I want to create a Sakis3G package, but I have 2 ways of doing it in mind.
The first is to create a single package that works both for stable and cooking with all the required packages inside it (including linux-dialup 2.6.30, 2.6.34 and 2.6.36). The advantage of this would be that you only needed 1 package you needed to install to get Sakis3G and all it's required packages. The downside is that it might break compatiblity with other other packages in the repository. This will also be harder to do because each package will require certain steps that it's installer must do.
The second is to create 3 different packages that rather require the packages available in the repository, while including the packages and files that does not come with SliTaz or that version. The advantage to this would be that it will keep compatibility with that repository, while providing the 'extras' that the user will need to get Sakis3G to run well. The Disadvantage is that the user will be required to download ALL the other packages by hand. They will also be split into an stable, cooking and latest cooking versions.
I now leave the choice with you guys. I'm willing to package Sakis3G so it will work without weird hacks or much effort on the user's side by using a proper sakis3g.conf to set all the options like using the iterm by default and having the package installer run all the 'by hand' hacks automatically. In the end, using Sakis3G will be simpler and smoother that my previous explaination.
I think, that concentrating on 4.0 is the priority - I think, that most modems are still supported with hal-info in 3.0, but it would get a problem doing nothing for SliTaz 4.0. Gnerally there is the question, if hal is needed any more - because the new xorg is based on udev and needs no more hal
That might be the case, but the older pcmanfm still requires hal. We need to move the completely rewritten beta copy. It atleast has better support for udev systems.
Another thing you probably should know is that sakis3g can run on hal free systems as well. It can successfully run on modern systems like Ubuntu 10.10 and the latest Fedora. It's not limited by an hal-info dependency like SliTaz's 'other' means of 3G configuration, which to be frank is lacking.
Some 3G modems will just refuse to work with a simple option/wvdial method which SliTaz depends on. Further more, there is NO guarantee that that 3G method would work with a brand spanking new 3G Modem. I know it because it did NOT work with mine even WITH a properly configure usb-modeswitch and option configuration. Sakis3G works from the get go and does NOT require hal-info, usb-modeswitch or wvdial to be pre-installed like the other method does.
By me building this package, it will also REMOVE the need to configure usb-modeswitch, wvdial or option by hand. Sakis3G can do it for you. By updating the configuration file I will include, you can even automate this step. Yep, getting online is as simple as typing sakis3g connect or clicking the right icon.
This is why I'm putting this effort in. If our users cannot get online with our distro then they won't use it. They'll also speak badly of it. Sakis3G gives us a quick, simple and easy solution to the problem.
I also do not mind replacing ash with bash. Bash is anyway more powerful than ash will ever be. Along with the proper getent, expr, ps and grep commands, I don't really see a problem here.
We shouldn't just be aiming for small size. We should also be aiming for usablity and ease of use as well.
From a user point of view i completely agree: " ... If our users cannot get online with our distro then they won't use it. They'll also speak badly of it. Sakis3G gives us a quick, simple and easy solution to the problem..." "We shouldn't just be aiming for small size. We should also be aiming for usablity and ease of use as well."
For the same reason i think that using by default a dock applications like docky o avant-windows-navigator could give to the distro a great appeal for getting more success.