Tying servers and desktops together (S5E9)

[R.Smith]

(As an (even further) aside, it would be interesting if you could tie this to the .install format that the Nokia tablets use, which automate the addition of new repositories too. If you could have a single install file that not only added repositories, but also picked up the version number/name of the distro and looked for the appropriate package, then you'd be laughing)

This isn't good from a security point of view, mind. It'd be all too easy to compromise a system.

[rmjb]

First of all I love the show and registered on the forum just to reply to this topic.

The specialization of features already happens in distros and it's what differentiates them from each other. Fedora 7 had a really great tool called VirtManager that I'd really love to have on my Ubuntu box, but it is not a straightforward port. A lot of underlying things in Ubuntu had to change for it to work, kind of what Chis was saying about kernel patches and so on. But it's all open source software and if enough people want it on Ubuntu they're free to do the work to add it (and they have https://bugs.launchpad.net/bugs/116897 ).
Similarly all those system-config- programs that RedHat did for itself, a lot of them are really good, but they all depend on how RedHat stores it's files in /etc (the use of the sysconfig directory under there). Is the onus on RedHat to make those tools easier for Debian to use if they wanted? I don't think so, as long as the tools are open source and the option is there.

Abstract that out to a muti-machine interoperability scenario such as client-server applications or even peer to peer. I think once the apps are open source it's up to the users/developers of the other distros to make the decisions to change their distro to suit, patch the app to suit, or work with the app developer to make it more portable... or make a fork...

Anyhow, that's my 2 cents. Keep up the great work gentlemen. And Aq, you all ready my iriver ogg email a couple episodes ago, I posted on LinuxReality's forums since his oggs always work, hope I get a response.

- rmjb

[Aq]

First of all I love the show and registered on the forum just to reply to this topic.

Cool! That's what we're hoping for

Aq, you all ready my iriver ogg email a couple episodes ago, I posted on LinuxReality's forums since his oggs always work, hope I get a response.

We're also experimenting with ogg encoding over in another thread: http://forums.lugradio.org/viewtopic.php?f=1&t=3684 -- your input would be very helpful!

[kit89]

Should distro vendors make their server and desktop products tie more closely together?

I don't think there's a problem for distros to create new features to tie the desktop & servers together. Especially if the source is released for other Distros to use & implement.

However I think the problem is what happens if one distro creates a new feature that does it one way. And another distro creates a similar feature but does it completely different way.

Then the problem comes down too, should you support both or just one? Or do you think they both suck & create your own similar feature.

With that you got fragmentation & more problems than what the feature solved.
(I'm sure that was one of the main downfalls of Unix, many versions and different ways of doing something but none worked together & there were no standards for it.)

[spamfree]

It is absolutely acceptable for any distribution to customize their server and desktop products to the fullest in order to make their use easier, more seamless, and more automated. If Mandriva, for example, want to make their server and desktop integrate using tightly coupled WinBind and mDNS/ZeroConf then they should do so. If it is desired by an individual or another distribution to integrate into that environment then it can be done by configuring your Puppy Linux(another example) distribution in a similar fashion. It's all good!

Indeed, it is the tying together of servers and desktops that make some distributions desirable over others. Novell and Red Hat for instance do just this as a means of distinguishing their products from the vast array of other Linux distributions. However, Novell and others make many of these improvements by using proprietary software that is not available to other distributions and lock in their customers in an effort to maintain their revenue streams. This lock in is what is not acceptable to me and an open source world.

When your server and desktop are tightly integrated with each other through the use of something like eDirectory but, you cannot easily integrate another desktop like Mandriva(another example) because, although eDirectory is distributed freely, some critical module is proprietary and even if you were willing to pay for it you can't get it compiled for the Mandriva platform then that would be an unacceptable tie in because it locks out Mandriva and any other distribution from having the same level of integration.

If the components are open source and available to all then you are not explicitly excluding anyone. Their exclusion becomes a matter of their own choice. However, when you hold back the proprietary bits, then you are explicitly excluding others and that causes me to explicitly exclude your distribution no matter how tantalizing your integration may be.

Taking it a step further, someone on the show mentioned integrating two products by fiddling with the TCP/IP protocol. This particular approach, fiddling with an established protocol, may be suicidal but, I still think it is acceptable if the fiddled bits are open sourced. If it offers something that is truly better or desirable, then everyone else can start using it and the protocol is advanced and improved. If it is not widely accepted then your now non-standard implementation will suffer as people shy away from it.

Any change you want to make to your distribution is completely acceptable. That's precisely why we have so many distributions today. But, thanks to the GPL none of them lock out any of the others because the open source code can be implemented by any and all of them if they so choose.

[TommyMac5]

I understood the question to be twofold; Should a server and desktop solution tie more closely together if it meant a smoother, more elegant experience for the user, and, should a distro be held back in it's development because it needs to insure compatibility with the "lowest common denominator" distro, thereby stymieing creative development.

The first example that came to mind for me was this:

What if Ubuntu developed a server app that would recognize a new PC that had just attached to it via network boot, and offered the user at the other end a one-time workstation remote set up? Imagine how a network admin's life would be if all he needed to do was put a box on the network, boot to the network card, and say "yes" - get a bite to eat, and come back to a fully configured workstation with all the configs set up properly and all the required apps in place?

I know this is way out there, but if it did, wouldn't they have an unfair advantage? It's not that they were predatory, it was they were on a quest to make life easier for everyone - and the result is that no other distro's would have a fighting chance in an environment like that. It's a fast, easy, low cost. low resource solution to a problem. Now, should Ubuntu be *prevented* from building that solution because it shuts out the other distro's?, it doesn't really, it just makes the decision easier for the admin. Any other workstation can still talk to the network, it's just that the admin has the "hassle" of setting up a non-native workstation, versus one that was designed for the system with business parameters already put in place. What would you do?

Is it wrong? I don't know. Is it against community standards and "fighting the good fight"? I don't think so. It's a perplexing problem that really does need to be talked about. But we're talking about advancing the Linux community as a whole, and if a distro has to be held back or shut out of a creative advancement because the others can't do it - doesn't that slow the whole communities advancement?

Just my thoughts.

[Mig21]

That's clearly a good goal, and we all agreed with that in the segment, I think. The question is this: if Red Hat invent some useful bit of software with client and server components, should they put the client on Fedora and the server on RHEL, and then let everyone else do the same if they choose to, or would you suggest that Red Hat should intentionally not do that until all Linux distros can do it at once?

Are you serious? Or is this really poorly phrased

I think all major (and many minor) players in the linux distro business try to do lock-in. If you think Novell, Redhat, and Ubuntu are in one big happy family, you're confused Ubuntu, being not-for-profit, may actually like to standardise things to ease switching between distros, but so far I haven't seen any efforts in this direction from them.

In fact since linux has such a ridiculously low market share, better standards would benefit all players except the ones who refuse to follow them - they will take market share away from windows/osx instead of each other. But I guess they're not smart enough to realize it, or they are just too many to be able to agree on stuff in reasonable time.

[bigredradio]

I just had a chance to listen to S5E9 and found your idea of a "just works" backup server interesting. I work for a company that makes backup software for linux. Not OSS, but I often take my knowledge I gain at work and put it into OSS or forum posts. I was thinking about how to do this and stumbled onto the Ubuntu Home Server project. I do not know if this is what you had in mind when you were talking about servers and clients working together, but this seems to be where the potential for lockin could occur. I was impressed with their concept and plan to contribute to their disk management team. I will try to stay out of backups as a dev. So thanks for the topic. It led me to a project that I think I can offer up some help.