Index: [Article Count Order] [Thread]

Date:  Tue, 24 Apr 2007 13:58:48 +0100
From:  Michael Stauber <bq (at mark) solarspeed.net>
Subject:  [coba-e:09695] Re: Upgrade to CentOS5
To:  coba-e (at mark) bluequartz.org
Message-Id:  <200704241458.49568.bq (at mark) solarspeed.net>
In-Reply-To:  <462D43DE.6040000 (at mark) planetcentral.net>
References:  <039701c785bc$64b751f0$3946460a (at mark) CIIC2> <462D43DE.6040000 (at mark) planetcentral.net>
X-Mail-Count: 09695

Hi Paul and Arthur and all others that have voiced concerns,

In advance I'd like  apologize for not answering all related message in 
detail, but there have been a few negative, false or misleading comments so 
far, where I have to jump in and say something. It will be a single answer on 
this this topic in general and after that I'll have nothing else to add.

> However, with what sounds like the "new" version of this solution coming
> out, no doubt the older version will continue to be updated and
> supported less and less 

To the contrary.

The "other project" uses BlueQuartz in the form that you can create virtual 
servers based on BlueQuartz. For those virtual BlueQuartz servers it uses the 
very same RPMs and code that you can download for free, or which you find on 
Brian's free CD. We had to make a few modifications and improvements of the 
existing code in BlueQuartz to make this happen. But these changes are going 
to be part of the general release as well sometime down the road.

We could as well have forked the BlueQuartz code tree and not offered to give 
something back to the community. Instead we renewed our commitment to the 
BlueQuartz project as clearly indicated by the recent move to Sourceforge and 
the addition of more code maintainers. And not all BlueQuartz code 
maintainers work on the "new project".

>  this means that unless BQ lives on under Centos
> 5 as a free offering, these things probably will come to an end

BlueQuartz will always be free and it will receive a lot of attention in the 
future to improve it beyond it's current capability.  I think on one issue 
(and that one issue only) I can lean myself out of the window a bit and speak 
for all code maintainers that work on BlueQuartz: 

We want to push BlueQuartz to a point where it really can compete on par with 
solutions like Plesk or Cpanel. 

To make that happen we'd have to provide compelling reasons to big ISPs (the 
really big players!) to use BlueQuartz over Plesk or Cpanel.

Right now they're not using it for a number of reasons which we are aware of. 
Partially that's due to missing features which other hosting panels have. Or 
that's due to CMU being the only "supported" method of doing backups. Another 
reason may have been the way how the BlueQuartz project presented itself to 
the outside. And there is also the point that if a company really makes 
itself dependent on a certain piece of software, it usually also wants to 
know how it is supported and who they can call at 2:30 a.m. on a Sunday night 
if they're in a bend with it. 

The virtualization approach allows to include a new backup method which 
previously wasn't possible. This scratches one of the old "faults" or 
weaknesses of BlueQuartz off the list for now and gives us the time to work 
on a good backup solution that works as well from inside the regular 
BlueQuartz as well. Likewise Brian and I can offer (and always have offered) 
commercial support for BlueQuartz itself for those that require it *and* we 
always have provided free support on this list, or by email or forums.

Of course there is the current push and shove underway to make BlueQuartz 
ready for CentOS5. Once that's finished, you can just run your "yum update" 
on your current BlueQuartz and leap ahead to BlueQuartz on CentOS 5.

The only thing that really has changed is that BlueQuartz is starting to get 
more viable and more attractive hosting solutions for people that so far had 
no compelling reason to use it. At the same time the current users will also 
benefit from the planned changes and upgrades to BlueQuartz itself.

So don't worry. 

There has been another undertone in the feedback so far where people 
complained about the pricing of "the other project" or that it costs money to 
begin with.

Lets play the number game and use small numbers which might not even reflect 
the real effort: Figure in 4h a day for two people working on something for 
half a year - weekends included.  If you round that up it's around 1500 man 
hours. This is time that could as well have been spent with the wife, or 
working on "a real job" or doing something equally rewarding or less 
stressful.

If you hire skilled programmers, take the average expenses for healthcare, 
insurance and wages and tally it all up, then the usual work hour costs 65,- 
EUR around here. Those 1500 man hours would cost something like barely under 
100thousand EUR. So $500 US won't cover that. It probably won't cover the 
monthly bill for pizza an diet coke for said programmers. ;o)

The last two weeks - so immediately prior to the release - I spent 250 hours 
in 14 days on that project. So that was a flat out 18h a day effort for 14 
days straight, weekends included. That's more than a 9 a.m. to 5 p.m. 
employee works in six weeks. It's also more than the unions or workplace 
related laws allow in most countries. And no, it aint healthy either. :o)

Now think about one thing: Without one form or other of a regular or irregular 
income we wouldn't have the time to work for free on BlueQuartz itself, nor 
would we likely have the desire to do so on this level of participation. 
Without making "the other project" commercial it simply would not be 
available now, nor would it be finished within months - if not years from 
now. Nor would code from it flow back to BlueQuartz itself.

Also: Do not forget that the RaQ550 GUI itself was commercial to begin with 
and remained that way for a very long time - until Cobalt was dead and 
burried. Even then it took quite some convincing by former Cobalt employees 
to get Sun to release it as open source.

Nobody will ever be forced to pay for BlueQuartz, nor are we taking anything 
away from it.  

If you really need clustering or virtual servers, you have plenty of choices 
as well. You can build something yourself in the same way we did it, or you 
can choose from various free or commercial projects. There are indeed plenty 
of choices. Usually people that *need* clustering do so because they have 
important revenue generating data that needs to be available at all times - 
or they loose business opportunities. So they should be able to afford a 
commercial and fully supported solution, or be able to build something on 
their own. If you can't afford it, but are still interested and/or need it 
for a charity project, then you can also ask nicely and in private.  

This will be all I have to say on this matter here. Thanks for reading this 
far. 

-- 
With best regards,

Michael Stauber
http://www.solarspeed.net