there's a few internal flows we want to improve first. I'm especially trying to reduce the number of sites people need to visit, and want to ultimately see it land in a common place where authenticated users also see the non-public / NDA aspects.
And seconding this advice on access. If you are a GCP customer, your account manager is able to add your account for access. My contact info is in my profile and I can reach your account team if you have difficulties.
My first reaction was very different from yours; the desire to dogfood feels excessive, to me, and negates the point in having a public roadmap in the first place.
Am I wrong for presuming that the people who would most care about GitHub's public roadmap probably aren't even GitHub users?
EDIT: Thanks for the replies, everyone. Seems like I hadn't fully comprehended the range of features available to paying members - given this knowledge, choosing to do the roadmap in this form makes more sense to me.
Am I wrong for presuming that the people who would most care about GitHub's public roadmap probably aren't even GitHub users?
Free tier users, maybe not.
But paying users care very much about what they can expect for their outlay of money in the years to come. Roadmap presentations for paid software are _de rigueur_; they help keep your paying customers on your platform.. They're just not often shared openly like this.
Github has a vested interest in attracting more and more free-tier customers though. Then they go to companies and influence tool purchasing decisions and help feed the paying customer funnel. So this seems like a risk with some postive calculus behind it. It could be that free-tier users didn't know they wanted to know the roadmap, and now that they do they'll be happy they know about it.
> Am I wrong for presuming that the people who would most care about GitHub's public roadmap probably aren't even GitHub users?
I think so. I'm a GitHub user and I like being able to view their roadmap.
Think about it this way. GitHub is a giant piece of software that developers use. Software changes, and developers are particularly sensitive to this. We're constantly fixing bugs and factoring code to support dependency shifts (or choosing to snapshot ourselves at some preestablished point of stability).
Given our sensitivity to change and how important GitHub's product is to our productivity it's crucial that we're not hit with surprises. A properly communicated roadmap helps avoid that -- thought admittedly it does take some work on our part as customers to parse it. That said this is no different than reading release notes and documentation for the software and libraries we use elsewhere in our day to day.
problem is, this roadmap doesn't contain any specifics in contrast whats required from startups. It's just a fancy way of saying "we gonna do more source code versioning stuff".
Can you explain why? I really don’t see the draw of public roadmaps. I buy a product because it works for me now. If there are improvements I need in the future the company either implements them or they don’t. A public roadmap doesn’t improve my experience at all. In fact, I might be less likely to buy the product.
This feels like it opens github up to the vocal minority to scream and yell publicly about features they want. Some of which could negatively impact other users. And github will have to succumb to the pressure of that minority.
I fail to understand how an open roadmap would make you less likely to buy a product. If you compare two products with the same current feature set but one has an open roadmap and one has none - are you saying physical access to a roadmap would deter you from buying that product and instead buy a product that has no roadmap?
A public roadmap functions as a communications tool and sounding board. However, it does not negate strong product leadership or moderation of course.
Interestingly, my small SaaS was using GitHub projects as a public roadmap for the last 2 years already! Even at our small scale it has already proven useful for finding beta users, getting feedback etc.
> Discussions live in your project repository, so they’re accessible where your community is already working together. Their threaded format makes it easy to start, respond to, and organize unstructured conversations. Questions can be marked as answered, so over time a community’s knowledge base grows naturally.
I can imagine "GitHub Discussions" [0] cutting off a significant number of Stack Overflow questions.
This looks really cool. A sweet spot for certain kinds of questions that people abuse Issues for, but never felt right. Very important things like "what is the best practice for X" or "I have this problem but haven't yet isolated the cause, can anyone help?".
This is quite some timing for me, since the StackOverflow announcement earlier today had me thinking what could replace it. This is a clear, promising possibility.
I'm fine with this, so long as the search engine optimisation is there. I often google for an issue with a particular framework or library, and rather than the first result being the documentation, the first result is somebody asking on SO about it. It would be great if the answer is "closer to the source", pardon the pun. It also means that related questions may be easier to find, and you may come across something else of interest while on the discussion page for that particular project.
It also means that rather than building up SO's capital, you're indirectly contributing to the open source knowledge base of a project. It'd be even better if the discussion were backed by a Git repo or something easily exportable so that you aren't tied to Github, though I suppose part of the motivation for this feature is doing exactly that.
