[mythtv] MythTV: Isaac Tivo: > 100 tech guys

[Isaac Richards]

On Tuesday 07 December 2004 07:36 pm, Kevin Elliott wrote:
> I would have originally agreed, having had used MythTV for
> many years now, but MythTV has lost a lot of forward progress
> over the last 6+ months. If you review the CVS commit logs, you'll
> notice that much of the commits to the code base are really just
> translation patches to other
> languages.

I suggest that you review them again.  Quite a bit has changed.  Certainly a 
lot of translation work has been done, but, that's to be expected, no?

I've been spending much of my free time in the past several months finishing 
my basement.  The rest of the time I have to spend on Myth is generally spent 
applying patches that people send in.  I really wanted to have been 
re-writing all the UI code to use the mythui library I wrote a while back by 
now, but I simply haven't had time yet.

> IMHO, and Isaac and other will flip a lid when I say this
> (like they usually do), but I attribute much of the development
> "slowdown" to Isaac's personal direction and resistance to
> third party recommendations and design strategies.

You're right - I won't put much stock into someone else tell me how to design 
something, especially when that person has no intention of writing any code. 

But, then again, I don't much recall someone suggesting any sweeping changes 
recently.  Well, maybe the multi-database support, but there was a nice 
compromise worked out, I believe.

> While I 
> agree that keeping a fast codebase is important, Isaac tends to
> prefer execution speed over proper design. 

I would _really_ like to see when I've said that.  The only time I can recall 
was when someone was trying to re-design the music database schema, and did a 
very poor job of it (ie, much, much slower than the current schema for large 
datasets).

> In the long run, it 
> has made things such as "communication between modules"
> and "simultaneous activity in modules" non-existant in MythTV.
> The simple fact that MythTV can not properly play music while
> browsing photos or searching for things to record is proof of
> this stagnation. It was over 16+ months ago when people attempted
> to work with Isaac to solve these sorts of problems, and he rejected
> most ideas in favor of his own needs.

I don't recall rejecting anything of the sort.

If someone were to propose a 'play music anywhere' extension at the moment, 
though, I certailny I would reject it.  Mfd (and mfe) exist in CVS for just 
this purpose, and will be merged in to the main mythtv/mythmusic trees 
whenever Thor is happy with his code.

I don't know what you're talking about with 'communication between modules', 
because such things can easily be taken care of by using existing message 
passing structure that has existed ever since the frontend and backend were 
separated.

> Certainly this is OK, because 
> it is his project. But Isaac has admitted on many occasions that his
> needs are met, and he has no interest in developing features that he
> is not interested in, and that others should. Excellent, that is how
> Open Source projects work, others will write functionality that
> others want. But Isaac (and a few others from his tight-nit group)
> have had no interest or time to assist with the conceptual direction
> in conversation. And those that have attempted to make design
> changes, were simply shunned away after submitting patches,
> deeming the patches unacceptable due to the wrong direction.

Yeah, that's why I'm completely rejecting multi-database support.

> Even more so, developing modules in general are overly complicated
> than what they should be. Even some of the once deemed "bad" Windows
> PVRs are ages ahead in that sense, some having tens of plugins, most of
> which are quite solid.

More people use Windows than Linux.  If I had started this project on Windows, 
I'd have a few orders of magnitude more users and contributors, just like 
those other projects you mention.

> Until the structure of MythTV is redesigned, I don't think there will be
> much growth by contributing developers.

Honestly, what's wrong with the current structure?

Isaac