Still even though it didn’t work out, I largely enjoyed the experience and am looking forward to implementing my next “big idea”.

It seems the main beef Mr. Hater has it that there is no common API/ABI across Linux distributions, so proper support and 3rd party applications are lacking. He also seems angry that the community won’t do anything about this. The rest of his blog seems to be various rants against real or perceived problems with Ubuntu (or as he calls it youbuntube).

If Mr. Hater really thinks it’s gcc/glibc/glib/gtk/etc and kernel versions that is holding back 3rd party apps, I think that is just naive. In 2008, who is writing consumer level applications in straight C/C++? Maybe the lack of ported apps has to do something with Windows/Mac developers using APIs not available on Linux at all (e.g. Win32, DirectX, ActiveX, various Core* stuff from Apple). Maybe they are depend on other 3rd party libraries/components that aren’t commerically available/supported either. That’s much more likely. It’s a chicken and egg problem as no one wants to port without a big userbase and people don’t want to switch if their apps won’t work.

Maybe Mr. Hater had some issues with his My Little Pony?

With the trends in virtualization, I don’t see why these problems won’t be solved with VMs in the near future. That is the direction that the market is going. If the application(s) you need must have Windows, then use a Windows VM and run it there. Technology like VMware’s unity is only making it easier. This stuff will filter down to Joe User eventually.

Alternatively, perhaps the average user will stop using a PC and miagrate to dedicated devices for Internet access? There are people looking for that kind of a solution.

Most non-technical people I know use apps like YouTube!, GMail, Facebook, etc. They install them by typing the URL into their web browsers. I can’t really say I know anyone offhand who’s bought any consumer-level commerical software recently.

For support, from what I can see, most ISVs either support RedHat and/or SuSE. Anything else and you are on your own. I think that eventually will consolidate the market and get rid of some of the lamer distro forks. That’s still in progress though.

As for Mr. Hater’s issues with Linux being difficult and non-intuitive compared to Windows or Mac OSX. I think that is kind of offbase too. Linux distributions have come a long way. I remember what things were like before HAL, D-Bus, udev, etc. Sure it’s not perfect, but it’s also not like people have stopped working on making these things better. Look at things like NetworkManager.

X11 was very stagnant with XFree86. It took time, but x.org is moving forward again. They are working on configuration-less X so you don’t need to edit xorg.conf. Their work has enabled other projects like Compiz-Fusion (sure some of those effects are useless eye candy) to push things forward too.

Also “difficult” and “non-intuitive” is rather subjective. I see people all the time who’s attitude towards computers is if there isn’t step-by-step instructions and a 120×120 icon on the desktop to click it’s not easy enough for them. You know these people who “hate computers” and are adverse/hostile to learning even the slightest bit about them. Windows, Mac or Linux isn’t going to make a difference to someone like this.

Someone who’s used Windows or a Mac for years and now switches to Linux is also not going to have an easy learning curve. Can’t argue with that. I think Apple makes switching look easy because they have total control of both the hardware/software. Maybe the solution is for distros to start only supporting certain hardware configs?

Over the years I’ve used Linux, I’ve seen things improve a great deal in regards to useability and functionality. I haven’t any huge (eg. non-working system) problems in a long time over many upgrades. I use it at work and I use it at home. I know I’m not the average user but I know I’m not alone in my experiences.

I will give him that Ubuntu/Fedora/OpenSuSE in their attempt to hide all the system complexity for end-users make it very hard for those same users to figure out what is breaking/wrong and correct it. I think Windows and Mac OSX suffers from this to some extent too.

It’s not all bad, some things Linux/GNOME has gotten right. I remember trying to use a friend’s Macbook last year, I wanted to make a copy of a data CD. I put it into the drive, the icon appeared on the desktop and I right clicked and selected Copy. Mac OSX (Lepoard) promptly created an copy of the disc icon on the desktop. I had to find the burning application and dump/burn from there. With Linux/GNOME, right click on the disc and select copy, “just worked”.

I don’t see any easy answers for making this stuff easier to everyone. That’s probably why you see GNOME guys always talking about who their tarket market/audience is.

At least Mr. Hater seems to have the attention of some of the main developers in GNOME/Varius Distros/Mono/etc. Hopefully the valid points he brings up will be addressed.

