Hi,
Open core, for those not familiar with the term is the business model of keeping some key features closed and selling a proprietary product where the "core" functionality is free and open source. Two features in the feature list are such software
http://fedoraproject.org/wiki/Releases/13/FeatureList
Zarafa groupware and IntelliJ IDEA, IDE for Java have a number of features that is only available in their proprietary product.
http://www.zarafa.com/content/editions http://www.jetbrains.com/idea/features/editions_comparison_matrix.html
Do we care? I am concerned about this.
Rahul
On Wed, 2010-04-14 at 06:33 +0530, Rahul Sundaram wrote:
Hi,
Open core, for those not familiar with the term is the business model of keeping some key features closed and selling a proprietary product where the "core" functionality is free and open source. Two features in the feature list are such software
http://fedoraproject.org/wiki/Releases/13/FeatureList
Zarafa groupware and IntelliJ IDEA, IDE for Java have a number of features that is only available in their proprietary product.
http://www.zarafa.com/content/editions http://www.jetbrains.com/idea/features/editions_comparison_matrix.html
Do we care? I am concerned about this.
It did bug me a bit about Zarafa. I'd rather prefer to promote eGroupware, myself...that's what I pushed for in the infrastructure group discussion on calendaring software.
Il giorno mar, 13/04/2010 alle 18.24 -0700, Adam Williamson ha scritto:
On Wed, 2010-04-14 at 06:33 +0530, Rahul Sundaram wrote:
http://www.jetbrains.com/idea/features/editions_comparison_matrix.html
Do we care? I am concerned about this.
It did bug me a bit about Zarafa. I'd rather prefer to promote eGroupware, myself...that's what I pushed for in the infrastructure group discussion on calendaring software. -- Adam Williamson Fedora QA Community Monkey IRC: adamw | Fedora Talk: adamwill AT fedoraproject DOT org http://www.happyassassin.net
And we do have NetBeans and Eclipse, where IDEA is just a cool editor in the current state. On the other end, people may use it and write great FLOSS plugins if they want.
Mario
Hey All,
2010/4/14 Rahul Sundaram metherid@gmail.com:
Hi,
Open core, for those not familiar with the term is the business model of keeping some key features closed and selling a proprietary product where the "core" functionality is free and open source. Two features in the feature list are such software
http://fedoraproject.org/wiki/Releases/13/FeatureList
Zarafa groupware and IntelliJ IDEA, IDE for Java have a number of features that is only available in their proprietary product.
http://www.zarafa.com/content/editions http://www.jetbrains.com/idea/features/editions_comparison_matrix.html
Do we care? I am concerned about this.
We had an interesting talk about the differences between Open Core and Open Source at LOADays last weekend. We went over some of the different "Open Core" models out there, and while we were pretty unanimous in that Open Source is better, one takeaway was that we need to differentiate between who is doing Open Source The Right Way (TM) and who isn't.
There are a number of issues we can identify, mostly orthogonal to each other, but combine defines how "Open" you might say the project/product/company is.
1) The source code is available freely, but not really usable without some tweaks, build hacks and possibly the sacrifice of $furry_animal_that_doesn't_offend_your_religion
2) The source code is available only to customers and is not freely available on the company's website.
3) There are modules you can purchase or get a support contract that adds critical value to the software, but are not available under an open source license; the community is forced to reimplement this functionality.
4) There are modules you can purchase or get a support contract that primarily focuses on interoperability with uncooperative proprietary software, and the code itself is encumbered by legal hurdles such as patents, trade secrets, NDA.
To different degrees, each situation here can be understandable and reasonable or not. I suggest that if this is an important issue, we go into exactly the different models of what's open core and what's just 'barely' open source, provide working recommendations on what works for us, and codify these standards.
In any case, as long as the code itself can be gotten Freely, there's no reason why it shouldn't be in Fedora. Perhaps people demanding more openness because they like the product can change the minds of the company that's creating it.
-Yaakov
On 04/14/2010 01:47 PM, Yaakov Nemoy wrote:
In any case, as long as the code itself can be gotten Freely, there's no reason why it shouldn't be in Fedora. Perhaps people demanding more openness because they like the product can change the minds of the company that's creating it.
Question was not about what should be allowed. It was about whether we should be advertising such software in our promo material.
Rahul
On 04/14/2010 11:12 AM, Rahul Sundaram wrote:
Question was not about what should be allowed. It was about whether we should be advertising such software in our promo material.
I don't feel good about delivering marketing for free to companies that do nothing for Fedora besides putting a spec/srpm into the the repos.
So I would say that the companies that produce the stuff should promote it but not the distro that has them in the repos.
Jan
Il giorno mer, 14/04/2010 alle 11.24 +0200, Jan Wildeboer ha scritto:
On 04/14/2010 11:12 AM, Rahul Sundaram wrote:
Question was not about what should be allowed. It was about whether we should be advertising such software in our promo material.
I don't feel good about delivering marketing for free to companies that do nothing for Fedora besides putting a spec/srpm into the the repos.
So I would say that the companies that produce the stuff should promote it but not the distro that has them in the repos.
Jan
But we can take it to the extreme so. Let's make a couple of examples.
First we have Eclipse. Eclipse itself is a core platform, everything else is based on plugins.
You can definitely use the core version, but may you depend of vendor specific plugins to do you job, just think about WebSphere or MyEclipse with support for the Matisse graphics builder in Eclipse.
I've been using closed things like that for my daily job, for example I have used the WindRiver Eclipse.
You may say that the core version is good enough, but in my opinion is not. So where's the boundary? It's completely user defined.
OpenJDK. Currently the state is of a fully functional product, but we had shipped it when it was not as ready, we shipped IcedTea based on version 7 of the early (day 0 infact) code drop of the JDK. We then added the missing features. You may consider those early days as a core version of the JDK, where support for extra functionality (including, back at the time, running programs like NetBeans, it took some time to get there) had to be found in the closed version.
But let's take it ever more extreme.
Fedora can be considered in some ways as the core version of Red Hat Enterprise Linux.
Red Hat has shipped in the past closed things as value added, like Acrobat Reader or the JDK itslef. Does it means we should not even develop Fedora anymore?
I know, it's a bit like comparing apples and oranges, but what I mean is this: if a feature, as in software, is nice to have, and there are no legal implications, that is, is fully conformant to our current rules and guidelines, I don't see why we should not support it. A vendor has the right to use and modify it's own software and sell it, with what you may consider added values (and thus, a selling point, and more selling means better chances to support the free code base).
On the other hand, it's true that maintaining a package is a costly task, so, if there are no maintainers that want to do this, and if there are better alternatives out there, if the code of those project is not after all so important for us, then it's ok to drop it. The code is out there, you can still use it if you want, so it's not that we kill the project.
But this is a cost-benefit analysis that is always valid, and has nothing to do with the morality of the FLOSS software.
I hope I explained my idea correctly.
Cheers, Mario
On 04/14/2010 03:49 PM, Mario Torre wrote:
But let's take it ever more extreme.
Fedora can be considered in some ways as the core version of Red Hat Enterprise Linux.
Not at all. Fedora is upstream for Red Hat Enterprise Linux and Fedora has more packages than Red Hat Enterprise Linux. Both of these are free and open source versions. At this point, you are very very far from the "open core" business model. The definition of "open core" is very specific.
http://alampitt.typepad.com/lampitt_or_leave_it/2008/08/open-core-licen.html
One of the important requirement of open core is centralized copyright control which none of the examples you cited have.
Rahul
Il giorno mer, 14/04/2010 alle 16.03 +0530, Rahul Sundaram ha scritto:
On 04/14/2010 03:49 PM, Mario Torre wrote:
But let's take it ever more extreme.
Fedora can be considered in some ways as the core version of Red Hat Enterprise Linux.
Not at all. Fedora is upstream for Red Hat Enterprise Linux and Fedora has more packages than Red Hat Enterprise Linux. Both of these are free and open source versions. At this point, you are very very far from the "open core" business model. The definition of "open core" is very specific.
Ok, maybe this was too extreme.
http://alampitt.typepad.com/lampitt_or_leave_it/2008/08/open-core-licen.html
One of the important requirement of open core is centralized copyright control which none of the examples you cited have.
If I understand what centralized copyright control is, OpenJDK has it, we had to sign a paper to Sun to contribute back, they have full control over the code I write for them. Even for GNU Classpath I had to sign this paperwork to the Free Software Foundation.
In any case, the point remains the same. If IDEA is cool to have, there's no reason in not having it because the closed features are of bigger benefit.
