User: Password:
|
|
Subscribe / Log in / New account

Meeks: Linux on the (consumer) Desktop

Meeks: Linux on the (consumer) Desktop

Posted Sep 12, 2012 19:07 UTC (Wed) by bojan (subscriber, #14302)
In reply to: Meeks: Linux on the (consumer) Desktop by robert_s
Parent article: Meeks: Linux on the (consumer) Desktop

I never claimed secret development model was required or desired. What is important is real techincal expertise and true dedication to the effort. In other words, a serious player.


(Log in to post comments)

Meeks: Linux on the (consumer) Desktop

Posted Sep 12, 2012 21:25 UTC (Wed) by khim (subscriber, #9252) [Link]

What is important is real techincal expertise and true dedication to the effort. In other words, a serious player.

Right. We need serious player which can herd the hardware vendors and pressure them to keep the platform whole.

Ok, suppose that we've got such player. Adobe, Facebook or may be AMD or Intel decided to become “real serious” about Linux. They find some billions to offer incentives to hardware vendors, they create large group which is supposed to develop the thing. Cool.

Now. They are real serious. They are big enough to push around tens of companies with total market capitalization measured in trillions and millions of employers (Samsung has over 300'000 employers, Foxconn has over million!). Great.

Now, why such a player will want to spent time and effort talking with “community” which can not do anything to it? It's significantly more logical to take what pieces you can from said community and send them packing. If you'll do that then you can seriously discuss release plans, there are no problems with leaks, etc.

Sure, you'll be losing support from said “community”, but we are talking about “serious player”, right?

And you'll succeed then community will come around, too: witness OPIE, Android, etc.

I never claimed secret development model was required or desired.

Secret is how “serious players” are doing things, sorry. If you have nothing to keep said “serious players” honest (for example ready-to-use platform which just need some relatively small push) then it'll be secret development first, code drop later. If code drop will happen at all.

Meeks: Linux on the (consumer) Desktop

Posted Sep 12, 2012 22:58 UTC (Wed) by bojan (subscriber, #14302) [Link]

> Secret is how “serious players” are doing things, sorry.

I find that untrue. Red Hat being the prime example.

Secrecy is absolutely not required, IMHO. What is, is dedication to the effort. This then provides the certainty to the ISVs and everyone else that would like to follow that platform.

Meeks: Linux on the (consumer) Desktop

Posted Sep 12, 2012 23:03 UTC (Wed) by bojan (subscriber, #14302) [Link]

> Secrecy is absolutely not required, IMHO.

BTW, do not confuse secrecy with direction/focus. Say a company actually decided to do this. The first thing to do would be to streamline what is and isn't supported. Prune the base, in other words. This kind of thing does not have to be done in secret or "against community". It just has to be done consistently and with a sense of purpose.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 0:23 UTC (Thu) by khim (subscriber, #9252) [Link]

I find that untrue. Red Hat being the prime example.

Indeed. Just not the example of openness you expect. Message from Matt just a few days ago: what happens October 26th, why it's so important for RedHat to do? Who knows…

Or another older, but significantly more famous example.

Sure, RedHat is quite open about their contribution to upstream, but when their real bread and butter is concerned… they are doing what other “serious players” are doing.

This means that if “community” wants to have any influence it must exercise said before code goes to serious player.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 2:15 UTC (Thu) by bojan (subscriber, #14302) [Link]

Yes, these examples are unfortunate and I wish Red Hat did not go down the obfuscation path for RHEL (actually, I don't know why Oracle don't just buy Red Hat instead and have the real thing - they are a public company after all).

Red Hat do develop the trunk in the open (Fedora), make sure that things do build and that there are real releases out there every 6 months. Community has plenty of influence and it's not like RHEL is an entirely different thing to Fedora.

You know, people cannot complain that Fedora is the perpetual alpha/beta of RHEL and that RHEL is not developed in the open at the same time. These things are mutually exclusive. What is true is that RHEL is a branch of Fedora. Once the branch is taken, the development is taken essentially in house - that much is true.

