I'd agree on Opensolaris being "too much work" right now for a common user. As
a student trying out opensolaris partly because of working as a newbie
sysadmin, and partly because of *wanting* zfs on my file / web / svn server, I
definitely find the learning curve as too step and unrewarding so far (it
reminds me to my beginnings with vim).
I find unreasonable to appeal to a student and his laptop mainly because of
hardware support and power consumption. Until very recently not even linux
(which I am quite experienced on) could reliably suspend and resume on most
laptops, and the battery lasted much less than on windows. That trend has
definitely changed, and it's one of the driving forces behind ubuntu adoption
in my university (in Madrid, Spain just FYI). I understand the difficulties in
getting proper power management support on any OS, but until that any student
using a laptop will really think twice using opensolaris if the battery lasts
half as long as with linux (which I, however, didn't try as I don't have the
free space).
The other driving force, by the way, has been so far *ease of use*. Long,
twitchy command lines don't scare a long time gentooer as me ;), but even I can
appreciate not having to spend half an hour trying to setup *anything*. Lately,
linux distros have evolved quite quick in that regard: printers work by
default, sharing using NFS or samba is quite easy, networkmanager handles
perfectly the wifi and ethernet needs, and installing packages *and finding
them* is quite easy with synaptic - even "complex" packages such as anything
apache-based, eclipse plugins, whatever. As well, system upgrades appear as a
tooltip with descriptions for the issues solved, etc. I personally find the
experience even easier than windows (true that I'm quite used to it by now).
Solaris, or Opensolaris, does not have any of those luxuries right now. I
greatly look forward to Opensolaris efforts to succeed, as I find the idea of
"making solaris more like debian" quite clever in order to get the benefits
linux does (mainly, ease of package management and upgrading, here greatly
enhanced by zfs magic). However lots need to be done: it's not reasonable for a
student *playing* with something to expect to read the whole beadm man page, or
to find by himself how to export by nfs by inmersing in the huge and highly
detailed doc pages from sun.
Small things don't help as well: if I'm just toying in my free time, all those
small annoying things can be really frustrating, making yourself abandon before
having had time to like the system. In particular, nv_sata forcing me to find a
usb cd drive, broadcom gigabit network card not being supported out of the box,
and not working with the driver provided forcing me to buy another pci one,
keyboard locale being outright *weird* for an es_ES user such as me when
compared to any linux (home / end / DELETE / control-cursor not working as
expected) ... and the general frustration of not knowing where to look for the
answers such as "how do services work compared to init.d", "where is everything
I'm used to such as /etc/exports", "what the **** happens to my
mostly-but-not-fully-right keyboard locale", or "why can't I find nowhere an
easy 1-2-3 guide to sharing with NFS or SMB"
Comparing that to the ubuntu experience I had in my dell laptop (just
installing and *everything* working right out of the box - webcam, remote
control, resolution, 3D support -but not compiz-, suspend and hibernate, great
default selection of packages...), those things really make the difference. I
understand most of them only work because the relatively huge masses behind
linux adoption and testing in all kinds of platforms, and as well because it
evolves quite quickly to adapt to new hardware while being mature enough on the
other basis such as APT.
And that leads me to the additional thing linux (or even other alternatives
such as freebsd, which I have also used a lot) have over OS right now:
community support. I can easily find howtos for pretty much everything
linux-based. Ubuntu and gentoo, in particular, excel at this. Freebsd handbook,
being easily readable and quite exhaustive, could also be mentioned here. Even
public bugtracking (submitting!) helps a lot here, as when you find a bug
another poster has the solution even before integrated in the distro, or
perhaps you can give a solution to a bug you've find quite easily. Community
development starts that way...
I'm still scratching my head on where to start to begin grokking this
OS without being such a pain as to discourage me. For sure if it weren't for
ZFS I wouldn't
even try, with linux being *good enough* for me as a unix development
platform, and quite easier to setup or modify (heck: apt-source
whatever, edit, rebuild package, or even editing by hand shell scripts)
Please don't take the whole rant as an attack or anything; I'm just a
frustrated beginner here who just wants to state some reasons a "common
computer science / engineering student" would throw in the towel, hoping that
stating some problems perhaps not too visible for the already experienced can
be useful to try to *improve* things.
And keep up the good work nevertheless ;), debian 2.2 or gentoo 1.4 were far
more painful to install ;).
--- El mar, 8/7/08, Roman Strobl <[EMAIL PROTECTED]> escribió:
De: Roman Strobl <[EMAIL PROTECTED]>
Asunto: Re: [indiana-discuss] new ISO soon?
Para: [EMAIL PROTECTED]
CC: "MC" <[EMAIL PROTECTED]>, [email protected]
Fecha: martes, 8 julio, 2008 3:06
Dennis Clarke wrote:
> [...]
> It is all too much like work. It doesn't matter if we are talking
> Linux or UNIX. Once upon a time there was no real choice and we were
> forced to install a UNIX system even for departmental level things.
> There was always some head geek that took care of that server because
> it is user hostile by design. Why would I think anything has changed
> in that respect.
>
I think we need to start with the geeks and Linux geeks on universities
are the right target audience to try to switch to OpenSolaris. Then we
can expand further as OpenSolaris becomes easier to use in future
releases. We can attract students/developers on the advantages of
Solaris (ZFS, Dtrace, SMF, SunStudio, ...). But OpenSolaris needs to get
easier to use to become accessible for wider audience (don't get me
wrong, I think OpenSolaris 2008.05 is a tremendous progress in terms of
ease of use compared to previous Solaris releases - it's the first
release that I personally can use ;)
> I look forwards to the Sparc versions. I have run headlong into a
> brick wall on a few occasions because a user looks at me and asks "if
> Sun makes the Sparc and the fancy flying Niagara that you love so much
> and this is their big open source thing then why don't they have a
> version for Sparc?"
>
> I don't know how to answer that at all.
AFAIK the reason for not doing SPARC was that there wasn't enough time
to get it done for 2008.05. The focus of 2008.05 were developers using
laptops - we need students/developers start experimenting with
OpenSolaris, and we needed something to get into their hands real soon.
One reason for this is that in the past Solaris was quite popular on
universities but this has changed with cheap Intel hardware and Linux.
2008.05 gives us an opportunity to re-introduce Solaris at universities,
because it will work on hardware students use (well mostly). SPARC is
very important for Sun of course but that's the next step when the focus
will move towards deployment.
-Roman
_______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
______________________________________________
Enviado desde Correo Yahoo! La bandeja de entrada más inteligente._______________________________________________
indiana-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/indiana-discuss
