What should the "PRD" process for editions look like going forward?

History

When we started Fedora.next, we had the idea of making “Fedora Products”. This presentation from Christoph Wickert at Flock Prague covers it really nicely – you can watch the whole thing for history, or jump to about minute 20 for the part specifically relevant here:

Or if you prefer reading to watching, these Fedora Magazine articles from the time are a good source:

What we asked for originally:

The accepted “Fedora Products” proposal to the Fedora Board is on the wiki. It was approved sometime in 2013. (As a project, we’ve struggled with recording and communicating decisions for a long time. “Slow progress towards the somewhat better!”)

That asked for two deliverables: a PRD which includes a statement of target audience and the role in the Fedora ecosystem, and a list of changes to Fedora procedures needed to make that thing happen. It’s also worth noting John Rose’s — that’s @inode0 — concern that “these three products are not set in stone for the next decade”, and the note from @misc that “we could had a yearly check of the relevance of the project or something like this”.

I can’t find if we actually defined in writing what we meant by “PRD”, exactly. We certainly talked about it a lot, and the resulting Cloud, Workstation, and Server PRDs — although different in structure — I think fit the intended idea well.

A couple of years later, we more formally set the ambitious goal of having annual updates. In retrospect, that didn’t work out.

So, now…

First of all, of course, we dropped the word “product” (for $REASONS). This makes “PRD” kind of a strange term, and perhaps even more esoteric to most people not familiar — and we have plenty of acronyms in tech already! Plus, in the world of product marketing, there’s some arguments about “MRD” vs “PRD” and so on. And, in an actual product organization, a PRD or MRD is a communication from Product Marketing, who does customer and market research to determine the problem to solve, to Engineering, who will then create a functional spec and implement. That clearly doesn’t cleanly map to what we’re doing in Fedora.

So, I’m up for calling this something else going forward. (Feel free to bikeshed on that below. I suggest, with no strong attachment, “Edition Target & Goals”).

Some meta-questions I think we should answer:

  • Why are we asking for this? What’s the point?
  • Who is the audience of the document?

And then, practical questions:

  • What must the document contain?
  • What shouldn’t it cover?
  • When, or, really: How can we structure the process so updating the document regularly is:
    • not burdensome,
    • practically useful to Fedora teams (both the edition WGs and others), and
    • something that actually happens?

Are there other question we should think about, both meta and pragmatic?

I am thinking that our eventual output here is in the form of a document with sections titled “Why”, “Who”, “What”, “When”, and “Where”. (That last one not just to complete the 5Ws, but also to make things easier seven years from now when Future Fedora is trying to look back.) You can reply to this however you like, but I think it might be useful to me if you use these headers in your responses.

I’m not sure, given the conversation that sparked this thread, that we want to keep the Editions as a concept. Certainly we don’t seem to be expressing what we intend to the community very well. So before we start changing the PRD process, maybe we should figure out what we want from Editions (or whatever concept replaces that) and then see how the current PRD process does or does not meet the needs.

This sounds like a great thing for that strategy meeting we’re planning to plan.

I think it’s good to have some wider conversations going into the planned planning plan. :classic_smiley:

I guess this all falls under the Why.

To me at least, the goals of the editions aspect of Fedora.next are still good. There’s a bunch of that in the links above, but basically as I’m thinking of it now:

One part of our mission is to empower our community to build tailored solutions for users. Emphasis on tailored — specifically fashioned and fitted. I was watching one of the Linus Tech Tips episodes where they’re trying Linux distros as “daily drivers” (thanks, @ngompa). From that episide:

Customization gets billed as a major selling point for Linux, and fair enough if that’s your thing, but speaking on behalf of “normies”, I don’t want a dozen novel ways to do the same thing. I want one, fast easy one.

Of course, Fedora Linux as a whole provides almost endless opportunities for customization — and like the quote says, that’s great for people who want that (and in fact, that’s really another part of our mission — we build a platform for that!). But, in order to reach a broader, bigger audience, we need products ¹ that are ready to go out of the box.

The original Products — Editions — idea was basically: there are three broad areas we’re already working on in Fedora which could reach that bigger audience if

  • we put in the basic product-marketing work to express the user needs in each area,
  • determine solutions to those needs, and
  • make a technical plan to approach each solution; and
  • empower dedicated teams to make independent technical decisions where necessary,
  • while still keeping to the fundamental Fedora foundations and addressing the broader mission.

This hasn’t been a perfect success, but it’s been a significant one. Honestly, I don’t think we’d be where we are today, talking about a surge in positive energy, if we hadn’t taken this approach.

Now, when we launched it, we knew that it meant some trade-offs. One that definitely frustrated a lot of people (hello, KDE SIG) was that making opinionated decisions (like, “we’re best equipped to deliver a polished Workstation experience on GNOME”) means people working on providing solutions in other ways, or working on other areas don’t get the same resources: top billing on the website and release announcements, quality assurance time, prioritization of Red Hat-invested work asked for by the Council², and more. I think part of our plan for the next five years — in which I modestly intend to call for doubling the number of active Fedora contributors — needs to address the areas we left behind.³

I don’t think we were wrong to make that call then. We were in actually pretty bad shape in terms of morale, resources, Red Hat confidence as a sponsor, and more. And just plain “hey QA and Rel-Eng, want to add all of this on top of everything else” wasn’t really an option. But now that we’ve gotten to where we are, I think we can do more.

Going forward, do we still need the big, broad, flagship areas at which we targeted Editions⁴ covered in the same way? Not necessarily (and certainly not necessarily the same target areas), but, going back to the quote I started with, I think it’s still a good way to grow the distro and the project with it. So, I think we should have something like it, at least.

Or, let me put it in negative terms. I think we should have the above, but I’m open to other ideas. I don’t think we can succeed with a “we give you options, you build it” model for users.⁵ I don’t think we should solely do "we build what we want and you can take it or leave it"⁶ — we’re only going to grow by caring about user needs, for users that aren’t just a mirror of ourselves. And I don’t think we should present users with an undifferentiated list of dozens offerings with no guidance on where to start, or on what’s best supported, what’s best tested, and so on.

So, we need some way to identify those user needs, and a way to empower our teams to deliver matching solutions. If that’s not Editions, what is it?


Notes⁷:

  1. The RH folks who told me this said they don’t mind if we use the word in our planning and development, they would just appreciate it not getting into our marketing and user conversations. Since Fedora doesn’t have a stark line between the two, it’s often easier to just always use “Editions”. But it’s not an edict to always contort our language or anything. So, I’m sayin’ products here.
  2. By the way, @bcotton, @amoloney, and I are working on a proposal for better feeding work requests and requirements to the CPE team. Stay tuned for more on this
  3. This footnote grew so long I’m making it into a separate thread, "Labs" and "Spins" — in the shadow of Editions?
  4. I know the other thread is long, but for context on the three areas see @sgallagh’s reply, comment #18.
  5. For users, no. But for Fedora project contributors and developers, yes. That is in fact our strategy of record.
  6. Again, in general. For some parts of the project, go wild with your personal vision and it’s okay if no one else cares!
  7. I really wish we had the Footnotes plugin for Discourse. If we decide to self-host or convince our sponsor to fund the enterprise hosting plan, that’s first on my list.

I agree (although it could be that I just haven’t yet heard an alternate proposal that sounds even better). But if we decide to change how we implement those goals, then that will potentially require changing what we ask of PRDs. So given that we’re about to start the former conversation, I think it’s a better allocation of our brain cycles to put a discussion of PRDs off until after that.

1 Like

I guess as I see it: we’re unlikely to build a new plan entirely from the top down (and if we were to, that’d be unlikely to be successful). And, as you note, we don’t really have an alternate suggestion, so this is probably going to be one of the components. Therefore, I think it’s useful to at least think about the “Why” at this level, because doing so will make it easier to fit into the bigger picture plan. I’m definitely open to alternate ideas that follow from the Why, or arguments for a plan with different Whys altogether.

1 Like

From my perspective as someone who works on the Workstation Working Group, we still really struggle with visibility and presence in the Fedora project. I find it striking how invisible we often are, whether it’s on the wiki, in the docs pages, or in this discussion space. Sometimes it feels like there’s an odd inverse logic at work - because we’re the “main” offering, we end up with less visibility.

Some of this is obviously down to the workstation working group itself, and I am trying to improve the situation as best I can, though I’m starting to realise that, given that the Workstation Working Group is primarily an engineering team, there are limits to what can be done.

The other thing that I’m very conscious of is identity (ie. brand) - Fedora’s identity, the editions’ identities. Having a positive identify around workstation is, I think, is key to growing the user base (my ultimate goal). Yet, we’ve never really developed a strong independent identity for workstation itself, nor has Fedora fully got on board with workstation being the main pillar of its own project identity. Which leaves me wondering - where is the identity we need going to come from?

Which is my way of saying - when we think about the role of each of the editions, we need to think about their role and presence in the community (not just organically - but by design) and how they function as brands.

@aday Can you provide an example what sort of visibility you mean?

To me it looks that the desire to be a default Fedora experience contradicts with the idea of a separate identity. Default Fedora identity is not the Workstation group alone, it is the identity of the Fedora Project. So what kind of visibility for a Workstation WG you are looking for apart form the fact that the deliverable the group creates is literally the one thing the user get on the front page of the getfedora.org?

Are you talking about the visibility on the contributor level? Or for the user?

Can you provide an example what sort of visibility you mean?

Sure! So one example is discussion.fedoraproject.org itself. It has a “desktop” category (for “discussion of Fedora desktop environments”), and there are categories for the desktop spins - Silverblue, Kinoite, plus the non-desktop spins and editions, like CoreOS, Server, IoT. The one thing that’s missing - Workstation.

Another example is project navigation. Go to docs.fedoraproject.org. In the user docs you will see Server, Silverblue, CoreOS, IoT. No Workstation.

Or go to the page on Fedora project organisation. Not a mention of Workstation, nor any link that will lead you to the workstation working group.

The wiki suffers from similar issues, too.

Sometimes it feels like the Fedora project actively tries to avoid talking about its main desktop offering. Are we embarrassed about it?!

1 Like

The issue here, i think, is that while workstation does use Fedora’s identity, workstation isn’t really present in that identity. Who can say how Workstation is key to delivering Fedora’s vision of the future? Or why Workstation is the primary vehicle for Fedora (as opposed to any other spin or edition)?

The other issue here is that, as more editions are added, and as more spins get recognition, it becomes more difficult to answer these questions. Fedora becomes more like an umbrella organisation for independent projects pursuing their own visions.

Ok, those examples make sense, though the conclusion you make actually surprises me.

I would expect that Gnome folks are quite familiar with the effect that if you make thing a default it becomes invisible. In the Fedora Survey we’ve done recently half of the respondents couldn’t name their image viewer program, and had to reply “whatever the default is” as they have never needed to call the program by name. So I don’t see why you have such a negative interpretation to this fact. No, there is nothing embarrassing, and nothing hidden, it is just a natural thing: people don’t separate Workstation from Fedora, because Workstation is the default Fedora.

Of course it doesn’t mean we can not work on that. And yes Workstation WG deserves a bit more publicity than it has now.

But I’d also say that rather than Fedora Council trying to make the WG visible, it is also up to the WG to become a bit more open. For example this discussion board: the categories here were created on request. And since this board competes with the mailing lists, we had to be careful and not create “dead” groups if the majority of the discussions on the topic happens elsewhere over the mailing list. Thus we have many groups missing from this board. The WG needs to explicitly decide to use the forum for discussions. Fedora Council did exactly that, for example.

Same for docs: the docs front page was not created by the council, it represents groups which submitted their doc sites into the common index page. So if you have a dedicated docs site for the workstation: you can submit it to the index page of the docs.fedoraproject.org too. There is no one else there to do it :slight_smile:

1 Like

Who can say how Workstation is key to delivering Fedora’s vision of the future? Or why Workstation is the primary vehicle for Fedora (as opposed to any other spin or edition)?

I like how Máirín puts it in A new conceptual model for Fedora as desktop Fedora experience is an entrypoint for a user into the “world of Fedora”. Or the first level in the pyramid.

It allows to not limit the project to desktop use, but it highlights the importance of the desktop experience of a user for Fedora.

Fedora becomes more like an umbrella organisation for independent projects pursuing their own visions.

This is a valid point. I think we don’t want to become something like Apache Foundation. There should be some unifying force which ensures that we actually work together rather than just next to each other. And it should be something different from “Red Hat has commercial interest in X”.

Some time ago I was trying to make non-modularized shared pool of integrated rpm packages to be that common thing, but it appears to be a more controversial topic than I originally thought.

2 Likes

Yeah, for this specific thing, that’s exactly it. From my point of view I’m just waiting for Issue #223: Workstation user and contributor community on discussion.fedoraproject.org - fedora-workstation - Pagure.io (and someone to ask me to do it!)

That’s fair, and it’s something that I’m personally working on. But we are constrained here, and the WG is primarily an engineering team and doesn’t naturally take on tasks like this.

My point being - whoever’s responsibility it is, the task of building out the editions we have is still very much an incomplete project.