But, my point was and still is something else. It is the commitment and expertise that counts. The engineering resources, the contract with OEMs and ISVs, the financial capacity and most importantly a stable strategic direction.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 2:28 UTC (Thu) by bojan (subscriber, #14302) [Link]

> I don't know why Oracle don't just buy Red Hat

Sidenote: I know that pretty much everyone hates Oracle these days (yours truly included), but an Oracle that swallowed Red Hat would be a good candidate for a company that could pull Linux desktop off. Sure, they would have to lose a lot of the pigheadedness in the process (i.e. lose the Orable moniker), but I do not see it as impossible.

PS. I cannot believe myself that I'm almost "advocating" such a thing here, but this is how big business works, I'm afraid.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 21:27 UTC (Thu) by khim (subscriber, #9252) [Link]

You know, people cannot complain that Fedora is the perpetual alpha/beta of RHEL and that RHEL is not developed in the open at the same time.

Yes, they can. Because both facts are true. Fedora is perpetual alpha/beta of RHEL and RHEL is not developed in the open.

High risk/high disruption things are done in the open. But then “enterprise features” are added behind the closed doors. So Fedora users are used as free alpha/beta testers but never receive the end result which is developed in secrecy. It's not too bad: if you really want the finished product without the support service then CentOS closes the loop. But to claim that Fedora users are not treated as alpha/beta testers for RHEL or that RHEL is developed in the open is to lie about the state of affairs.

And this is RedHat! One of the most FOSS-friendly companies around!

These things are mutually exclusive.

Absolutely not. Google does the same with ChromeOS BTW: 99% if code is developed in the open, but 1% which makes it possible to, you know, watch the videos on Netflix or talk with your peer in Google Talk... these pieces are developed privately.

But, my point was and still is something else. It is the commitment and expertise that counts. The engineering resources, the contract with OEMs and ISVs, the financial capacity and most importantly a stable strategic direction.

Right. And this is where all the secrecy comes from. You need to herd the cats: ASUS, Acer, HTC, Motorola, and Samsung may want to jointly develop OS—but they don't want to ever show the unique differentiators before release! And of course if they need changes in the system core they don't want to show them to the cheap chinese competitors, too. That's how all the complexity Android's VCS is born.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 22:45 UTC (Thu) by bojan (subscriber, #14302) [Link]

> But then “enterprise features” are added behind the closed doors.

And released as open source.

Of course Red Hat backport a whole lot of drivers and fixes to RHEL kernels - that is what customers pay for - to run and old, binary compatible, patched code on existing/new hardware for a long time. And this is how this "branch" differs from Fedora (why would they be porting all this to Fedora kernels, which are always based off fairly recent mainline, that contain all this, really escapes me). It is unfortunate that they are obfuscating the source of RHEL kernel by not releasing the patches, but it sure is not closed.

However, the tip of the development is Fedora. It is done entirely in the open and is a "trunk" (of sorts) for RHEL.

> So Fedora users are used as free alpha/beta testers but never receive the end result which is developed in secrecy.

You are talking as if these are some sort of proprietary features that only exist in RHEL. Can you please tell us which ones they are?

Of course Fedora users never receive the backports of the old stuff. What would be the rationale for that? Why backport what exists upstream already?

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 23:06 UTC (Thu) by bojan (subscriber, #14302) [Link]

Just one more point here:

> Yes, they can. Because both facts are true. Fedora is perpetual alpha/beta of RHEL and RHEL is not developed in the open.

> But then “enterprise features” are added behind the closed doors. So Fedora users are used as free alpha/beta testers but never receive the end result which is developed in secrecy.

How are Fedora users being used as alpha/beta testers of the code that does not exist in Fedora (i.e. enterprise features that are added only to RHEL behind closed doors)? Or more specifically, how am I right now testing RHEL7 code that is not in my F-17 installation? I would really like to know by what magic this can be true.

Answer: I'm testing it, because it _has_ been added to Fedora.