Still, SO has been king for too long. And the king's crown has been looking rather tarnished for a little while now.
I really like the idea behind this feature. Being able to add a discussions 'forum' to my main project could be an excellent way to start building a community around the project.
Looking at the Discussions issue page - https://github.com/github/roadmap/issues/104 - it looks like this will be delivered to projects not in the Beta testing group sometime before Christmas?
Out of interest, are there any project admins around who can comment on how easy it is to moderate the Discussions threads? My one concern would be managing flame wars, and minimising inaccurate answers participants may offer each other.
I'm looking forward to GitHub Discussions too. As well as helping foster community and contributers, it's also a clear separation from Issues, where I often see non-dev and peripheral questions being asked.
> Out of interest, are there any project admins around who can comment on how easy it is to moderate the Discussions threads?
There's a soft NDA on people who are in the Discussions beta. It's also early enough that there's a direct pipeline to GitHub Staff on issues that are truly urgent, so it's really too early to say. The Issues moderation things like locking and org-level blocking definitely still exist and work in Discussions, though.
I've seen a few startups (For example https://www.getoutline.com/) using the GitHub Discussions beta as a defacto support forum. I like that use case and it seems like it will help both developers and users.
For devs, you'll get a forum setup without requiring time/money to set up your own hosting, which is big for small or bootstrapped projects. It will also help move specific questions out of your issues section/ticketing system.
For users, it lowers the friction associated with getting support since many already have a GitHub account.
This is welcome, but anyone else feeling that GitHub is moving in the more-stuff-lower-quality direction? In particular I've wasted quite a bit of time over the past year on scarcely documented features and misfeatures around GitHub Actions. (To be clear I like GitHub Actions a lot.)
One example: just yesterday I found out that public images on GitHub Packages Docker registry can't be used in GitHub Actions' jobs.<job_id>.container, since the former is gated by auth for whatever goddamn reason and the latter can only pull from public registries. Think about it, their (probably #1) container-related feature can't use their own registry. Apparently people have been complaining for almost a year now, yet nothing has changed.
I do wish they were holding a slightly higher quality bar on each release, yeah.
My tiny example was the release of statuses. It included a pop-up to tell me it was there -- everyone knows developers find those obnoxious, just let me discover the feature myself. Plus, the feature just isn't relevant to me as someone who isn't actively participating in open source or looking for a job right now. The little animations were also slightly janky, and the button didn't feel like it was in a neat and orderly place - just sort of tacked on so that it'd be seen.
My hunch is that this comes from organizationally overemphasizing shipping something splashy instead of something amazing. Is the company celebrating "discerning people find this to be uncompromisingly good" (even for small or old features) as much as "this thing was noticed by a lot of people and then this other thing was noticed by a lot of people too"?
Overall I feel that under Nat, GitHub has walked this line quite well - just feels like it slips a little sometimes as a natural pendulum swing
I think more-stuff-lower-quality is kind of inevitable, but I am happy with that general direction. I have run into a lot of annoying things like and including your example. This roadmap would be better if it had a simple way to submit and vote for feature requests like other public roadmaps.
Thanks, @ianwalter. We plan on having a public feedback repo soon - but in the meantime, if you want to submit ideas, please check out https://support.github.com/contact/feedback.
My biggest beef with actions there's huge conflation between action and workflow. I've spend many hours swearing at their documentation...
Also I'm not sure it's a great a idea to use their actions actions as it creates dependencies on someone elses repos and Github itself. I feel I should just write my tests in a container, but I think I've hit a wall there pretty quick too.
I love that the issues each share a common format and provide concise yet clear documentation. The concise bit here is key -- I've had trouble perusing similar, open product roadmaps from other products because of the verbosity / level of detail.
Kudos to GitHub's team for doing such an excellent job in communicating and managing the roadmap. Particularly given the size or the organization -- this is no small feat.
It's 2020, the cost of an ipv4 address is more than $20 an address. Can there be some focus to include ipv6 support so that perhaps those people in the developing world who can't afford an ip address can access github and contribute?
Such a mechanism would be an actual meaningful step for enabling access to poorer communities.