Networking

First item to deal with was networking. I recently bought a Linksys WRT54GL router and flashed it with DD-WRT v24. This allowed me to use the router as a wireless bridge to my local network.

My other router, a Linksys WRT350N was already running DD-WRT. Only issue I had was that I couldn’t get any wireless security to work. In the end, I just enabled MAC filtering. Hopefully this gets resolved by a DD-WRT update in the future.

Once the bridge was established, I could connect my Pioneer BDP-95FD Blu-ray player on to the network. I configured it with a static address in the initial setup screen. I was able to ping the player from my other machines and the player was able go on the Internet and download/install the latest firmware.

I was only able to use one device on the bridge though because everything behind it was being masked with the same MAC address. Seems to be a limitation of the bridge itself.

Sharing

My Blu-ray player supports Windows Media Connect and DLNA for media sharing on a network.

Windows Media

Using KVM I loaded a Windows XP Professional virtual machine and installed the Windows Media Connect software on it. It was easy to setup. I told it what I wanted to share, what devices to authorize and started the service.

I turned on the player and it saw the sample photos and sample MP3s on the VM. It didn’t however see my collection. It doesn’t seem to be able to share a mapped network drive. I tried messing with Samba on the Linux box to allow anonymous access, but no go.

Nothing I tried worked expect using the local storage in the VM. I didn’t want to have to copy GBs of MP3s into the VM though. I also didn’t want to have to depend on Windows if I could help it.

DLNA

I found uShare which is a free UPnP A/V & DLNA Media Server for Linux. I’d seen Ubuntu users talking about how they had used this to share media on their XBox’s and PS3’s.

I found an ebuild for Gentoo and installed it on the server that held the MP3 collection.

When I turned on the player, at first it didn’t see anything when I used the DLNA profile in ushare. I switched to the XBox profile but then it said it wasn’t authorized. I didn’t see any way to fix that.

After some more debugging I found that the DLNA profile was the correct one to use. The player could see and use the service. The command line I use to run it as:

ushare -n server_name -i eth0 -c /data/mp3s -D

To debug problems I use:

ushare -n server_name -i eth0 -c /data/mp3s -d -v

Final thoughts

uShare works good and gives me a Linux based DLNA solution. I can run it on my server in the background and it’s always available to the player. If I had to use Windows, I’d need to keep a VM running and duplicate copies of my stuff.

Hopefully in future uShare versions they will improve support for bigger libraries and provide more metadata. I found that Windows Connect would show album covers and read the ID tags off the MP3s. DNLA was more like a file browser.

How does something like that even pass QA?

You can follow the work in my git repository: git clone git://www.lbtechservices.com/lat.git

If you notice any problems, let me know.

The first issue is MonoDevelop. In the 1.3 branch, I started to make more use of MonoDevelop features. I switched from building the interface in Glade to MonoDevelop’s Stetic UI builder.

This made development much easier and I was able to clean up a lot of interface issues. However, I found new MonoDevelop releases wouldn’t compile the project any more. I reported the bug and some things got fixed but overall it stopped development cold.

As of MonoDevelop 0.15 I can’t even open the project anymore. It just crashes try to load the solution file. Even with the stable 1.2 branch, I can open the project but it refuses to compile with some strange errors.

The second issue is Mono, specifically the Novell.Directory.Ldap assembly. It doesn’t support non-Latin characters as far as I can tell. TLS support is broken. There is no support for GSSAPI for Kerberos intergration.

I can’t implement the features I want with the way things currently are. I guess I could try and fix it myself but it took me months to get one SSL bugfix accepted and then I had to wait for the next two Mono releases before I could use it. The TLS bug I reported has been open for over a year now. It doesn’t inspire confidence.

I’m not really interested in re-doing all the UI work in Glade and then switching back to a Gedit/Terminal/Glade workflow. Overall LAT does what I need it for and my itch is basically scratched. So as it stands, I don’t think I will be doing any more releases of LAT past 1.2.3 which I just pushed out today. If someone is interested in taking over the project, let me know.

• BUGFIX [#90] Fix bug with samba SID being stripped during edits
• BUGFIX [#88] Creating new POSIX user causes crash

Downloads: tar.gz

• Building a Mail Gateway (ODT)
• Building a Mail Gateway (PDF)

Comments/suggestions are welcome.