I would see it from the other side instead, drop it because the open features are not worth the efforts.
Btw, IDEA is just an example here, I personally do not care. The fact is that I would not like to see that a policy is being defined here, because there are other projects where the difference in features is not so huge, they may be worth supporting then. This should be a package by package consideration, not a general rule in my opinion.
Cheers, Mario
On 04/14/2010 04:36 PM, Mario Torre wrote:
If I understand what centralized copyright control is, OpenJDK has it, we had to sign a paper to Sun to contribute back, they have full control over the code I write for them. Even for GNU Classpath I had to sign this paperwork to the Free Software Foundation.
Yeah but Sun is not selling proprietary extensions to OpenJDK.
In any case, the point remains the same. If IDEA is cool to have, there's no reason in not having it because the closed features are of bigger benefit.
Again, nobody is talking about dropping anything. We are only discussing what we want to *promote*
Rahul
Il giorno mer, 14/04/2010 alle 17.00 +0530, Rahul Sundaram ha scritto:
Again, nobody is talking about dropping anything. We are only discussing what we want to *promote*
Rahul
Yep, I think I misunderstood this sentence:
"I don't feel good about delivering marketing for free to companies [...]" (Jan Wildeboer).
The "delivering" part, hinted to me that this was the case.
In this case I completely agree, it doesn't make sense to promote as our key features when we do have more open alternatives.
Cheers, Mario
On 04/14/2010 12:19 PM, Mario Torre wrote:
You can definitely use the core version, but may you depend of vendor specific plugins to do you job, just think about WebSphere or MyEclipse with support for the Matisse graphics builder in Eclipse.
The core question here stays: Should we *promote* openCore stuff in Fedora Marketing material. Please stay focused.
Do we *promote* Eclipse in that sense?
You may say that the core version is good enough, but in my opinion is not. So where's the boundary? It's completely user defined.
Yep. So we should ship it but nor promote it. Good to see that these companies/projects care for Fedora and we should praise them for that, but we should not *promote* it. There are so many interesting real open projects in Fedora that definitely deserve to be promoted.
I know, it's a bit like comparing apples and oranges, but what I mean is this: if a feature, as in software, is nice to have, and there are no legal implications, that is, is fully conformant to our current rules and guidelines, I don't see why we should not support it. A vendor has the right to use and modify it's own software and sell it, with what you may consider added values (and thus, a selling point, and more selling means better chances to support the free code base).
Again - the question was on *promoting* these projects/companies. Nut about support or having them in the distro.
IMHO Fedora should use its marketing power to promote Open Solutions. Not comapny driven community stuff.
Jan
On Wed, Apr 14, 2010 at 12:47:02PM +0200, Jan Wildeboer wrote:
On 04/14/2010 12:19 PM, Mario Torre wrote:
You can definitely use the core version, but may you depend of vendor specific plugins to do you job, just think about WebSphere or MyEclipse with support for the Matisse graphics builder in Eclipse.
The core question here stays: Should we *promote* openCore stuff in Fedora Marketing material. Please stay focused.
Do we *promote* Eclipse in that sense?
You may say that the core version is good enough, but in my opinion is not. So where's the boundary? It's completely user defined.
Yep. So we should ship it but nor promote it. Good to see that these companies/projects care for Fedora and we should praise them for that, but we should not *promote* it. There are so many interesting real open projects in Fedora that definitely deserve to be promoted.
I know, it's a bit like comparing apples and oranges, but what I mean is this: if a feature, as in software, is nice to have, and there are no legal implications, that is, is fully conformant to our current rules and guidelines, I don't see why we should not support it. A vendor has the right to use and modify it's own software and sell it, with what you may consider added values (and thus, a selling point, and more selling means better chances to support the free code base).
Again - the question was on *promoting* these projects/companies. Nut about support or having them in the distro.
IMHO Fedora should use its marketing power to promote Open Solutions. Not comapny driven community stuff.
Originally the story behind Zarafa was that one of our volunteer community members worked with an ISV to get appropriate packaging together for the Zarafa code. It's a nice illustration of Fedora as collaborative environment, but perhaps there's less appeal on the purely technical/code side.
The talking points process happened on this list as an open, transparent process, so we should definitely be considering this issue for F13 GA and future releases.
