UNIX admin wrote:

Did you ever consider that one possible reason is
that Solaris X86 simply doesn't have all the drivers - I have a laptop, and it's less than 2 years old, which Solaris will install on, but there simply are no network drivers for it. So it would take much more of my time to get to a usability level on Solaris X86 with that machine that it would with JDS/Linux, hence I'm sicking with the latter for the moment, it doesn't rule out Solaris X86 - I still have it installed and if I manage to locate drivers for it, I'll move, but it's the time investment needed that I simply don't have right now.

In my 20 years of experience working with computers, if there is one thing that 
I've learned, then it's

[I]get hardware that supports your software, not software that supports your 
hardware[/I]. Now what does that mean? It simply means, if for example Solaris 
has built in drivers for a RTL8139 ChipSet, I'll make sure I get HW that has 
that ChipSet. If the SW supports printing to a PostScript printer, I'll make 
sure that the printer I buy understands PostScript.

This is a lesson I learned very well during my Commodore Amiga days, and it 
stuck ever since. As a result, I've been relatively [I]unscathed[/I] in regard 
to all this stuff.
That's all great in theory - but it's not always like that - while there may be support for one version of a device, if that device gets up-reved the whole chipset can change (just look at how wireless drivers are these days). For Small to Medium enterprises it's not always possible to purchase the best devices, or even Sun hardware, and then when they go to purchase something they can afford - all to often it using the latest versions of h/w which quite frequently is not supported by Solaris X86 - so what do they do - buy someone else's scrap because it happens to have the supported devices they want?

What you suggest is entirely unrealistic for everyone - and most likely why there is considerable interest in getting a device driver community under way - and the more that Open Solaris is being used should also encourage h/w developers to be more encouraged in providing the drivers themselves - with our help.

Be that as it may, when was the last time you installed Solaris on the x86 
platform? Solaris10 has very good HW support for most common HW, and the OEMs 
are starting to release [I]Solaris native packages[/I] of their drivers right 
on the 'install CD' along with Windows drivers!
I've installed the latest Solaris Express on my laptop a number of times to try get the latest drivers - still doesn't work ... I also know (and work with) plenty of people who have tried to also install on desktop machines and other laptops that have wasted their time doing the same.

I can only hope that the number of people producing Solaris drivers increases - we in Sun are certainly trying to encourage it, and as I said, Open Solaris will help here.

For example, I just about fell on my behind when I found Solaris native 
packages of drivers from a [I]vanilla Taiwan manufacturer[/I] of the machine I 
bought.
I think I would have too... :)

Further more, the laptop I'm writing about is a Compaq AMD Athlon "run of the 
mill" thing. It has a Savage 3D graphics adapter, RTL8139C, and some sort of AC '97 
audio. All these things are supported by Solaris10 out-of-the-box, Savage 3D included. 
And if my audio won't work, I'll just get the audio drivers from

http://tools.de/solaris/audio/

And I'm done. Easy as pie. I could probably even get Solaris9 to work on that 
thing. Where's the problem?
The problem is that you had to do this at all - it should be possible to get these as part of Solaris, especially the common stuff (which I'm assuming it is if someone went to enough trouble to write it).

Solaris10 on x86 has really come a long way, and its support for x86 is rapidly 
increasing -- in fact, I've never seen any platform's support start to build 
momentum so rapidly.

I totally agree on this - but you have to be realistic and accept that sometimes it simply does not work out of the box, and things should easier than they are at present.

As stated in other threads - it's crazy the amount of
extra libs that get downloaded if you pkg-get something like this - and in this case it was intended to be a learning experience w.r.t. what is involved in building such applications as well.

As I stated before, Blastwave currently suffers from lack of engineering and 
quality control. But they have come a long way, and I'm confident that 
Blastwave will only get even better as the time passes.

What "Nekoware" is to SGI IRIX (http://www.nekochan.net/downloads.php)
"Blastwave" is to Solaris. Blastwave is already the #1 Freeware provider for 
the Solaris platform. It can only get better from this point on.

As an interesting side note, the Nekoware guys said they "liked the comparison" 
when I pointed out Blastwave to them. These two projects are very, very much alike, only 
for different platforms (SGI and Sun).

pkg-get is a great idea, but I think unless it's part
of the core, and includes the core packages - in other words a person can update their entire Open Solars based distro via pkg-get - then we will always end up with problems of duplication of packages, etc.

Ah, but the question is, [I]what is considered to be the core, and by whom[/I]?
The distro maintainers (and Solaris is going to be a distro of OpenSolaris) are the ones that decide it.

