fedora culture clashes?

October 29, 2008

I’ve been seeing a bunch of threads over the last few dev cycles that are leading me to believe we are in the middle of a culture clash in fedora – maybe in linux distros in general. The cultures are roughly divided along these lines (in no particular order)

A. Sysadmins, old-school linux-y, retrogrouch folks.

B. Laptop-oriented developers.

Group A sees computers as a means to an end. Change is something they have to act as a proxy for to their users. Their users are a large and diverse group. They control many of the views of the computing world and access for their users. They meter access and privileges. When they think of users the idea of one of the users also being an admin is anathema. Change and disruption of the status quo for their users is generally considered a bad thing. This group is more likely to think of a computer in the same way most people think of a desk, a chair, a pen and paper. It’s something that is expected to be there, expected to work in a specific way and significant changes in how these items work are generally met with irritation. Recurring change is met with outright anger. At some point in time this group got comfortable with how bits were working and they are resistant to changes that don’t make things immediately easier either for them or for their users w/o a loss of functionality for someone else.

Group B sees computers as an exciting interface to the world. A place where we are not hampered by older assumptions or paths. A zero-physical-cost to replacing and revamping designs. They live in a rapid-prototyping world. The idea of throwing-one-away is something they know quite well. They believe they are working on new ideas and in some cases they are.  They look at the other operating systems and interfaces than the ones on linux with an interest in bringing some of that ease of use and smoothness to linux and X in general. The release of MacOSX was an exciting driver for many folks in this group. Their concept of a user is someone buying a laptop or a netbook-style or other smaller device and wanting to use free software on it. They think that mobile and ubiquitous computing is happening in laptops, netbooks, MID devices and phones. They are able to change rapidly with design revamps b/c if they are not the ones involved with the revamp they are often close to those who are.

Let me give you a fictional example of not understanding each other:

Member of B: We’re going to be moving around where the tabs work on the default interface. They’ll all be on the right side, rotated vertically, like reading down a bookshelf. We’re sure it will help make things easier to the user. If you want them back the other way, just move them back by dragging them over or setting your preference for tab location

Member of A: You’re just going to stick that on us? That’s nuts and reckless!

Member of B: It’s progress, we need to do this to enable a vertically-functional desktop! It’s not a big change!

Member of A: I Can’t believe you’re ignoring your users like this!

What Member of A doesn’t understand is that Member of B has been looking at how various interfaces has evolved and noticed that a lot of vertical placement is comfortable (or some other explanation) for users and allows them to ‘vertically file things more efficiently in their mind’ (I’m making this up, it’s just an example)

What member of B doesn’t understand is that Member of A is going to have to deal with hundreds of end users. Some of them will find the change interesting and maybe even more efficient. Most of them  are going to react with “CHANGE BAD, MAKE CHANGE STOP, GO PUMMEL LOCAL SYSADMIN UNTIL CHANGE STOPS”.

So member of A is a product of their (abusive) environment. That environment is no less valid, it is just very different from member of B.

Member of B is also a product of their environment. They work and communicate with other forward-looking developers who are welcoming and comfortable with change like this.

The problem for fedora is not about which of these groups is better or right in any sense of either of those words. The problem is there is only so far we can go in fedora trying to target BOTH groups at the same time for our default release. At some point the infrastructure changes you need to drive our interface changes make user and admin noticeable impact on core functionality in the distribution. So, fedora needs to definitively and authoritatively say what the target is for the system development that is being done. This has to keep  a couple things in mind:

1. it’s current user base

2. distros downstream of it (OLPC, rhel, centos, etc)

There may be a medium somewhere between group A and group B. That is a hard place to find. However, I think it is best for us to understand the difference between these two groups if we’re going to be able to speak and comprehend each other w/o it devolving into flame wars.

Please remember, try to look at the world from the perspective of the people you’re disagreeing with. Their perspective isn’t yours and it is no less valid.

14 Responses to “fedora culture clashes?”

  1. Nicolas Mailhot Says:

    I think you forget an important point. Left alone group B will never finish anything. His members will always start tearing down things when the 80% easy and cool bits are done and it’s time to do the 20% hard and boring and un-exiting part. (in kernel land, that gave use perpetually unfinished and not-merged gfx drivers). And you’ll end up with something free-wheeling out of control like Vista was.

    While often overly cautious, group A is actually directly in contact with users (that’s called help desk, support, admin) and relays their aversion to gratuitous change. It should not be dismissed lightly.

    When group B tells you it’s building the car of the future, what it really intends is to build a concept car that will be paraded in the next auto show and never see mass production. For mass production to happen group A needs to be listened to.

  2. Máirín Says:

    I have to agree with Nicolas here, the key difference here is the level of contact with the end users. I do think though that administrators’ user data will obviously be colored by their position, but even ‘colored’ user data is better than NO user data.