Meeks: Linux on the (consumer) Desktop

Posted Sep 14, 2012 12:20 UTC (Fri) by khim (subscriber, #9252) [Link]

Answer: I'm testing it, because it _has_ been added to Fedora.

You have a point. Just… please remind me when exactly libsdp was added to Fedora? It's included in RHEL5 released half-decade ago.

Meeks: Linux on the (consumer) Desktop

Posted Sep 15, 2012 8:56 UTC (Sat) by bojan (subscriber, #14302) [Link]

> Just… please remind me when exactly libsdp was added to Fedora?

http://lists.fedoraproject.org/pipermail/test/2005-Novemb...

> It's included in RHEL5 released half-decade ago.

And dropped from RHEL6 as well, because it was dropped from Fedora.

So, this is your big proof of secret development. A non-inclusion of an open source library for migration of existing apps - something you can compile and LD_PRELOAD yourself.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 16:52 UTC (Thu) by mjg59 (subscriber, #23239) [Link]

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 1:32 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

Can you kindly point me at their current kernel development tree?

Thanks.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 1:53 UTC (Thu) by bojan (subscriber, #14302) [Link]

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 2:16 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

Bzzt! Wrong answer.

I'm talking about their next 'enterprise' RHEL kernel (which is also used by CentOS). For which we don't even get patch-level info now.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 2:29 UTC (Thu) by bojan (subscriber, #14302) [Link]

Yes, you cannot see that "branch". Only the "trunk" (i.e. Fedora).

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 3:57 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

Nope. Their enterprise kernel is most definitely NOT a branch of Fedora's kernel.

Sure, they influence each other, but they are NOT the same.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 4:59 UTC (Thu) by bojan (subscriber, #14302) [Link]

Of course they are not the same. Once the Fedora is branched off to become RHEL, changes are made on the branch that are not in Fedora.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 5:19 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

Wrong. RHEL is not branched off of Fedora.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 5:46 UTC (Thu) by bojan (subscriber, #14302) [Link]

Please be serious. Are you honestly suggesting here that Red Hat engineers first build Fedora, then forget all about it and build RHEL in an entirely different way? How do you then explain changelog messages from Fedora packages in RHEL packages?

Sure, they do put some stuff that is not in Fedora in RHEL. But they are far from being idiots, so they reuse vast majority of _already_ _tested_ code from Fedora.

Here is a good article for you to read:

http://www.redhat.com/resourcelibrary/articles/relationsh...

Quote from it:

"The size and expertise of the Fedora community make Fedora an ideal incubator and proving ground for features that eventually get incorporated into Red Hat Enterprise Linux. To meet the quality and reliability requirements that make Red Hat Enterprise Linux the choice for mission-critical applications, Red Hat puts Red Hat Enterprise Linux through its own set of tests and quality assurance (QA) processes that are separate and distinct from those of Fedora."

I hope you are not suggesting that Red Hat engineers "rewrite" those features for RHEL just to make it different.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 5:52 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

I did say that both kernels influence each other. However, RHEL kernel is very clearly NOT branched off of Fedora's kernel. They share many patches, but RHEL gets a lot of stabilization work and backports.

That stabilization work makes it special. And it's not available to the public for exactly the same reasons - community does not help much in this case.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 6:13 UTC (Thu) by bojan (subscriber, #14302) [Link]

> They share many patches, but RHEL gets a lot of stabilization work and backports.

Well, yes. And I agreed with you there.

> However, RHEL kernel is very clearly NOT branched off of Fedora's kernel.

So, your theory is that Red Hat have a completely independent RHEL tree in house and they have some secret society test just those kernels for years, before they become real RHEL kernels. I think you are mistaken. They would not have enough man power or wide enough configurations available for that. Their "super stable" RHEL kernels would suck big time if they did that.