Plus, this is very much [I]conceptually wrong[/I]. Why? Because "updating" in 
the sense you're writing about could end up [I]removing and replacing entire software 
subsystems[/I], and that [B]rules it out for production right on the spot[/B].
There will be times that you would want this (complete machine upgrade) - but in the more general sense the distro would decide this and simply not do it if (as in Solaris) it is likely to cause major customer problems. Most of this is focused on the server, but for the desktop user - quite often they don't care as long as they have a very functional (does not just mean eye-candy) desktop.

This is the exact same approach that Linux takes. The only way to patch a software 
subsystem through the OS interfaces is to do `rpm -u` which goes and [I]replaces the 
entire software subsystem[/I] in order to "update" it.
This is very much busted. For example, SuSE is most notorious for breaking 
production systems in this way, especially when they do kernel updates with 
YaST.

Again, I think this depends the uses of the OE - server or desktop - Open Solaris is aimed at getting recognition - not just making money - the more people that see Open Solaris the better - and this also goes for the notion of a Debian distro - you have a choice - if you want stability take Solaris (as I say it's a distro based on OpenSolaris), if you want the latest of everything - take another distro (eg. SchilliX, Debian or whatever other distros might come along).

What Alo was suggesting was having a Debian distro
using an Open Solaris kernel - why is this such a bad thing? I can only be

Because that would mean you'd have a Solaris kernel surrounded with GNU tools. 
It would effectively behave like a Linux distro, only instead of the Linux 
kernel you'd have a different kernel.

The people that make these kind of suggestions don't understand Solaris, more 
importantly, they don't understand UNIX! They are proposing to start fitting a 
square peg into a round hole, because the only experience that they have is 
working with a round hole!
That's just plain wrong. They should learn about the "square peg", instead of trying to 
make it a "round peg". That's just BUSTED. I can't even begin to explain how wrong it is.

But you are totally blinded by the "production" environment - Solaris is aimed at that market, yes, but something like Debian is not - and as I say it's about choices...

A Debian distro based on Open Solaris is a perfectly valid idea, and for some people would be perfect for them - this is exactly why people like RedHat and Novell/SuSE have the concept of an enterprise distro and a Fedora Core or Novell Linux Desktop - for people that want stability and control (as in a production environment) use the enterprise version (of OpenSolaris this would be Solaris) but for the people who like to be on the bleeding edge - take the distro that suits them.

Personally I am one of those people that likes to be on the Bleeding Edge - for my desktops - so I would be only too delighted to see a less stable distro that provided this but even better be based on Open Solaris - not Linux.

good for Solaris / Open Solaris users since it will encourage the thousands of opensource software packages to be easily compiled on Open Solaris because developers who prefer the Debian way of doing things can still do so while also being able to benefit from a world class
base like Open Solaris.

We really don't want Linux "developers" on board; those guys are hackers that are going 
to bust stuff by "improving" it. What we want are professional developers picking 
(Open)Solaris as a free, better alternative to everything else out there.

That's a totally unjust statement - most developers are hackers at heart, even professional ones - and there are many professional developers that give their time to Linux (are you suggesting that Linus Torvalds is not a professional developer??). OK, there are probably some inexperienced developers out there on Linux but that's why they are there to learn - if we introduce Open Solaris to them, then we can encourage them to learn even more - if the interest is there encourage it (most parents will already know this).

I think it's this "business as usual" that isn't
actually so usable by end-users - based on you're e-mail title, I'm assuming you're a sys-admin - so maybe this is fine for you, but as an end-user this really isn't good enough - and I think this is were Debian (and the off-shoot Ubuntu) have done really well.

What exactly do you find "not so usable"?

Is it the fact that:

a) you have to do `su -`

or

b) type in `pkgadd -d /path/to/package [all|package name]` or pick a choice 
from a MENU presented upon executing the command

or c) something else (lack of eye candy GUI, for example)?

Yes exactly - why should the average desktop user have to do this? That's exactly what is broken on Solaris. Everyday tasks should not require you to drop to the command-line.

I think you are being blinded by your (assuming bad)
experiences with

VERY BAD experiences (as compared to Solaris, IRIX and HP-UX, even AIX). And I 
did Linux professionally, for a living!

Open Solaris is not aimed at just system admins -
it's being targeted at the desktop user, more so now than ever - so live with it, you are going to have to accept that some people simply don't like the way things are right now on Solaris and would like to have a desktop machine that doesn't constantly require you to bring up a terminal to make even the slightest change to the systems configuration.

How will you use your eye candy GUI to make "slightest change" to the system 
when the system HAS NO GRAPHICS ADAPTER but only a TERMINAL or is connected to a REMOTE 
SERIAL CONSOLE MANAGER, and/or there is NO X-window run time environment on a system? You 
will be in trouble. A lot of trouble, in fact, because you'll only know how to do things 
from a GUI and you'll have no clue how to do it on the command line.
Again I think you're missing the point - if you are a DESKTOP user, you WILL have a graphics adapter. Out of the box you SHOULD have X running (if you don't that's a problem in it self that should be solved by the distro not the user).

MANY _DESKTOP_ users, even power desktop users, have trouble with the command-line - why should a user have to run a command-line tool to change their wireless network settings? or to create a printer queue for a USB printer? These are everyday tasks for a user (especially a mobile user). You seen to forget that the majority of people that use desktops are not even trained in IT - they are just given a machine, with a word processor to type up documents, etc.

See, this is exactly what I mean when I write that the new influx of Linux 
converts doesn't really understand the core UNIX concepts.
You obviously don't understand the market in which IT operates - it's not entirely server based! Most of the employees in Sun use desktops, usually running Solaris, for their everyday work - they can manage it fine most of the time but that shouldn't preclude them from being able to have better functionality like hotpluggable devices that just work.

How many Linux machines did you ever see managed in the "lights out management" 
mode, i.e. with an ALOM or an RSM or a CMS?
Again - DESKTOP users!

You are the one who has to understand that not everything revolves around the 
desktop: it's a 50-50 proposition.
It is also rediculous the come in and say "I'm here; I'm a convert; now cater to 
me!"
So what??? If you don't like it, you can always go down to GRUB/LILO and select 
the Linux or Windows distro you have installed on that PC at home.
Totally wrong - MOST people use desktops (certainly not 50-50), not servers in everyday life - why do you think Windows has just a large volume - certainly not just servers Are you telling me that enterprises have more servers than desktops!. Are you telling me that you would consider your laptop to be a server??? Maybe you have it configured that way - but in reality it's a desktop machine.

I cannot believe how closed minded you are - I'm certainly not just a convert - I've worked on Solaris (and SunOS4) in excess of 12 years - so I'm not coming to this totally fresh - I've had to use Linux and Windows in conjunction with my work and can certainly see that Solaris has to do some catching up on the desktop. I at least am open enough to know when Solaris has problems and interested enough to actually want to fix them.

But as somebody pointed out, all the goodies are here, in Solaris, not >>over 
there<< in Linux, so either you learn about Solaris, or you go back. Why was it OK to 
spend the time and effort learning Linux, but it's suddenly not OK that you actually have 
to LEARN REAL UNIX?

Again with the insults - is that how you manage to work? Are you one of those BOFH that people speak of?

Precisely why the Linux community can also benefit
from Open Solaris - which can be achieved via the Debian proposal.

Why should the Linux community benefit? What have they given to Solaris except "it 
works on my Linux system, if it doesn't work for you, use Linux! Everybody should just 
use Linux!"

Linux has opened up unix to the masses - and shown people that the really is a possible alternative to Windows on your desktop. Linux has provided a couple of decent desktops (GNOME/KDE) for people, it has helped expose Microsofts lock-in mechanisms by being there in the desktop space.

Linux has had many problems in it's day - so has
Solaris for that matter - but one of the reasons for putting out Open Solaris is also to enable sharing of technologies - Debian based upon Open

No, that was NOT the reason why Solaris has been released to the public! 
Solaris was not released so that the Linux crowd could bastardize it into 
another Linux distro, but to evolve further as Solaris!
What does it matter to you if some people decide to have yet another distro based on Open Solaris - you obviously aren't interested in one so if you don't like it then don't use it, simple as that.

Open Solaris has been released to encourage developers to get involved - Linux has got to where it is now because developers were encouraged to be involved, it increases innovation, and another platform will also encourage people to look at the bigger picture. Without developers there are no new applications, with out the applications people want, there are no more users.

better, so it would probably stay in place. Why do
you think it's bad that anyone could get the best of both worlds? Isn't this what the opensource community is really about?

Because Linux as a whole does not have anything good to offer, except that "it's 
free". So is Solaris, so that's no argument. It has a big following, but the 
*majority* of people involved are just plain incompetent from a professional point of 
view. They are no engineers, and it shows. Solutions are not well thought out or thought 
through, mainly because of lack of experience and understanding.
So let's teach them - Linux is growing up too, are you telling me that you came out of the womb knowing all? No, you had to learn from experience - Linux has learned a lot, it's made mistakes too, but so does everyone. Many of these engineers are really interested in learning - so let's teach them how to do things properly, and I'm sure in the process we can also learn a lot from them.

Darren.
_______________________________________________
opensolaris-discuss mailing list
opensolaris-discuss@opensolaris.org

Reply via email to