    By playing in a sandbox unencumbered by user research and user data, group B is treasure-hunting, group B is not making safe bets, and furthermore group B is treasure-hunting with no way to measure their progress (or regression) if they are not using user research to confirm their hypotheses before unleashing them on group A and that group A’s users.


  3. I have a practical example. I wanted to use ClarkConnect which is based on RHEL4. I wanted to use the print server, and I have an Epson Stylus R320.

    Short story, the printer stuff in RHEL 4 is too old, for my printer, and the only real way to get it working, is to move to something based on something newer.

    I’ve considered Amahi, but that doesn’t have lots of the established stuff I like about ClarkConnect. I could make it happen in Fedora 9/Amahi, but I already know that CC is stable and I don’t have to mess around with it too much.

    However, on my desktop, I’m much more inclined to use recent versions of Fedora. I don’t mind shaking up my desktop, but that’s really only because my server is running pretty stable. I can’t really depend on my server to run something like Fedora, that’s constantly changing…

    Or, can I? Maybe we just need a way of separating data from the interface better. I’m way more tolerant of an interface or desktop change, if I know my data is safe. Maybe if Fedora had a better way of interacting with (backup, GUI remote management, applications) a dedicated “home server” type of box, users would be less upset about changes.

  4. Stephen Smoogen Says:

    Twin!.. but I didn’t go as far on the analysis of who the groups are. I am definately in the first group. My views are conservative because when stuff breaks I have 100’s or 1000’s of desktops to deal with. I need to take a second back and realize that desktop developers usually are dealing with a 1:1 relationship.


  5. I think another way to summarize is:

    A: everything new is bad

    B: everything new is good

    I guess as most people I consider myself somewhere between these two groups. But I recently made the mistake of subscribing to fedora-devel and it seems that Fedora is just going for the B type. Maybe so far Fedora always got the balance right and it is just now that a shift is happening and this is the cause of all the hurt on the lists.

    What B also doesn’t realize is that it easy to do new things and throw old stuff out, it is very difficult to go back to the old version, even if the new way doesn’t prove to be good. Instead the new bit is hammered into place until it fits or the next new hype turns up.

    To me it looks like Fedora 10 might be my last Fedora version after too many annoyances, but I haven’t decided where to turn to. Maybe just CentOS, as I am a bit of a RPM junkie.

  6. Bill Rugolsky Says:

    IMHO, the single biggest fear of retrogrouchs like myself is that folks in group B, in an effort to promote ease-of-use ala MacOS X or Windows, will lose sight of the multi-user (or, for single-user laptops, multi-role) heritage of UNIX. So many of the security and configuration mgmt issues with Windows arise from its historical non-networked, single-user orientation, and the assumptions were baked in at such a low level that they are incredibly difficult to fix.

    I participated in an unpleasant thread about NetworkManager, and I’m happy to see that the rewrite addressed some of my complaints.

    Nevertheless, this grouchy old UNIX hand has been impressed with the direction taken by PolicyKit; it demonstrates that one can satisfy a large number of traditional functional requirements with a completely re-architected implementation that also serves modern requirements.

  7. Kai Willadsen Says:

    While it’s nice to have a more level-headed let’s-all-just-get-along perspective on recent flamewars, there’s a problem. The main problem with your hypothesis is that group A is usually found (at least on fedora-devel-list) complaining about problems that have nothing to do with what their users see.

    These people don’t complain when the default Gnome theme changes, or when KDE gets upgraded. They don’t complain when some unintuitive piece of interface gets exposed and they start getting support calls. They don’t complain when a Firefox update breaks user-installed plugins.

    They complain when you threaten to make them install sendmail themselves.

    They are currently complaining about a purely cosmetic change that will (very slightly) affect them, but will probably not affect a single one of their users.

    I’m not saying that some grouchiness isn’t reasonable, or that there isn’t some reason for their apparent abhorrence of change. However, the public behaviour of group A doesn’t support the “it’s all about the users” theory.

  8. jjmcd Says:

    Hehe Seth, you can sure tell which side of the fence you’re on!

    There is a lot of value to both points of view, and perhaps in Fedora we get exposed to it more than most because Fedora is right on the bleeding edge, and intended to be. While we need to be moving quickly, as the upstream for RHEL, we also need to be watching supportability and maintainability. And I think we do a pretty decent job. (OK, maybe I shouldn’t say “we”, as a relative newcomer to fedoraproject my contributions have been trivial, but still I feel like I need to defend!)

    A few months back I asked a question at one of the IRC meetings along a similar vein. Not exactly the same, but really addressing the same groups. As a past big corp I/T guy, things like a new editor written in every language under the sun just because it is cool to have an editor written in Haskell, or python, or ocaml, or whatever, I mean what is up with that? Why on earth do we need wallpaper that changes with time of day? Each theme the art team needs to do four times and make sure they all fade nicely into each other. Where is the value in that?