> It's 2020, the cost of an ipv4 address is more than $20 an address.
But arbitrary layers of NAT are free, so in practice this shouldn't actually block anyone. I haven't had a globally reachable IP address for any of my computers ever, and yet here I am on the web.
It may not matter in the developed world where this cost doesn't matter because infrastructure is cheap. There are still limits to how much this can scale as a solution no?
For a Country like India, surely this cost will matter especially for it's poorest citizens. The limits of this infinite scaling NAT will matter. That and the cost of an IP which I imagine will inflate much higher than $20 an address.
It's 2020, why isn't this a priority on the roadmap. I think it would have a much more meaningful impact to those that need it.
Are we so sensitive to racism that the word "master" is offensive? There is a task to change the default branch to "main".
Note: there is no "slave" branch.
Intent matters, not the literal meaning of the word taken a specific orthogonal context. Not a single person in the millions of developers ever had a perverse notion of what master branch means.
> Not a single person in the millions of developers ever had a perverse notion of what master branch means.
Did you ask them all? I also never noted that it could be offensive. But "main" seems to be a good alternative and as long as master still works as a synonym, nothing should break changing it. The only contra to changing it would be the money Github spends on development. But that's their decision. Overall, I fail to see the problem.
I really don't need a lot of fancy stuff in order for GitHub to be valuable to me or my organization. Honestly, I did not find much of the roadmap to be compelling.
For me, the core features that comprise my GH user story are:
- Code
- Pull Requests
- Issues & Labels
We also have limited use of Actions (for check builds) and heavy use of the API/Webhooks for integration with our own custom CI/CD tooling.
In my opinion, the biggest place where value should be added is in these 3 areas above. Some of the simplest ideas are the most powerful. We get an incredible amount of mileage out of basics like issues & labels. If there were additional aspects to issues similar to labels that could further enhance this experience, I would be very interested. Our entire development process revolves around issues to track tasks.
One of the things I've had in mind would be a way to build a markdown-defined webform template that can be used for populating highly-structured issues without requiring the user to edit a complex document each time.
For example, you could have CustomerTroubleTicketTemplate.md in a .github/issue_templates/ path, and then when you go to click New Issue, a down arrow could be provided that pops a list of all defined issue templates. Selecting one would present the user with a webform (as defined in the template markdown) that collects all required fields to create that issue. The issue could be created with labels/assignments/etc as defined in the markdown document. This would likely require enhancements to GFM.
Bonus points if there is some way to expose specific issue templates through public GH pages so that your customers can submit tickets against your private repositories even though they don't have direct access to your issue buckets. I think a very small step in this direction could obviate a lot of use cases people find with products like Zendesk (it would for us, anyways - we just funnel ZD tickets back into GH issues).
It's a longtime practice of Azure DevOps (former TFS) team [1][2] and many from that team are now working under GitHub. Maybe that's a culture transfer from MS.
Readit News
A whole other level of evolution to do it while you’re large. Imagine Apple suddenly publishing their roadmap. This type of organizational neurogenesis done as an adult is impressive.
EDIT: I am in admiration of the change, regardless of opinion on the value of public roadmaps. It’s just rare to see big companies make big changes. Sign of youthful vitality and health, in my opinion.
Ex. https://docs.microsoft.com/en-us/power-platform-release-plan...
https://support.google.com/a/table/7539891?hl=en
The upcoming launches section is literally just an HTML table with two columns: service name on the left and a brief summary on the right like "Add support for X Y.Y". It's almost funny how so much effort goes into fancy blog design when the "real" news is just distributed through a basic plain text email :)
---
To a certain extent, roadmaps become redundant for enterprise customers too in that you start to influence feature request priorities via your dedicated account manager(s)
Vendors may often prioritise a feature requested by medium enterprise to keep them content since if they don't address their features in a timely manager, they may go elsewhere when their contract comes up for renew.
---
I suppose that's why UserVoice type of forums are always a wasteland because the big players can get all the attention?
It's also behind the scenes so perhaps naturally, the result you're left with is acres of "9999 votes for feature X, submitted 14 years ago" and a Product Owner saying "Sorry, no news to report on this just yet"
[1] https://cloud.google.com/blog/topics/inside-google-cloud/ann...
