Feature Wishlist

From MythTV Official Wiki
Revision as of 21:49, 10 July 2012 by Parent55 (talk | contribs) (Platforms)

Jump to: navigation, search

Important.png Note: Before making any feature requests to the MythTV developers, one first needs to understand the most basic truth about MythTV: "MythTV is a project by developers, for developers." If you look at things in that light, comments that get made by developers to users who submit feature requests ("Sounds good; I look forward to your patch") make a whole lot more sense.

The developers of MythTV work for free (obviously), in their spare time. Most/all of them write software for a living, where they work all day long on things that other people want them to work on. When they work on Myth they focus on what's important to them. A feature gets implemented because a developer wants it bad enough to spend his spare time writing it and testing it, and believes in it strongly enough to defend it from the other developers (which helps to avoid the feature creep common in some projects). Bugs, especially crash bugs, get worked on by all of the devs as they encounter them as those impact everyone.

That's not to say that the users don't matter, or that the developers never implement something that comes from a user. It's just that unless a developer says either "why didn't I think of that" or "I could knock that out in a couple hours" it will be a much lower priority.


MythTV logo square.png Join us making your favorite media center even better than it already is today.

Requests from people who have contributed back to the project in some way carry a LOT more weight. Developers by and large tend to be rather blunt. People often mistake being terse and to the point for being insulting. Then they start a flame war on the mailing list (which is pretty much certain death for a feature request) all because a developer either a) didn't take 3 paragraphs to tell them "I'm not going to work on this" or b) they think that the developer should drop whatever they're doing because *they* want it done. A lot of the devs are a bit defensive when it comes to requests, largely due to previous bad experiences.

And, finally, remember this: if there's something you want badly enough, you can always offer to pay someone to do it.

Editor's note: Even moreso than on most pages, if you add a note here that a wishlist item is now available or in process, please note which version that information applies to ('added in 0.19', for example). Remember: wiki pages live forever. --Baylink 17:00, 12 February 2006 (UTC)

How to make a suggestion

If you've got an idea for a feature that you'd like to see implemented here's some guidelines for submitting it:

  • Before anything else, check the Current Version release notes to make sure it's not already in there (yes, this happens).
  • Clearly indicate that it's a request, not a demand.
  • Understand that code speaks louder than words. If you can, write a patch and submit it to TRAC
  • Be very clear with how you envision your idea working. The more details you have in your request the better chance you have hitting that magical "knock it out in a couple of hours" mark.
  • Make sure you're not repeating a previous request. (Search the wiki and mailing list archives)
  • Do NOT be offended if a developer responds with "sounds good; I look forward to your patch".
  • Requests from people who have contributed back to the project in some way carry a LOT more weight.

When you're ready to submit your suggestion, add it to the end of the relevant page and section listed in "The Wishlists" section below.


The Wishlists

  • Setup and Configuration -- Anything for the setup process including hardware support. These are things you want mythtv to do prior to actually running mythtv.
  • Backend Addons -- Backend additions are things you want mythtv to do when you're not around.
  • Frontend Addons -- Frontend additions are things you want to do in mythtv while using recordings.
  • Plugin Addons -- Plugin additions are things you want to do in mythtv while not using recordings.
  • New Plugins -- New plugins are new things you want to do in mythtv while not using recordings.

Platforms

  • MythTV Frontend running as a Wii Channel (There is a small bounty for this on FOSS Factory. You can contribute to the bounty if you want.)
  • MythTV Frontend (TV client only) running natively on GlobalScale Linux D2 Plug ARM Marvell PXA510 – 800MHz see http://www.globalscaletechnologies.com/c-8-d2plug.aspx - ** UPDATE June '12 ** New D2 Plug availability occurred in October 2011. We have received product and are now implementing the platform. Some delays due to other work. We are having good and bad results. We can compile code on the D2 plug, we are having problems with decoder (OpenMAX is not supported on ARM cpu) we are looking at the options. We still think that it is doable. Send suggestions to mythtv@quickcycle.com
  • MythTV Frontend running on AppleTV (doable: see Installing MythTV on an AppleTV and a thread on the developer mailing list)
  • MythTV Frontend (TV plug-in only) running natively on AppleTV i.e. using Apple's FrontRow for menus. See AwkwardTV Wiki and BackRow Developers' Kit. This would create a single, modeless AppleTV front end / user interface.
  • MythTV Frontend on Windows Mobile. Since a frontend is working on Windows, a Windows Mobile version could be promising, One challenge might be a touch interface.
  • MythTV Backend ported to ARM SoC i.e. NVIDIA's Ka-el and/or Denver projects. While it is unknown if graphics accelerated drivers (VDPAU) will be made available to run a frontend the raw horsepower should be sufficient for a backend with storage, comm flag, and USB based tuners. Imagine your 35W TDP Core2 Duo (T7200) CPU replaced by a 1W TDP CPU/GPU that has vastly less performance...

Platforms are the hardware and operating system setups that actually run the Myth applications. This section should track the request to support platforms other than the default x86 PC environment. This would include console ports like X-Box and PS3.

Will Not Implement

These are suggestions which for one reason or another will not be implemented in MythTV.

MythTV is designed as a PVR, which by definition necessitates that all TV be recorded and buffered. As mentioned, the frontend/backend split was designed with this requirement in mind, and MythTV would require a significant rework of the livetv code in order to reduce this buffer.
fast channel changing is also an unachievable goal due to hardware constraints. Tuners and cable boxes take time to lock, mpeg encoders need to fill video buffers, digital tuners need to wait for stream information and then the first I-frame. All of this adds up to MythTV's internal lag becoming a less significant portion of the problem.
  • Peer-to-peer sharing of shows with friends who also use MythTV
MythTV is designed for use with commercial TV. Fair use precedence allows you to record this for your own personal use, but does not authorized you to share or redistribute in any fashion. MythTV will never implement a feature whose sole purpose is to facilitate this, or any feature which would primarily be used for this.
For those using package managers, your distro should provide this service automatically.
  • Migrate to use VLC, or gstreamer, or some other codec library.
MythTV was written from the start to be based off of ffmpeg. Any move away from that would require vast changes throughout MythTV, and is not likely to ever happen.
  • I demand MythWeb be rewritten in Python. Because that's just the way open source works.
This was actually emailed to a developer. It's a prime example of when to go back and try it again. Demanding things will never endear you and your cause to someone whom you expect to do your work for free.
  • MythTV Frontend and backend runnning on a Casio fx7000G (1985) pocket calculator
This platform is too limited to support a frontend or backend.
While running a backend on Playstation 3 is doable, the hardware is too limited to run the standard frontend satisfactorily. It is possible to use MythTV's UPnP support to play back most MPEG-2 recordings on the PS3.
  • MythTV as a replacement for DVR/PVR interfaces, such as the Philips DVDR3455H/37. The Rockbox.org is a UI replacement project for music players, and MythTV could do the same for DVRs.
The hardware on commercial DVR's is too limited to run MythTV. While these are getting more powerful every few years there is a second problem in that many of these devices use hardware without Linux drivers or run only signed binaries.
  • Have MythTV setup and define network interface
The network interface is outside of MythTV's purview. Most Linux distro's and other platforms MythTV runs on provide decent tools for doing simple network configs automatically or through a GUI and provide command line tools for more complex setups (ifconfig/ipconfig).
  • mythbackend (Ubuntu x86_64 v0.20.20070821-1) appears to catch signals (all, most?) and will exit with return code 0 even when core dumping. I suggest to reserve exit code 0 for a deliberate, graceful shutdown, and instead using a non-zero exit code for other, non-graceful shutdowns and crashes. This allows the user, or any process monitoring mythbackend (in my case I'm Using_pcsk_to_Supervise_mythbackend) to distinguish between the two situations.
Please make sure you are checking the 8th bit of the exit code, Linux normally sets that to one and puts the signal in the low order bits. MythTV does not catch the segfault signal so the operating system is setting the return value.
  • Optionally disable the sync loop on the backend; comments say it's for NFS.. on a combined front end + back end this short-periodic sync can wreak havoc with file allocation, and cause excess fragmentation.
The comments are out of date, this is required on Linux with the CFQ and AS elevator algorithms. Since CFQ is the default on most Linux distro's the sync will stay for the time being. The need to disable it is too low to justify another setting. Commenting it out is fairly simple to do if your one of the few mythtv users changing elevator algorithms for greater backend performance. NFS does not require the sync on any modern Linux kernel.
  • Support for USB-UIRT - IR blaster
This is a job for LIRC.