Re: [Userops] What is "userops", and who does it target?

[Dave Crossland]

On 12 Apr 2015 2:26 pm, "Christopher Allan Webber" <address@hidden> wrote:
>
> Blaise Alleyne writes:
>
> > But maybe that's not the point... maybe the target audience for userops is
> > *still* a bit more of a technically savvy audience? But the goal is to lower the
> > bar from "developer/sysadmin" to "somewhat technically savvy computer user"?
> >
> > This changes my thinking (misconceptions?) about userops... from the
> > impossibility of making something that the average user could actually use,
> > towards maybe just significantly lowering the bar.
> >
> > For me, recognizing that there are still many users for which a total dream
> > userops setup would *still* be too complex and inaccessible, it highlights for
> > me that userops efforts also still really need to be about "self-hosting" for
> > *other* people too...
>
> This is a good point.  First of all, it's worth noting that "userops" is
> a vague term, totally undefined at this point, more of a rallying cry
> for people working on similar issues, but it's probably worth defining.
>
> Here's attempt one at a definition:
>
>   Userops is about deployment solutions that target users rather than
>   developers or businesses.
>
> But wait, that's not right.  Many times, developers *are* users, if we
> think in the "user freedom" style of users.  Sometimes, users are
> hackers.  So, try two:
>
>   Userops is about deployment solutions targeting user freedom, rather
>   than developer freedom.
>
> This one I think has more bite: it's still a bit loose, but I think it
> gives us a better picture of things.  One might say that the whole state
> of the current "free software dark ages" has probably come from a shift
> to making sure that developers have a maxmimum of ability to write
> software and employment opportunities, but not really worrying about how
> that affects the end-user.  But if we're worried about end-user
> freedoms, that changes our focus.
>
> But okay, hackers can be users, but are we also only interested in
> freeing hackers, what are we achieving?  That sounds like freedom for
> the technically privileged.  But true, maybe hackers can help run some
> services for their family and friends, which currently already is pretty
> hard, so that might at least make some movement towards success.
>
> Thus, I suggest there are several "milestones" of categories that
> userops systems could succesfully target.  (These are loose, abstract
> categories.  Of course it's not true that people just neatly fit into
> these roles.)
>
> 0) Big companies and developers with a high level of technical privilege
>    and enormous amounts of time, one way or another, to come to learn
>    and understand whole systems and keep them deployed.  I'd say this is
>    the current "devops" focus.  Hopefully "userops" systems aim for
>    better!

I think this is like git vs fossil. Git is designed for big hierarchical professional teams like kernel developers, and fossil takes the concept of dvcs and designs one for smaller, flatter teams.

> 1) The people on this list deploying their own servers for themselves,
>    family, friends.  I assume this list is also full of the highly
>    technically privileged, but maybe we want to be able to reuse each
>    others' deployment strategies and build tools that reduce the amount
>    of time and effort required for us to self-host.

Right, because we probably only have 3 or 4 sites (home, work, school/club, family?) with a half dozen machines at each, max. And median might be 2 sites, 4 machines total.

This means some aspects of devops tools made by and for (0) are inappropriate for us.

I believe the ephemeral nature of containers is appropriate, and missing from your earlier post.

> 2) Your "average" free software operating system user.  Intentionally a
>    bit vague!  Assume they know how to install packages, maybe through
>    something graphical like synaptic usually, but can follow some
>    tutorial online and drop down to the command line if they really need
>    to.  This milestone is all about making your userops system "as easy
>    as" running synaptic or apt-get/yum, which is to say, it could be
>    better, but most of the system has been simplified for you.
>
>    (Note: If you think "don't distro package managers already solve
>    this?" I'd encourage you to watch the talk Deb and I gave at FOSDEM
>    this year:)
>
>      http://ftp.belnet.be/FOSDEM/2015/devroom-distributions/can_distros_make_the_link%3f_lets_package_the_customizable,_free_software_web_of_the_future!_.mp4

I think if we can measure success at making it easy for (1) by the extent the solutions work for this group; which is to say i think this is a reasonable user group to design for.

> 3) The "tech-savvy" family member.  This person is the person who
>    usually installs the wifi router for the household and usually gets
>    asked to help other family members deal with basic problems.
>
>    Most family members have someone who knows well enough to help others
>    debug.  Maybe your system has some tricky edges to it, so that person
>    may help other family members come up to speed, or install something
>    for everyone else on their behalf.
>
>    (Aside: if you reference this example, please don't use sexist/ageist
>    tropes when describing roles.)

Is this typically someone who has heard of "open source" but is unlikely to be able to define it in a way a software freedom activist will be content with, and is unlikely to run any GNU based systems?

This seems like a person who might *buy* a gnu based server to put behind the sofa (or rent remotely) with userops capabilities, along with the roku or other such systems they've picked up in bestbuy, but these people aren't going to  become (2) any time soon.

Thus this group seems out of reach.

> 4) A very non-technical user.  This person may be able to install
>    applications from (sorry to use this term) "an app store" or etc, and
>    interacts with computers, but lacks a lot of self confidence or
>    skills regarding understanding how the technology they're working
>    with works at all.

And this is typically someone who is unlikely to have heard of "open source," no longer owns a desktop and is laptop/tablet/phone only, doesn't buy any sofa servers as uses the one supplied by the cable company... UserOps is a lost cause :)

> As such, I think that's a spectrum of usability... anything along 1-4 is
> helpful in some ways, though the closer we can get towards 4, the more
> we can help individuals be autonomous.

I can't understand how you could get close to even 3 without founding a (cooperative, 501c3 backed, etc etc) business that ships preinstalled servers to people.

> Personally, the system I am working is being architected for an end
> result of maybe hitting 3, but the initial implementation will be
> targeting 0-1 with a roadmap for hitting 2-3 along the way.
>
> What do people think?  Is that a reasonable way to frame things?

If this software will not run on OSX and Windows, your addressable market is the 1% who use gnu based systems, mostly Ubuntu.

>  - Chris
> _______________________________________________
> Userops mailing list
> address@hidden
> http://lists.mediagoblin.org/listinfo/userops