    Paul’s response, I thought, was actually very good. We move quickly, we have the latest and greatest, and we do things that are sometimes more about fun or coolness than utility because that attracts developers. Yes, we sometimes need to work at striking a balance, and perhaps because we are on the leading edge that gets a little tougher. We are the upstream for RHEL and some other distros that tend to be more “serious”, so we need to exercise some caution, but that is part of the rent for our niche in the world.

    And there should be debates. Maybe even heated debates. Thats how you get to good decisions. Better to have the debate then to go off doing something stupid, or doing nothing at all. Sure, sometimes the debate isn’t as comfortable as we all would like, but each of us has something in there we are passionate about. How can you respect someone who doesn’t care about anything? But at the end of the day, we simply need to respect each other, and the process, and move on.

  9. henshaw Says:

    Disclaimer: I don’t have a horse in this race, so any subtext you read into this is probably unintentional.

    “What Member of A doesn’t understand is that Member of B has been looking at how various interfaces has evolved and noticed that a lot of vertical placement is comfortable (or some other explanation) for users and allows them to ‘vertically file things more efficiently in their mind’ (I’m making this up, it’s just an example)”

    Perhaps group B should be more forward in presenting the research to justify their changes (ideally, wikis and hyperlinks are good for this). And accept that they will need to present their case (at least in link form) over and over again, when questioned by group A.

    I wouldn’t be surprised to find that group B often have good answers to some of group A’s concerns. Communicating this effectively, to reduce group A’s complaints to the essential sticking points, is more difficult.

    Group B won’t be able to stop all the complaints, whatever they do. But they can forestall most of them.

  10. skvidal Says:

    jjmcd: I don’t think I’m really on one side or the other. I was a sysadmin over a lot of systems and a lot of users for a decade, so I definitely understand group A pretty well. But I’ve been on the fedora board and working on fedora for the last 5 years and I deeply understand the need to unload old bits and move ‘forward’.

    I didn’t write this post to incite anger, I wrote it to explain that the situation is not very cut and dried to either perspective. No matter what the specific flashpoint is.

    It’s not about X and tty7, it’s about change and not-change.

  11. Harry Says:

    I just want my usb digital camera to work when I plug it in! ( it hasn’t since fedora 7 and yes i’ve filed a bug )

    Plus, could you please make the fonts on your site here just a little bit smaller, I’m studying subatomic components and structure inside the atom and I would like to use your font size as comparative reference.

    Thanks,
    Harry

  12. Stephen Smoogen Says:

    I am not sure its just change/no-change. Its also about environments people work in, and how change is managed. Every other release there is are a lot of questions about “why isn’t Fedora used more in the Enterprise?” and while the answer is usually too short of a support cycle.. its deeper in that the rate of change for Fedora is too high for even many small businesses to keep up with. By the time you have gotten used to the changes that were done in F8, F9 has a whole new set you have to document, test, get into training, which you get done when F10 comes out.

    While most might think that oh its just a small change here, a small change there.. it adds up in a non-linear fashion. I feel that “Fedora has become an OS for people who live Fedora 24×7” which means that there isn’t a place for people like me who can only ‘live it every now and then’. Which basically says its time to get out of the way.

  13. Kevin Kofler Says:

    @Christof Damian: This is not a new phenomenon at all, there have always been flamewars about changes in Fedora. It may just have been the unlucky timing of the changes which is keeping the noise higher lately: Fedora 8 was fairly quiet, all the big changes (new X.Org X11, KDE 4 etc.) were in Fedora 9, so we got all those flamewars at once. Now Fedora 10 is making a set of minor changes which have in common to change decades-old Unix behavior (sbin in PATH, X on tty1, and other such stuff like dropping or replacing sendmail and replacing SUID with capabilities is also being discussed, though AFAIK neither change was made for F10) and that is angering a certain group of people (probably composed mainly of long-time users), and the flamewars are probably made worse by the coincident timing of the changes (and meanwhile, flamewars about KDE 4 and other F9 changes are still not over).

  14. mvpittman Says:

    Can Fedora do anything to reduce the cost of the high rate of change? Maybe that sounds derivative, but one example I’ve found useful is pre-upgrade and the USB stick scripts. Easy to upgrade, and easy to try out, which is great, as long as you don’t need guaranteed availability of your data.

    I’m thinking disconnecting the data from the changes, an improving that whole “backup my data” experience and making it easier to “roll back” would make happier testers and might even make the bleeding edge a more acceptable place to be.

    If you’ve got a spare box, you can install Fedora on it, but you use your main box, not your spare one. If you make your spare box something that contains backups, and will do a bare metal restore of your main box, you can play on your main box (that has all the cool stuff that is meaningful on that bleeding edge that we care about) But, when it goes sour, you’re not constantly worried that you won’t be able be productive while you “fix” it.

    You can substitute “spare box” with “spare hard drive” or “flash drive” or “live CD” or whatever, but it is simply too expensive to live on the edge. I’m suggesting that making that experience less perilous (not necessarily less-bleeding edge) could alleviate tension, and get more “testers”. Which are *required* for more and better feedback.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: