[mythtv] [mythtv-commits] mythtv branch master updated by jpoet. v29-pre-84-g5933e03

Piotr Oniszczuk piotr.oniszczuk at gmail.com
Tue Jun 14 08:20:05 UTC 2016


> Wiadomość napisana przez Roger Siddons <dizygotheca at ntlworld.com> w dniu 14.06.2016, o godz. 00:33:
> 
> On 13/06/16 20:51, Piotr Oniszczuk wrote:
>> 
>> 1.ppl preferring when myth intelligently hints them what to watch
> 
> I'd rather squash that now.
> The current/old watchlist uses (bizarre IMO) rules to predict what the user wants, which just makes it irritating.
> With some exceptions (which can be turned off) my proposal simply orders oldest to newest recording - the "Catch-up TV" scenario.
> The intelligence is limited to selecting which episode of a series to show, but that isn’t going to do anything for movies…

Heh - I agree regarding old watchlist. It was working so distantly to my expectations that I never use it more than 1..2 tests. This is however definitely subjective thing…nevetheless just my experience.

With current implementation I really started to see sense to use it so it is now definite improvement :-)

> 
>> 
>> Next sort case might be sorting by category then by bookmark creation time (for ppl who watching recording in multiple short sessions but they i.e at morning prefering documentary while at evening films).
>> 
>> Maybe we can distinguish major types by prefix:
>> Intelligent: classic
>> Intelligent: LimitedOldest
>> Simple: RecentlyWatched
>> Simple: MostProgressed
>> Simple: RecentlyWatchedCategories
>> 
>> Let me know what You think about this….
>> 
> 
> I see 2 problems;
> 
> 1. Currently the sort settings ("strategy" for watchlist) are all buried elsewhere in the menu, so it's not possible to change easily. As the options increase and become more specialized, the more often you'll want to change it.
Do You mean needed changes in code or rather not easy way to provide ability to change by user in UI?

If it is code - then sure - let’s not reshuffle half of mythtv code.
Maybe we can try do add another, single sort strategy. 
In such case I would choose between „RecentlWatched” or „RecentlyWatchedCategories”….

If it is UI problem - then we can do this in 2 steeps. 
Initially add sort strategies in setup screens. 
Later (when acceptance stabilizes and users will prove their interest) - lets add selecting sort strategy to UI menu.
    
> There are intentions to make them more accessible, I believe, but until that's done, adding more options is of limited use.
> 
> 2. Your scenarios are very valid. But once you go that route, other people want different things and the UI gets over-complicated very quickly. Category is a good example. Do you mean SD genre, EIT genre, EPG category ? (I’m not even sure how many different ones there are)
I would propose: lets use exactly categories used in rec. group filters.
This will IMHO provide consistency in rec. category usage/meaning when dealing with filters/sorts.

 
> 
> Recently Watched is an interesting idea, but I’m wondering if that should include deleted recordings ?
IMHO no.
When user deletes given rec. - then he probably thinks: „no more interesting for me. Delete it”

> 
> We're only talking about sort algorithms here so, longer term, I was wondering about a plugin interface where the sorting could be delegated to Python scriptlets. Then users could choose which ones to install and even write their own. Haven’t had time to experiment though.
Sounds interesting. 
But I think good initial usability/UI design (which we maybe trying do here) may be so efficient that use-cases for so sophisticated feature will unfortunatelly become rare/v.rare.
I think having so talented dev. like You  might be „consumed” in more spectacular(effective)* way :-))))

(*): forgive me selfishness here. 
I’m defining efficiency as: Eff = (all_myth_users_happiness_from_new_myth_features) / (feature_devel_time)
Your gallery work is perfect example of really high Eff :-p





More information about the mythtv-dev mailing list