A public GCP roadmap is also in longer plans but there's a few internal flows we want to improve first. I'm especially trying to reduce the number of sites people need to visit, and want to ultimately see it land in a common place where authenticated users also see the non-public / NDA aspects.
And seconding this advice on access. If you are a GCP customer, your account manager is able to add your account for access. My contact info is in my profile and I can reach your account team if you have difficulties.
Am I wrong for presuming that the people who would most care about GitHub's public roadmap probably aren't even GitHub users?
EDIT: Thanks for the replies, everyone. Seems like I hadn't fully comprehended the range of features available to paying members - given this knowledge, choosing to do the roadmap in this form makes more sense to me.
Free tier users, maybe not.
But paying users care very much about what they can expect for their outlay of money in the years to come. Roadmap presentations for paid software are _de rigueur_; they help keep your paying customers on your platform.. They're just not often shared openly like this.
Github has a vested interest in attracting more and more free-tier customers though. Then they go to companies and influence tool purchasing decisions and help feed the paying customer funnel. So this seems like a risk with some postive calculus behind it. It could be that free-tier users didn't know they wanted to know the roadmap, and now that they do they'll be happy they know about it.
I think so. I'm a GitHub user and I like being able to view their roadmap.
Think about it this way. GitHub is a giant piece of software that developers use. Software changes, and developers are particularly sensitive to this. We're constantly fixing bugs and factoring code to support dependency shifts (or choosing to snapshot ourselves at some preestablished point of stability).
Given our sensitivity to change and how important GitHub's product is to our productivity it's crucial that we're not hit with surprises. A properly communicated roadmap helps avoid that -- thought admittedly it does take some work on our part as customers to parse it. That said this is no different than reading release notes and documentation for the software and libraries we use elsewhere in our day to day.
This feels like it opens github up to the vocal minority to scream and yell publicly about features they want. Some of which could negatively impact other users. And github will have to succumb to the pressure of that minority.
Interestingly, my small SaaS was using GitHub projects as a public roadmap for the last 2 years already! Even at our small scale it has already proven useful for finding beta users, getting feedback etc.
https://github.com/checkly/public-roadmap/projects/1
I can imagine "GitHub Discussions" [0] cutting off a significant number of Stack Overflow questions.
[0] https://github.com/github/roadmap/issues/104
[1] https://github.com/vercel/next.js/discussions
This is quite some timing for me, since the StackOverflow announcement earlier today had me thinking what could replace it. This is a clear, promising possibility.
It also means that rather than building up SO's capital, you're indirectly contributing to the open source knowledge base of a project. It'd be even better if the discussion were backed by a Git repo or something easily exportable so that you aren't tied to Github, though I suppose part of the motivation for this feature is doing exactly that.
Still, SO has been king for too long. And the king's crown has been looking rather tarnished for a little while now.
Looking at the Discussions issue page - https://github.com/github/roadmap/issues/104 - it looks like this will be delivered to projects not in the Beta testing group sometime before Christmas?
Out of interest, are there any project admins around who can comment on how easy it is to moderate the Discussions threads? My one concern would be managing flame wars, and minimising inaccurate answers participants may offer each other.
There's a soft NDA on people who are in the Discussions beta. It's also early enough that there's a direct pipeline to GitHub Staff on issues that are truly urgent, so it's really too early to say. The Issues moderation things like locking and org-level blocking definitely still exist and work in Discussions, though.
For devs, you'll get a forum setup without requiring time/money to set up your own hosting, which is big for small or bootstrapped projects. It will also help move specific questions out of your issues section/ticketing system.
For users, it lowers the friction associated with getting support since many already have a GitHub account.
One example: just yesterday I found out that public images on GitHub Packages Docker registry can't be used in GitHub Actions' jobs.<job_id>.container, since the former is gated by auth for whatever goddamn reason and the latter can only pull from public registries. Think about it, their (probably #1) container-related feature can't use their own registry. Apparently people have been complaining for almost a year now, yet nothing has changed.