On 04/14/2010 08:26 PM, Paul W. Frields wrote:
The talking points process happened on this list as an open, transparent process, so we should definitely be considering this issue for F13 GA and future releases.
So, do we want to mention Zarafa and IDEA in the Fedora 13 GA announcement?
Rahul
Hello Rahul,
On Wed, 14 Apr 2010, Rahul Sundaram wrote:
Open core, for those not familiar with the term is the business model of keeping some key features closed and selling a proprietary product where the "core" functionality is free and open source. Two features in the feature list are such software
http://fedoraproject.org/wiki/Releases/13/FeatureList
Zarafa groupware and IntelliJ IDEA, IDE for Java have a number of features that is only available in their proprietary product.
I'm sorry to say that, but it looks very much like you never used Zarafa so far nor do you have somehow a deep clue about the features or how it works.
Please explain why you consider the Microsoft Outlook support (a *.DLL MAPI connector), the Active Directory toolkit, support for Blackberry Enterprise Server and the Auto deployment tools as key features? Is Fedora 13 going to support Microsoft Windows clients and Microsoft Outlook? No. And that's why it doesn't matter that Fedora doesn't ship all these features. Because they require and depend on Microsoft Windows, otherwise they won't work.
The High Availability support is only human support for DRBD and Heartbeat; they are both used to get the High Availability support to Zarafa. Nothing you couldn't setup yourself, you're just paying for human support here (see the tooltip).
The remaining three key features are Bricklevel backup, Advanced multi user calendar and Multiserver support. Seriously, I never needed any Bricklevel backup/restore the last three years (where I began with Zarafa), the soft- delete satisfied all user needs so far. For the other two features, I would say, that there needs to be a tiny difference between paying and non-paying people. Especially the Multiserver support doesn't make much sense without having a support contract with Zarafa anyway. Imagine, that something hooks up in your productive environment with 2000 users. Do you seriously want to debug that yourself and on your own? No.
On the other hand, the Zarafa Open Source Collaboration as Fedora is ships it right now, is a full-featured collaboration software 100% based on MAPI and providing a native MAPI implementation for Linux. If you are a Fedora (that means a Linux user), you can send and receive e-mails (via Webaccess, POP3/IMAP), create calendar entries (Webaccess, CalDAV/iCal), create group invitations (Webaccess), create contacts (Webaccess, there's no VCardDAV standard until now). The Zarafa Webaccess is made to be a full replacement for Microsoft Outlook (there are even prominent paying Zarafa customers who only use the Zarafa Webaccess because they like it more than any Outlook). And if you use Z-Push from the RPM Fusion Package Review, you even can sync your mobile device/smartphone - of course it needs to support ActiveSync over-the-air, but there are clients for nearly each smartphone nowadays.
And if you don't believe me, install Zarafa and verify it yourself. Or ask somebody with Zarafa knowledge and experience.
So which key features are you exactly missing in Fedora? Who the fuck needs seriously support for Microsoft products in Fedora? Or do you want to start a "Fedora Windows" spin for Fedora 14? ;-)
Greetings, Robert
On Wed 14 April 2010 1:42:58 pm Robert Scheck wrote:
Who the fuck needs seriously support for Microsoft products in Fedora? Or do you want to
start
a "Fedora Windows" spin for Fedora 14? ;-)
Somebody who uses Fedora (or if this hits RHEL 6, RHEL too) in an enterprise environment where they _have_ to use microsoft products?
Ryan
Hello Ryan,
On Wed, 14 Apr 2010, Ryan Rix wrote:
Somebody who uses Fedora (or if this hits RHEL 6, RHEL too) in an enterprise environment where they _have_ to use microsoft products?
if you go Enterprise, use an Enterprise product please. And no, I'm pretty sure, that Zarafa won't get part of RHEL 6. It will get part of EPEL 6 for sure, but that's something different. Or would you use CentOS 6 for your Enterprise environment? No, you wouldn't, because you want to get support and all the fancy stuff your subscription provides you. Why should things be different for Zarafa? If you want Enterprise, get a subscription - and for RHEL 6, Zarafa will surely provide official supported packages with the proprietary stuff.
Greetings, Robert
On 04/15/2010 02:12 AM, Robert Scheck wrote:
I'm sorry to say that, but it looks very much like you never used Zarafa so far nor do you have somehow a deep clue about the features or how it works.
Not the case.
Please explain why you consider the Microsoft Outlook support (a *.DLL MAPI connector), the Active Directory toolkit, support for Blackberry Enterprise Server and the Auto deployment tools as key features?
Any large scale deployment would very likely use these features. If they were not key features, the organization behind Zarafa wouldn't be holding back these features as proprietary ones. The whole business model depends on enticing users to buy the proprietary version. Now the open source version might be usable by itself for some users and I don't disagree with that but it is certainly worth asking whether this is a model we want to promote.
So which key features are you exactly missing in Fedora? Who the fuck needs seriously support for Microsoft products in Fedora?
Believe it or not, we do spend a lot of time working on interoperability with Windows. Think Samba for instance and Red Hat's investment on openchange. Do watch your language. There is no need for rudeness when the question being asked is simple and straight forward.
Rahul
On Thu, 15 Apr 2010, Rahul Sundaram wrote:
Believe it or not, we do spend a lot of time working on interoperability with Windows. Think Samba for instance and Red Hat's investment on openchange.
Well...the time for interoperability with Microsoft products is spent in the end for Enterprise environments, not for the typical Fedora usecases. Nobody in a real and huge Enterprise environment is using Fedora, they're using RHEL. And they're using RHEL and not CentOS. Fedora is and will be a playground and space for latest, newest, greatest in Open Source Software, but nothing for real Enterprise in mission critical environments. Or with other words: A testbed for RHEL.
Having this said, I'm pretty sure, that only less Fedora users care about whether Zarafa in Fedora has support for Outlook or not, because most of our Fedora users don't use Windows as their main operating system. Maybe you should see the benefit of what Zarafa open sourced so far: There is a full-featured Open Source MAPI implementation and library by Zarafa, while OpenChange provides in comparision only a few less percent of the overall functionalities and features - they're providing basics, but did you ever try something more extended? It doesn't work or it crashes. OpenChange is not usable in Enterprise so far, while Zarafa (even without the proprietary features) is.
By the way, wasn't it you some time ago promoting the Omega Fedora Remix containing closed source software via the Fedora mailing lists?
Greetings, Robert
On 04/15/2010 03:23 AM, Robert Scheck wrote
Having this said, I'm pretty sure, that only less Fedora users care about whether Zarafa in Fedora has support for Outlook or not, because most of our Fedora users don't use Windows as their main operating system.
We really don't know much about our users. They might be using Windows or they might be working on a environment where a number of *other users* are using Windows.
Maybe you should see the benefit of what Zarafa open sourced so far: There is a full-featured Open Source MAPI implementation and library by Zarafa, while OpenChange provides in comparision only a few less percent of the overall functionalities and features - they're providing basics, but did you ever try something more extended? It doesn't work or it crashes. OpenChange is not usable in Enterprise so far, while Zarafa (even without the proprietary features) is.
The whole point of MAPI is to talk to folks using Windows. So you can't really argue that Windows is irrelevant to our discussions. Even in a groupware deployment within Red Hat, blackberry integration was considered a blocker for example. No doubt that open source features are useful by itself but it should be taken into consideration that a large scale deployment very likely would have to be buying the proprietary version and IMO that should factor into our considerations of what we promote within Fedora regardless of whether you consider your work in Fedora as a "test bed". I understand you are associated with Zarafa and worked on integrating it into Fedora but my position on this concern around the model we are implicitly promoting rather than anything against the software itself. The question boils down to one simple thing: Should we promote "open core" software prominently within Fedora? You can take Zarafa completely out of the equation and still answer that meaningfully.
By the way, wasn't it you some time ago promoting the Omega Fedora Remix containing closed source software via the Fedora mailing lists?
Common confusion. There is no "closed source" software in Omega. What I include by default is the some software from the RPM Fusion - free repo which are entirely free and open source software. Being patent encumbered in some regions doesn't make it non-free. Anyway this is off-topic to the current discussion. If you have questions on Omega, feel free to mail me offlist about it and I can discuss any concerns in length.
Rahul
marketing@lists.fedoraproject.org
-
Adam Williamson -
Jan Wildeboer -
Mario Torre -
Paul W. Frields -
Rahul Sundaram -
Robert Scheck -
Ryan Rix -
Yaakov Nemoy