Custom carpc distro - Page 4

kev000 · 12-24-2006, 03:16 PM

/cheer

I've successfully built the minimyth from scratch (only took 8+hrs to build). I had to go in and change the makefile in the /scripts/utils/file folder to read file-4.18 because 4.17 no longer exists from the download site. I also updated the checksum file. This problem probably doesn't apply to the SVN of minimyth but I made a post on their forum just in case.

Now the questions is... what do I do next? the completed build didn't create a .iso file that i could burn so I don't really know what to do next. Maybe I could copy the built root file system to a usb drive and use syslinux to make it bootable. Nevertheless, I included that question in my post to the minimyth people. Hopefully I'll have more updates soon.

Toe · 01-03-2007, 01:15 PM

Quote: Originally Posted by mccdeuce

What we need for this community is a simple, professional looking, operating software.

I agree with mccdeuce completely. Making an os be the frontend would be better than making another os with a small footprint to run frontends. We all want similar features, so a decent skinning system should be adequate enough replace the desire to have multiple frontend support.

Is this the direction you are planning to go with the distro? I'm not sure I completely understand the goal yet, but I have a feeling it could easily morph into something I'd be willing to put time into.

kev000 · 01-03-2007, 04:46 PM

Quote:

I agree with mccdeuce completely. Making an os be the frontend would be better than making another os with a small footprint to run frontends. We all want similar features, so a decent skinning system should be adequate enough replace the desire to have multiple frontend support.

The idea was (according to my interpretation) just that. This car PC specific OS would boot up to X and whatever frontend was decided to run. It wouldn't have a window manager like gnome or xfce and the capabilities would be limited to what the frontend could do and what the user could do in the command prompt. Minimyth (the OS that was decided to use as the blueprint) was just that. It was an operating system built around running the mythtv frontend. You could even say that mythtv was the OS if you really wanted to super generalize things.

I think monkeybox wanted multiple frontend support because as of right now, IMHO there isn't a real suitable solution. He knew that the OS would be flexible enough that it could support whichever frontend was decided as THE frontend. Blk and myself (others are welcome to help too) are working on a solution for that.

Eubey · 01-09-2007, 08:44 PM

would be cool if we included all the available linux front ends with the distro, and have a settings window where the user could choose which one they wanted.

Geese · 01-09-2007, 11:19 PM

I've read this thread in it's entirety, and I've come to these conclusions:
- There is an abundance of Hardware, Linux, and Programming talent in this forum. This is great.
- There is a common consensus that "we" the mp3car.com Linux forum need a solid platform on which to build solid, usable, free, open carpc solutions.
- We can never agree on the requirements.

Monkeybox, or whomever wants to lead the effort, should decide the supported hardware, distro, software/libraries, and frontend that will be used. If feature X isn't included, it goes on the to-do list. If it doesn't go on the to-do list, contribute or develop your own distro.

After the distro reaches a state of being able to be built repeatably and quickly, the frontend developers can begin to write apps that will run on a known-stable platform.

This natural separation leads me to my final point: we would need two leaders that would work together. One to lead the initial distro creation process and another to lead the frontend process. As long as the dependencies needed by the frontend are in the distro, the two could work independently toward stable releases.

I really wish I could take my original ideas and make them a reality on the distro side, but I can't in the near future. Instead, I only offer my thoughts on how I would approach this from my experience. If the leaders kept each effort focused and on track, it could really make for an interesting carpc experience.

Thanks for reading

monkeyBox · 01-10-2007, 06:36 AM

Unfortunately, due to extenuating life circumstances, I probably won't be able to help out as much as I planned, let alone lead the project. I'm hoping that you all will take this project and run w/ it. Perhaps sometime in the future I can re-join the efforts.

kev000: Thanks for taking the initiative to get minimyth built. The next step would simply be for someone to decide what frontend to use. Personally, I liked the prospects of a flash-based frontend, mostly because it's easy to develop, and it will attract more creative people. Also, I believe the new flash for linux comes with a standalone player. Anyways, the next step is also to just start taking stuff out that we don't need, and adding stuff in that we do need. As far as booting, I've only had experience net-booting minimyth, but I know it can be booted from a flash disk. The minimyth guys are definately a great reference for any questions you have.

kev000 · 01-10-2007, 03:03 PM

Quote:

Unfortunately, due to extenuating life circumstances, I probably won't be able to help out as much as I planned, let alone lead the project. I'm hoping that you all will take this project and run w/ it. Perhaps sometime in the future I can re-join the efforts.

Sad to hear.

I would love to take on this project. I have some time to spare. But as far as using minimyth, I have a few concerns. I propose that we could use a ubuntu based OS. This seems to be the simplest and most user friendly way of doing things. We could use Reconstructor to make a carpc customized livecd/install cd of ubuntu. By stripping the unnecessary packages and by using a fluxbox/xfce/icewm window manager we could get the image <300mb. By adding repositories we can get easy access to awesome stuff after installation: suspend2, etc. everything else is just an apt-get install away.
A minimyth based distro would have a lighter footprint, however it requires a lot more work to get it customized to the point that it is usable in the car. After that it is very limited on what you can do to it after it is installed.
Contributions can be made by creating packages, usplash images, etc.

I would love to hear what everyone here thinks of the proposition. If agreed I can set up a page on the nanonymous webpage and create a forum for it.

sama · 01-10-2007, 03:15 PM

I will gladly offer any help in customizing velocity for the os, should that be desired.

Geese · 01-10-2007, 06:17 PM

I propose that Kev000 and sama lead the distro and frontend, respectively. You both seem to have a good idea of what needs done. Here's to your efforts!

sama · 01-10-2007, 08:11 PM

whilst it's flattering, I don't think I'm quite the right guy for the job

I'm a Java developer and know just enough about Linux to get myself by.

I understand that this distro is destined to run any frontend. I'd certainly be willing to get velocity optimised for the distro (not to include any already included modules etc), but don't think I can lead otehr frontends!

Geese · 01-10-2007, 11:06 PM

That's ok. In fact, once the VM is installed, you're probably set anyway in terms of dependencies. Kev000, what do you say?

kev000 · 01-10-2007, 11:25 PM

I'll do my best. The dependencies for the velocity frontend shouldn't be difficult at all. I'll need some help on some other parts, but we've got a great community here that I'm sure would help out. If what I'm thinking works out it'll support nghost, SW2, and velocity out of the box.

eugenen · 01-11-2007, 06:24 AM

Since there are already so many Linux distros out there, why make another? Why not create a build script for one of the existing distros, it would cut down on a lot of the testing as software is updated, just get the latest version of that distro and run the script to do the custom install.

Dave1216 · 01-11-2007, 06:21 PM

It does seem like a bad idea to fork mini myth, as unless we maintain the core install ourselves we will need to regularly merge changes in the mini myth release.

For this reason I second ubuntu as the distro. I suggest using the 'server' edition (lightweight, no X) with the addition of ratpoison as the window manager. It's the lightest on the market that doesn't really support a mouse (hence ratpoision

). This seems ideal for a carpc.

nasa · 01-13-2007, 11:40 AM

Hey Guys,

Take a look here:

https://www.timekiller.org/carpc/index.php

maybe interesting....

Nasa