[OpenIndiana-discuss] [oi-dev] [illumos-Developer] OpenIndiana and illumos, part 2

[Christopher Chan]

On Saturday, November 20, 2010 05:40 PM, Jose-Marcio Martins da Cruz wrote:
> Christopher Chan wrote:
>> On Saturday, November 20, 2010 07:56 AM, Gary wrote:
>
>>>
>>> I'm replying to this thread here instead of on the developer lest
>>> someone issue me a netiquette citation for being off topic. How do you
>>> quantify something like that? Even if you have some industry confirmed
>>> sales numbers comparable to IDC tracking desktop PC and notebook
>>> sales, how do you figure out just how many users a server has
>>> regardless of its operating system? Does a web server have a half
>>> dozen users because there are two sysadmins, two content providers,
>>> and two developers? Or does it have 10 million unique visitors every
>>> day and therefore have ten million and six users? Whenever I see this
>>> comment it boggles my mind -- especially when in the context of Unix
>>> systems regardless of flavor. For example, the commercial OSes that
>>> have sold licenses based on 10 users or unlimited users. Ten users of
>>> what? Shell accounts? Ten entries in the password file? What does that
>>> mean and how can you claim that one OS has more "users" than any
>>> another?
>>>
>>
>> I think we can safely assume this to mean installations. Number of
>> people that actually use the installation would seriously inflate the
>> numbers. If we go by the latter, you have more than 750 users of
>> OpenIndiana already from just my installations alone.
>
> Thanks Chris, you've perfectly understook this even without knowing the
> context I said it.
>
> <rant on>
> In some communities it's becoming really hard to open your mouth without
> risking to be flamed...
> </rant off>

In what way did I flame Gary? If expressing my opinion equates flaming 
then I feel very sorry for you. In fact, if you want an example of a 
flame, maybe what seems to be a sarcastic reply higher up seems to smack 
of a flame more than my reply since I did not imply anything about Gary.


>
> But in the context, I told that if OI wants to innovate, a support from
> some big companies is a requirement.
>
> And to explain what innovate means, in my mind, I'm thinking about
> things like improving the kernel threads model, or creating new
> features, *from scratch* as did Sun, features like ZFS, DTrace, zones
> and so...
>
> Examples of inovation mentionned at oi-dev list are adding KDE, or
> removing the question "are you in a sub net" when using "zlogin -C" for
> the first time. In my mind, these are just examples of integration
> solutions, hacks, or similar things, not innovations...

So they are not innovations that sprung out of nothing but when doing 
something new by integrating existing technology to bring about a more 
comprehensive experience sure counts in my book.


>
> So, to really innovate, at research level, you shall be funded and
> supported by someone, with a team big enough and with required skills...
> Not just a small hand of integrators, as it seems we have here.
>
> Examples of FOSS supported by big companies are are Fedora, postfix,
> sendmail (for some time) ...
>
> And, as I said, if the number of, say, installations, is very low, it's
> harder to get some support from big companies...

postfix was written from scratch without any existing user base by 
Wietse for IBM. Upstart in Ubuntu likewise for Canonical. So too 
reiserfs for an example of something in a kernel for DARPA. Linux itself 
had zero commercial support in the beginning. The number of 
installations or the number of users does not necessarily have any 
contributing factor to whether some 'big company' will support the 
research and development of something. The Linux kernel was offered an 
enhancement feature by a single person who was not a C programmer by 
trade. I am not saying that this is the way to go but that we should not 
preclude innovation (features from scratch as written in your book) 
coming from seemingly impossibly resource constrained sources.


>
> Just a complement, I'm not expecting OI to remain, in the future, fully
> compatible with Solaris, as I think it will undoubtely diverge. I expect
> OI to be just an alternative to Oracle Solaris. In the same way that
> creating a new OS 100 % compatible with Microsoft Windows is something
> nobody is looking for.
>
> Well, I'll close my mouth, from now...
>

We can't have that. If everybody stays mum then how can we get a list of 
ideas for vetting? For now I think we should stop worrying about where 
innovation will come from and concentrate on keeping Openindiana relevant.