Re: Competition

[Scrap Happy <Scrap%GMX.com@localhost>]

Hi Magnus,

[please cc via email as I don't subscribe to -advocacy]

On 8/24/2011 10:24 PM, Magnus Eriksson wrote:
> On Wed, 24 Aug 2011, Scrap Happy wrote:
>
> > > So why does this community seem to think that it's somehow a bad
> > > thing trying to increase the pool of users and potential developers?
>
> > I think you (we) have to decide what you (we) want in terms of users.
> > Sure, more users *can* be better. But, what do they bring to the
> > table -- besides their chair (seat)?
>
> I don't think I get this line of reasoning at all...
>
> My view is that more users will mean more future developers, more
> feedback, more community. Not everyone will be interested in getting
> involved of course, but all things being equal, more means more in the
> long run.

Users add costs to everything.  They *can* add benefits.

If the sorts of users you attract represent costs without
benefits (e.g., requiring support but never doing anything
that attracts other developers, vendors, etc. to the market),
then you have to support *their* needs ("How do I install
MythTV?") without getting anything positive (besides a seat)
out of the transaction.

Note that I'm not trying to pidgeon hole these folks.  Rather,
I'm acknowledging that there are costs to users that would
(hopefully) be offset by *benefits*.

> > > just going with what people around you are using, and being nudged by
> > > the occasional "hey, this stuff is really cool".
>
> > > So lets talk to people about the stuff that is cool once in a while.
>
> > But "cool" means different things to different people.
>
> I'm not saying we should have to agree about the definition of "cool",

Understood.

> just that it won't kill you/us to occasionally say that "my NetBSD box
> does X, which is cool". (Which might influence others who share the same
> view of cool.)

Yes.  That's "advocacy"  :>  (which need not be evangelism)

When I show folks a deployed system and the questions turn technical,
it's fairly obvious to them that "It's not Windows".  So, the next
idea that crosses their mind is "It runs Linux?".  "No, this is
NetBSD based."

> > Secondly, they like to *play* with their computer. They don't mind
> > spending a week/weekend setting up MythTV
>
> I believe you're describing tinkering, not playing. (Which I agree can
> be fun, though, as long as things keep working...)

Yes, tinkering, playing, etc.  My point was they don't ever expect to
*do* anything "productive" with the machine.  Just "tinker" endlessly.

> > To these folks "cool" is synonymous with "toy". And, we all know
> > how quickly children outgrow toys... :>
>
> I guess we're from different planets. What is working on an open-source
> operating system, if not exactly tinkering, for fun? It just happens to
> be a kind of tinkering that (I'm guessing) both of us approve of.

No, you missed my point.  I'm talking about passive users who are
just playing with features.  Like getting speech output working
just to show that they can make their computer "talk" (yet never
needing or using it other than to amuse themselves or some other
friend/acquaintance).

> > I point this out because you spend your (limited) efforts where it
> > will give you the most return for the type of users you want to
> > attract.
>
> So what type of users *do* you want to attract then? Me, I want all kinds.

I don't know.  I'm just saying that, I think, you have to identify the
types of users you are trying to draw in.  E.g., if you want "lots of
seats", then you probably want a working, integrated "Desktop" with
a turnkey installer so folks can just throw a CD in and click "install"
(note that I said *click* not "Hit Enter")

OTOH, if you want a platform for folks to design/deploy embedded
systems, you probably end up with fewer *developers* but *many*
more "deployments" ("This toaster powered by NetBSD")

Different efforts, different types of users.  If you have unlimited
talent available, then you can address all of the above.  If not,
you have to pick and choose your battles more carefully.

> > E.g., I use very few of the multitude of packages available. I
> > don't care if GIMP runs well or if the latest version is
> > supported on NetBSD -- I'll drag out Photoshop and be done
> > with my work before the package installs itself and all its
> > dependencies. Likewise, OO is more investment than I would
> > care to make -- FrameMaker handles my DTP needs quite well.
> >
> > OTOH, every hour put into tweeking and improving the kernel
> > itself rewards me directly. Folks don't care if I was able to
> > prepare my images (Gimp), documents (OO), etc. on the system;
> > what they want is a system that *runs*... *my* code.
>
> I don't quite follow. So if you actually want stuff done, you don't use
> NetBSD at all, but kernel work somehow still benefits you? (What is "my
> code"?)

No, if I need certain types of "things done", I pick the tool that
gets me to "Done" quickest/most efficiently.  Often that is NOT
running on a *BSD.

I do most of my software development on NetBSD (though sometimes
"type in the code" on another platform and transfer the files over
prior to compilation).  And, *deploy* NetBSD-based systems (hence
the benefit of changes to the kernel, etc.)

> On the other hand, if I was picking an OS specifically for, say,
> embedded use, I wouldn't necessarily want to restructure my life around
> using it for absolutely everything either. Is that what you mean?

Yes, but my comment was intended to be even more general.  I pick
whatever tool solves my problem best.  I'm not a "zealot"  :>

> > Again, I think it depends on what audience/user base you want to target
> > and what you want *from* them (since you are investing something to
> > support that group). You have to consider the costs and their
> > potential benefits.
>
> I'm getting this weird elitist sense from your mail. Something like;
> people that want to use NetBSD for the right reasons are welcome, but
> lusers who just want "that other stuff" (MythTV, a desktop box, ...)
> should go elsewhere. Is that close?

No.  All I am saying is that you have to (IMO) identify the types of
users you expect to draw with <some_development_activity> and decide
if it is an efficient use of your (limited) resources.

I was commenting that many (esp younger) users tend to be drawn
to bells and flashing lights (e.g., MythTV).  So, if you want lots
of "buzz", effort expended on a killer-app along those lines might
be worthwhile.

If you want lots of *seats* (with less "pizazz"), you might want to
work on a Desktop.  Not much "glory" but lots of users (you then
need to be sure they can get the answers they need *easily* else
The Boss is going to see your solution as counterproductive)

If you want lots of "deployments", you might want to work on
more ports and/or ways to modularize the kernel so that tinier
and more (resource) predictable implementations are possible.

*I* don't know what the right answer is.  I'm just commenting
on my observations from the local "Free UN*X Group", my "peers",
etc.  Showing *them* a wireless X server running on an iPhone
gets a "cool" -- since they can make the next intellectual step
to what that capability gives them.  OTOH, showing them MythTV
running on "yet another OS" is old news (unless it was faster,
less resource intensive, easier to install/configure, etc.)

A live CD of a complete Desktop would probably qualify as an
eye-opener!

> Apologies if I've misrepresented you.

--don