My tiny example was the release of statuses. It included a pop-up to tell me it was there -- everyone knows developers find those obnoxious, just let me discover the feature myself. Plus, the feature just isn't relevant to me as someone who isn't actively participating in open source or looking for a job right now. The little animations were also slightly janky, and the button didn't feel like it was in a neat and orderly place - just sort of tacked on so that it'd be seen.
My hunch is that this comes from organizationally overemphasizing shipping something splashy instead of something amazing. Is the company celebrating "discerning people find this to be uncompromisingly good" (even for small or old features) as much as "this thing was noticed by a lot of people and then this other thing was noticed by a lot of people too"?
Overall I feel that under Nat, GitHub has walked this line quite well - just feels like it slips a little sometimes as a natural pendulum swing
Thanks, @ianwalter. We plan on having a public feedback repo soon - but in the meantime, if you want to submit ideas, please check out https://support.github.com/contact/feedback.
Also I'm not sure it's a great a idea to use their actions actions as it creates dependencies on someone elses repos and Github itself. I feel I should just write my tests in a container, but I think I've hit a wall there pretty quick too.
I guess the good news is that the majority of the items on the roadmap are github actions related?
I love that the issues each share a common format and provide concise yet clear documentation. The concise bit here is key -- I've had trouble perusing similar, open product roadmaps from other products because of the verbosity / level of detail.
Kudos to GitHub's team for doing such an excellent job in communicating and managing the roadmap. Particularly given the size or the organization -- this is no small feat.
Dead Comment
Such a mechanism would be an actual meaningful step for enabling access to poorer communities.
But arbitrary layers of NAT are free, so in practice this shouldn't actually block anyone. I haven't had a globally reachable IP address for any of my computers ever, and yet here I am on the web.
For a Country like India, surely this cost will matter especially for it's poorest citizens. The limits of this infinite scaling NAT will matter. That and the cost of an IP which I imagine will inflate much higher than $20 an address.
It's 2020, why isn't this a priority on the roadmap. I think it would have a much more meaningful impact to those that need it.
Note: there is no "slave" branch.
Intent matters, not the literal meaning of the word taken a specific orthogonal context. Not a single person in the millions of developers ever had a perverse notion of what master branch means.
https://www.theregister.com/2020/06/08/developers_renew_push...
https://sfconservancy.org/news/2020/jun/23/gitbranchname/
https://github.blog/2020-07-27-highlights-from-git-2-28/#int...
and some previous discussion: https://news.ycombinator.com/item?id=23968417
The word is not necessary, so why does anyone have to be offended if it's changed?
Did you ask them all? I also never noted that it could be offensive. But "main" seems to be a good alternative and as long as master still works as a synonym, nothing should break changing it. The only contra to changing it would be the money Github spends on development. But that's their decision. Overall, I fail to see the problem.
For me, the core features that comprise my GH user story are:
- Code
- Pull Requests
- Issues & Labels
We also have limited use of Actions (for check builds) and heavy use of the API/Webhooks for integration with our own custom CI/CD tooling.
In my opinion, the biggest place where value should be added is in these 3 areas above. Some of the simplest ideas are the most powerful. We get an incredible amount of mileage out of basics like issues & labels. If there were additional aspects to issues similar to labels that could further enhance this experience, I would be very interested. Our entire development process revolves around issues to track tasks.
One of the things I've had in mind would be a way to build a markdown-defined webform template that can be used for populating highly-structured issues without requiring the user to edit a complex document each time.
For example, you could have CustomerTroubleTicketTemplate.md in a .github/issue_templates/ path, and then when you go to click New Issue, a down arrow could be provided that pops a list of all defined issue templates. Selecting one would present the user with a webform (as defined in the template markdown) that collects all required fields to create that issue. The issue could be created with labels/assignments/etc as defined in the markdown document. This would likely require enhancements to GFM.
Bonus points if there is some way to expose specific issue templates through public GH pages so that your customers can submit tickets against your private repositories even though they don't have direct access to your issue buckets. I think a very small step in this direction could obviate a lot of use cases people find with products like Zendesk (it would for us, anyways - we just funnel ZD tickets back into GH issues).
[1] https://docs.microsoft.com/en-us/azure/devops/release-notes/...
[2] https://dev.azure.com/mseng/AzureDevOpsRoadmap/_workitems/re...