Instead, Red Hat let new kernels into Fedora for community to test (for instance, people running F-17 are testing 3.5.x for them right now, people running F-18 what will become 3.6.x etc.). Then, at some point, dictated by internal release schedules of RHEL (i.e. the mysterious dates RH employees sometimes slip into public domain), they branch that off and start stabilisation work, based on various patches from older/newer development.

Otherwise, what's the point of having Fedora? They are not doing it only out of being nice, I am pretty certain of that.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 6:21 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

>So, your theory is that Red Hat have a completely independent RHEL tree in house and they have some secret society test just those kernels for years, before they become real RHEL kernels.
That's EXACTLY what happens. RHEL kernel is branched off about a year before the first beta and 1.5 years before the final release. With lots of backports, of course.

Case in point: RHEL 6 which was released on 2010-11-10 is based on 2.6.32 which was released on 2009-12-03 (and that's unusually quick turnaround for RHEL).

And they're supporting it until 2018 (at least), with backports of new features and bugfixes.

>I think you are mistaken. They would not have enough man power or wide enough configurations available for that. Their "super stable" RHEL kernels would suck big time if they did that.
They employ a lot of kernel developers precisely for that very purpose.

And that's the reason why RedHat is a big company.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 6:27 UTC (Thu) by bojan (subscriber, #14302) [Link]

> RHEL kernel is branched off [...]

Branched off what exactly? Fedora, of course.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 6:29 UTC (Thu) by Cyberax (✭ supporter ✭, #52523) [Link]

Actually, no. They begin with the mainline kernel and start applying patches on top of it. Some of these patches, of course, were in Fedora.

But by the time they start a new development cycle, Fedora is already several kernel versions in the future.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 6:45 UTC (Thu) by rahulsundaram (subscriber, #21946) [Link]

"Actually, no. They begin with the mainline kernel and start applying patches on top of it."

Where did you get that idea?

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 7:02 UTC (Thu) by bojan (subscriber, #14302) [Link]

Here is a changelog message from one of the RHEL5 kernels:

* Fri Jul 07 2006 <name> <email>
- Unified rhel and fedora srpm

> Actually, no. They begin with the mainline kernel and start applying patches on top of it. Some of these patches, of course, were in Fedora.

This is just playing semantic games. All the patches that Red Hat do on recent mainline kernels end up in Fedora, which are then (attempted to be) pushed upstream. So, by the time this RHEL kernel is based on some x.y.z upstream release, many of those Fedora patches won't even apply (because they are already part of mainline). Of course, features that RH decide not to support in RHEL (that would be the streamlining thing I was talking about) will not appear, will not be patched and will not be compiled.

> But by the time they start a new development cycle, Fedora is already several kernel versions in the future.

I do not see why that would be surprising at all. And, of course, this is where the "trunk" of new RHEL is being forged: by applying Fedora specific patches, testing them in the wild, getting feedback and pushing them upstream. This is how they _leverage_ the community. If they did not do that, that would be really stupid. And I think Red Hat's balance sheet speaks volumes about them not being stupid.

PS. If you are trying to suggest that RHEL kernel is physically not a branch of some SCM Fedora system, then you may even be correct. But, that is not what I was talking about when describing the non-secretive trunk of the Red Hat's distro development.

Meeks: Linux on the (consumer) Desktop

Posted Sep 13, 2012 7:30 UTC (Thu) by bojan (subscriber, #14302) [Link]

> If you are trying to suggest that RHEL kernel is physically not a branch of some SCM Fedora system, then you may even be correct.

Just downloaded a RHEL6 kernel SRPM. The spec file mentions the word "fedora" (case insensitive) in 12 lines. So, even technically, you are probably wrong.

Meeks: Linux on the (consumer) Desktop

Posted Sep 15, 2012 18:21 UTC (Sat) by pboddie (guest, #50784) [Link]

Now. They are real serious. They are big enough to push around tens of companies with total market capitalization measured in trillions and millions of employers (Samsung has over 300'000 employers, Foxconn has over million!). Great.

Employees. Unless you're of the opinion that these companies "work for you".


Copyright © 2018, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds