Kodi Community Forum
FanArt and Thumbnails Naming-Standard and File-Structure Convension Rationalisation? - Printable Version

+- Kodi Community Forum (https://forum.kodi.tv)
+-- Forum: Discussions (https://forum.kodi.tv/forumdisplay.php?fid=222)
+--- Forum: Feature Requests (https://forum.kodi.tv/forumdisplay.php?fid=9)
+--- Thread: FanArt and Thumbnails Naming-Standard and File-Structure Convension Rationalisation? (/showthread.php?tid=49801)

Pages: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16


- xexe - 2009-06-27

I still say a central repository organized by scraper id. As far as I can tell it solves every problem still being faced with the only down side it wont work in file view mode (but it will work in file view mode if you also have library mode set up).

Based on what I am hearing XBMC is becoming more library orientated anyway.

IMHO its the way to go.


- AnalogKid - 2009-06-27

xexe Wrote:I still say a central repository organized by scraper id. As far as I can tell it solves every problem still being faced with the only down side it wont work in file view mode (but it will work in file view mode if you also have library mode set up).

Based on what I am hearing XBMC is becoming more library orientated anyway.

IMHO its the way to go.

I don't think this is really about file mode vs library, I know for me personally, library mode is the far superior.
This is really more about taking a brand new installation and having XBMC gather all the media and art into the library, preferably without relying in a scraper, and certainly without having to manually configure stuff.
Today, even with a superbly organised library (no art), a scraper will pull art for a movie (but it's often not the choice of art the USER wants, so they have to then manually override the scraper choice).
Also, it's great that XBMC can export it's library, but... the thumbs it exports are low res, not original artwork.

A lot of the additional scraping tools really come from a world (that I support) where folks simply cannot trust IMDB etc to always be around, and so we like to scrape the best quality artwork to our systems and know they are 'safe'.. we have them forever then.
I have seen instances were art that WAS available online has been removed, and potentially lost forever.


- digitalhigh - 2009-06-27

Jezz_X Wrote:Because as much as you AEON fanboys might not like to admit it AEON and djh don't dictate poilcy. Its been called fanart from day one all the sites we scrape it from call it fanart so fanart it shall remain

So I have to ask why ?

Sure its nice to have a bunch of extra images and you can allready do most of this now with hacks and background loading but we don't need to support every variation of a format that someone comes up with.

EG: tvshows the default for many years was Wide Icons (and it still is) djh_ then decided to use posters instead (thats fine) then he changed his mind again and went for 16x9 cropped fanart style with a logo. Where does it end ? how many different types of icons should we force users to fetch just so XBMC can look decent.

Thats the whole point of a "standard" you have the default and thats what should be used. If skins like AEON decide to go against the "standard" then its up to their users to fix what ever so they can use it. Its not the responsability of XBMC to make sure it can fetch every single different Idea that anyone can come up with past future and present. Its the resonsibility of the skins to follow "the standard" and if not thats their issue

P.S everything above is my personal opinion and does not reflect the team in general

As the biggest Aeon Fanboy out there...lol...

You're right. I devised some interesting hacks to get the coverscans and disc images to work in Serenity. But, in order to do so, the files have to be "locally stored". I recently got mad at my HTPC and grounded it to the closet and putting my Xbox back in service. This means that all my "coverart" doesn't work anymore, because I can't map a network drive on an xbox.

So, my hack is now useless for me.

Also, there is no easy way of going <visible>image.exist(IMAGENAME)</visible> so that I can have configurations based on what media images I had for that specific movie. Having this stuff committed to a DB would (theoretically) make it very simple to determine if that image were available. Then you can ensure you have no blank images, you know?

As far as "forcing users to fetch images"...that should really be up to the user. I would really like to see a day where you can have either/or as far as banners/posters. I believe I even made a feature request for this a long time ago, and it was shot dead. This schema would allow for a person to have banners and posters side-by-side and have them available to a display without having to go through and replace everything.

I understand where you're coming from...but with the advent of HD, I think that the current standard is (like AnalogKid implied) starting to become a tad cludgy. It works, yes, but it could work a lot nicer. That's all we're trying to do.

And honestly...if enough users are going against the "standard", can it really be considered standard anymore? I don't think the rules should go out the window entirely...just be expanded a bit.


- tetsuo55 - 2009-06-27

xexe Wrote:This is what we keep coming back to. There really is no such thing as standard naming. What you are proposing is IMDB specific naming for your country. It is to specific to be scalable and is not a good identifier as IMDB could change their policy on (TV),(V) type naming at any moment (if they cared to) breaking this entire scheme. Also MANY user simply will not rename folder or video files. I am one of those users, renaming is the simple way out, there is always a better solution.

I think we need to take a deep breath and re-evaluate the problem and the responses from XBMC devs to the proposed solutions.

I have been here long enough to know the current solution path is not going to be accepted.
I'm used to detection based on hash codes like CRC, this doesn't work for movies though. The only thing left really is a file/folder name.

The way the file name is chosen is something that has to work 99% of the time.

The basic regex we are using now "name" "(date)" is pretty good.
How do you chose the name can be decided through several conventions, the current one being IMDB

Can you explain why any would not want to name the movie accurately?


- xexe - 2009-06-27

AnalogKid Wrote:I don't think this is really about file mode vs library, I know for me personally, library mode is the far superior.
This is really more about taking a brand new installation and having XBMC gather all the media and art into the library, preferably without relying in a scraper, and certainly without having to manually configure stuff.
Today, even with a superbly organised library (no art), a scraper will pull art for a movie (but it's often not the choice of art the USER wants, so they have to then manually override the scraper choice).
Also, it's great that XBMC can export it's library, but... the thumbs it exports are low res, not original artwork.

A lot of the additional scraping tools really come from a world (that I support) where folks simply cannot trust IMDB etc to always be around, and so we like to scrape the best quality artwork to our systems and know they are 'safe'.. we have them forever then.
I have seen instances were art that WAS available online has been removed, and potentially lost forever.

File mode artwork relies on either having one movie per folder and using whatever artwork is there OR using regex to match the artwork to the video name.

However if you use library mode it will, using various mechanisms we dont need to care about, know what the movie is based on the scraper. Once you know what the movie is you can easily imagine a situation where you have your own central repository of images organised by movies ids (id based on whatever your favourite scraper is).

This is better as you do not need to worry about movie identification from filename as part of this project. It is also perfectly scalable and by definition 100% accurate.

The driving goal as I see it is to have human readable, 100% accurate, organised artwork that other tools can manipulate. Absolutely nothing says they have to live beside the video files and it makes alot more sense to have them in one place rather than spread all over the FS.

Once you have the central repository other tools can manuipulate it in any way users like. if they want them beside the videos then other tools can do it. If they want something else something else can do it. Having the data centrally makes alot more sense.

It also covers situations you guys havent even discussed yet, like the fact XBMC is multi user capable via profiles. By the time you add all current and future variables to you artwork filenames its going to be completely silly. I could go on all day finding examples that cause problems like hows this for a legitimate movie name that wont work on fatx:

Night of the Day of the Dawn of the Son of the Bride of the Return of the Terror (1991)

But i predict that someone will say "well how often does that happen?". Well my answer is once is too many.

Or "This is a silly movie no one will have".... Well how about "To Wong Foo Thanks for Everything, Julie Newmar (1995)" nominated for two golden globes.

As i have said before take a step back and look at what I am saying.


- Gangsta - 2009-06-27

I think there was a suggestion a few pages back to handle the example that you mention long filesname on fatx, or any movie names with strange/foreign/system characters in them.

The centralised system that you mention definately has appeal Smile


- AnalogKid - 2009-06-28

xexe Wrote:File mode artwork relies on either having one movie per folder and using whatever artwork is there OR using regex to match the artwork to the video name.

However if you use library mode it will, using various mechanisms we dont need to care about, know what the movie is based on the scraper. Once you know what the movie is you can easily imagine a situation where you have your own central repository of images organised by movies ids (id based on whatever your favourite scraper is).

This is better as you do not need to worry about movie identification from filename as part of this project. It is also perfectly scalable and by definition 100% accurate.

The driving goal as I see it is to have human readable, 100% accurate, organised artwork that other tools can manipulate. Absolutely nothing says they have to live beside the video files and it makes alot more sense to have them in one place rather than spread all over the FS.

Once you have the central repository other tools can manuipulate it in any way users like. if they want them beside the videos then other tools can do it. If they want something else something else can do it. Having the data centrally makes alot more sense.

It also covers situations you guys havent even discussed yet, like the fact XBMC is multi user capable via profiles. By the time you add all current and future variables to you artwork filenames its going to be completely silly. I could go on all day finding examples that cause problems like hows this for a legitimate movie name that wont work on fatx:

Night of the Day of the Dawn of the Son of the Bride of the Return of the Terror (1991)

But i predict that someone will say "well how often does that happen?". Well my answer is once is too many.

Or "This is a silly movie no one will have".... Well how about "To Wong Foo Thanks for Everything, Julie Newmar (1995)" nominated for two golden globes.

As i have said before take a step back and look at what I am saying.


Honestly, I believe folks are considering this and taking a step backward... it's not like it's not been documented about long filename issues (and we already have solutions for this).
I think the issue many folks have with tying artwork to scraper sources / ID's is the following:

Media and Media artwork have a clear relationship, typically a one to one relationship, and if being clever a many to many relationship.
The relationship of artwork to a scraper source is far more tenuous. And I believe MOST folks see a scraper as a means to an end. A means to obtain the artwork, that's it.
Scrapers and scraper source are transient in nature and out of our control, and so linking to them seems risky (I've seen artwork removed / renamed on scraper sources).

I absolutely agree that any solution cannot be based on 'good fortune', but must be based on fail safe (at least) and hopefully infallable logic).
I think however, that you're glossing over a few existing facts today - that scraper logic is fallible, and that XBMC artwork detection logic is fallible.

As for profiles, it's irrelevant. This schema doesn't compromise profiles one jot, it's absolutely no different than the current implementation in that regard. The logic of determining which art belongs to which media is entirely divorced from user profiles.

Again I will reiterate, the schema has nothing at all to do with filemode, it's about library mode and how XBMC can locate the artwork. It's absolutely a 'small' thing in many ways.

We are not looking to find the movie name from the filename, we are stating that XBMC itself does precisely that.

XBMC does the following:
1) Scan for media files
2) Deduce the 'assumed' movie name from either the filename OR from the folder name
3) Offer the 'assumed' movie name to the scraper, and either find an exact match, or allow the user to choose the correct movie
4) Stores the correct movie name in the database (or the assumed one if the scraper failed)
5) Seek out any associated artwork on the local FS based on the FILENAME (or a few coded files e.g. folder.jpg, movie.tbn etc)
6) If no artwork is found, download default artwork from the scraper instead.
7) Create an optimised cached version of the artwork in the cache,and store a link to it in the database

All we are now doing is discussing step 5) and indirectly, step 7)

Now to step 5)
Step 5 works reasonably well today IF the user obeys some rules and some of those rules are naming rules. So the user is FORCED to rename files or it doesn't work. This is a fact.
Today "Movie.avi" will simply NOT work with any scraper in the world. The user has to rename that movie, OR store in a folder using the movie name.

You're also massively overlooking Music...
Today, if you have 1 compilation album in your collection, with say Madonna as an artist on one track. It's impossible to have a local thumb for Madonna.
The only way you can do this is a) Forced to scrape or b) Have a Madonna track on Madonna album

Regarding step 7)
It's entirely up to XBMC if it wants to cache artwork... but we are mindful of the fact that if more artwork is available, XBMC must choose wisely which artwork to cache.


- xexe - 2009-06-28

If profiles are irrelevant how do I add 3 new users on the same network after I have set up all my own artwork? Do you propose that user2 has to rename all my artwork or demands that I do it so they can add their artwork... or do they simply replace mine?

You say "Movie.avi" will simply NOT work with any scraper in the world. Actually all you need is Movie.nfo and it does work. I have these simple URL nfos for every movie and tvshow. My video files and folder names often (but not always) bear no decodable relation to themovie name. As it simply does not matter to me using library mode are you suggesting i rename and rescan my entire collection to comply with this naming or that the scehme is just not for people like me? What about scene users that simply wont rename files or folders ever since its against scene rules. You see my point.

I have little XBMC music experience so I cannot comment on that.


I am not trying to be a PITA but I am just one dude, wait til 10,000 more users find out hey have to change how they organize to use this feature. We better have covered all the bases covered justifying why they have to do all this manual renaming before that point or there will be IRC hell to pay Tongue Especially since XBMC quite happily handles how they have their stuff named this now.


- Gangsta - 2009-06-28

I think that were starting to go off at a tangent here.

From what I understand of the discussion so far what <moviename> is, is really rather unimportant - as long as the artwork has the same <moviename>

Its the artwork files on your system that would need to be renamed to the new schema.

From this you can see that it wont affect scene filenames atall - if they call the movie

TMNT.CAM.HD.VIVIDUX.avi or whatever.

The important thing would be that if they had a poster for it that it be called TMNT.CAM.HD.VIVIDUX.avi.movie.poster (or whatever gets agreed on)

The only reason that the actual movie name was brought up was in relation to the length of the names (and the need for abbriviation of the schema) and a brief discussion about having a nicely organised library.

I reiterate - you do NOT have to rename your movies, pictures, music, tvshows for this to work.

You would only be renaming your artwork that goes with the movie, which lets be honest - very few folk would have downloaded it manually, named it manually etc. It was most likely done by a scraper.

AND as has been mentioned since the start of the topic, a renaming script (FOR THE ARTWORK) would be included, that would rename <moviename>.tbn to <moviename>.movie.thumbnail etc

Hope that clarifies things a bit (or at least shows how I see things)


- xexe - 2009-06-28

In that case you need to make it a requirement that EVERY app can handle and read each users system stacking regex and user addon stacking regex realtime).

You also have to teach user how to use stacking regex, for most it is magic.


You also have to deal with cross FS incompatibilities

You also have to accept that there may be duplicate artwork names for different movies across the FS so to backup your artwork you need to have a backup structure that takes into account the original path of the file.

Those are fairly tricky thing to do (but not impossible)

I know this for a fact having been helping users with REGEX via IRC for more than a year now. Your relying on one of the hardest XBMC config features for this to work. That is doable but its going to be hard work I can assure you.

Edit: Been thinking about it since I posted. I was quite interested in this project when there was a chance it was going to end in globally unique and individually identifiable artwork image names (i.e. user could centrally store and share artwork removing the scraping burden and allowing for hand chosen artwork packs to be created etc). As the current consensus is just another on disk user specific naming scheme it doesn't really interest me so much. I likely wont be posting much on it again. Good luck Smile


- Gangsta - 2009-06-28

xexe Wrote:Edit: Been thinking about it since I posted. I was quite interested in this project when there was a chance it was going to end in globally unique and individually identifiable artwork image names (i.e. user could centrally store and share artwork removing the scraping burden and allowing for hand chosen artwork packs to be created etc). As the current consensus is just another on disk user specific naming scheme it doesn't really interest me so much. I likely wont be posting much on it again. Good luck Smile

Hi again xexe, I do hope you come back and read this, as it seems that we are only in disagreement over the filename issue. (eg the name of the actual movie, as opposed to the name of the artwork - which is whats really up for discussion here) and your input is invaluable.

The idea of 'media packs' for a film sounds good, the ability to post up a single archive that has been carefully crafted for a particular movie.

By media pack (or artwork pack), however we describe it, I mean a set of posters, thumbs, fanart, etc, etc that have been carefully selected to contain similar style elements, colours and the likes for a single film, that will be appealing to look at (when displayed together in a skin). As opposed to having to 'fix' all the stuff from the scrapers when they just download the first available images from each site.

Just an idea, but perhaps the archive could contain an xml, defining what movie the pack is aimed at this could be the IMDB, TMDB, thetvdb, or all of the above. Somekind of simple loader script, possibly a python script running inside XBMC looks through the archives available - and build a list of these ID's. Next it looks at the XBMC database to see if you have any of the films, if you do, then it copies the contents of the archive out (into a folder you specify) and adds the filename *That you already have* to the start of it.

Sorry im really crap at describing things. But i'll try an example.

Media Pack:

Code:
I Robot.mediapack.zip
--mediainfo.xml
--.movie.poster.landscape
--.movie.poster.portrait
--.movie.fanart.HiRes
--.movie.fanart1.HiRes
--.movie.fanart.LoRes
--.movie.widebanner
--.movie.widebanner1
......
and so on

the XML would contain something like
Code:
<mediatype>movie</mediatype>
<mediaids>
  <mediaid>
    <source>IMDB</source>
    <sourceid>tt7846383</sourceid>
  </mediaid>
  <mediaid>
    <source>AllMovie</source>
    <sourceid>286093</sourceid>
  </mediaid>
...........
etc

The important thing is that the name of the archive is totally IRRELEVANT, it is merely a placeholder for a human readable name so that you can find and download packs easily - IT DOES NOT need to match the name of the media file on your hard disc. The script takes care of matching the archive to your media file.

In the example above, say your movie filename was

I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi

then once you had run the script, it would create the files

Code:
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.poster.landscape
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.poster.portrait
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart.HiRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart1.HiRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart.LoRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.widebanner
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.widebanner1
......
and so on

therefore packs could be created on one machine and easily deployed to another.


- xexe - 2009-06-28

It is an interesting way of looking at it but I still think having semi unique ids as the pack names i.e. the zip name is not the way forward.

For instance choosing English immediately makes them not human readable for a large percentage of the XBMC user base. I stand by my opinion that going from a unique scrapeer id to the human readable movie name is a task that could be easily acomplished by any application.

The last thing we want is a zip file containing the same potential content named differrntly all over the palce i.e.

La guerra de las galaxias .zip
La guerre des étoiles .zip
Stjärnornas krig .zip
A Guerra das Estrelas .zip
Adventures of the Starkiller: Episode 1 - The Star Wars .zip
Csillagok háborúja .zip
Guerra nas Estrelas .zip
Guerre stellari: Episodio IV - Una nuova speranza .zip
Gwiezdne wojny .zip
Hvezdné války .zip
Hviezdne vojny .zip
Hviezdne vojny IV - Nová nádej .zip
Krieg der Sterne .zip
Krieg der Sterne - Episode IV: Eine neue Hoffnung .zip
La guerra de las estrellas .zip
La guerra de les galàxies .zip
La guerra de les galàxies: Una nova esperança .zip
La guerre des étoiles - Un nouvel éspoir .zip
Ratovi zvezda - Nova nada .zip
Ratovi zvijezda .zip
Star Wars IV: A New Hope .zip
Star Wars: Episode IV - A New Hope .zip
Star Wars: Episode IV - Neue Hoffnung .zip
Star Wars: Epizoda IV - Nová nadeje .zip
Star wars IV - Tähtien sota: Uusi toivo .zip
Star wars: Épisode IV, la guerre des étoiles .zip
Star wars: Episodio IV - Una nueva esperanza .zip
Stjörnustríð .zip
Stjernekrigen .zip
Sutâ wôzu: Aratanaru kibou .zip
Tähtien sota .zip
The Adventures of Luke Starkiller as Taken from the 'Journal of the Whills': Saga I - Star Wars .zip
The Star Wars .zip
The Star Wars: From the Adventures of Luke Starkiller .zip
Yildiz savaslari .zip

and thats just the offical Star Wars list. In theory you could have another 6000 language translations of the same filename.


- AnalogKid - 2009-06-28

xexe Wrote:If profiles are irrelevant how do I add 3 new users on the same network after I have set up all my own artwork? Do you propose that user2 has to rename all my artwork or demands that I do it so they can add their artwork... or do they simply replace mine?
This proposal is NO different than using folder.jpg, it's just using a different name, and we have more options.
However XBMC implements profiles will remain untouched by this proposal.
If I'd said "lets rename folder.jpg to folder2.jpg" there wouldn't be a fuss... so I fail to see how renaming that file to anything else can affect profiles.

Quote:You say "Movie.avi" will simply NOT work with any scraper in the world. Actually all you need is Movie.nfo and it does work. I have these simple URL nfos for every movie and tvshow. My video files and folder names often (but not always) bear no decodable relation to themovie name. As it simply does not matter to me using library mode are you suggesting i rename and rescan my entire collection to comply with this naming or that the scehme is just not for people like me? What about scene users that simply wont rename files or folders ever since its against scene rules. You see my point.
I am correct "Movie.avi" will NOT work with any scraper.
You are indeed correct that if the user creates an NFO, they can resolve this... but this is ADDITIONAL work the user must do. In fact, most users do not create NFO files, they download them, based on the correct identification of the movie.

We have NEVER suggested users must rename their existing media files. We have highlighted that in come cases, a user MAY rename a file to overcome a conflict, but in parallel we've got other solutions that don't require that.

You have argued against yourself a number of times here, in that you have firstly required the user do not do any additional work in terms of renaming a file, and yet, you are happy for them to have to create an NFO which as much more complex for them to do.

Quote:I have little XBMC music experience so I cannot comment on that.


I am not trying to be a PITA but I am just one dude, wait til 10,000 more users find out hey have to change how they organize to use this feature. We better have covered all the bases covered justifying why they have to do all this manual renaming before that point or there will be IRC hell to pay Tongue Especially since XBMC quite happily handles how they have their stuff named this now.

It's good that you are! we need counter argument... it helps us fix issues.
However, I think you're mistaken in thinking "it all works today", it absolutely does not. You cannot just throw any unorganised media at XBMC and expect it to work. XBMC makes the user do some things...
a) You have to store your movies in folders (where the movie name is the folder name)
b) OR you must name your movie files with the name of the movie in flat folders
c) OR you must MANUALLY create an NFO file

You cannot use "the user has to do something" as an argument against this scheme when already the user has to do something in the current implementation!


- AnalogKid - 2009-06-28

Gangsta Wrote:Hi again xexe, I do hope you come back and read this, as it seems that we are only in disagreement over the filename issue. (eg the name of the actual movie, as opposed to the name of the artwork - which is whats really up for discussion here) and your input is invaluable.

The idea of 'media packs' for a film sounds good, the ability to post up a single archive that has been carefully crafted for a particular movie.

By media pack (or artwork pack), however we describe it, I mean a set of posters, thumbs, fanart, etc, etc that have been carefully selected to contain similar style elements, colours and the likes for a single film, that will be appealing to look at (when displayed together in a skin). As opposed to having to 'fix' all the stuff from the scrapers when they just download the first available images from each site.

Just an idea, but perhaps the archive could contain an xml, defining what movie the pack is aimed at this could be the IMDB, TMDB, thetvdb, or all of the above. Somekind of simple loader script, possibly a python script running inside XBMC looks through the archives available - and build a list of these ID's. Next it looks at the XBMC database to see if you have any of the films, if you do, then it copies the contents of the archive out (into a folder you specify) and adds the filename *That you already have* to the start of it.

Sorry im really crap at describing things. But i'll try an example.

Media Pack:

Code:
I Robot.mediapack.zip
--mediainfo.xml
--.movie.poster.landscape
--.movie.poster.portrait
--.movie.fanart.HiRes
--.movie.fanart1.HiRes
--.movie.fanart.LoRes
--.movie.widebanner
--.movie.widebanner1
......
and so on

the XML would contain something like
Code:
<mediatype>movie</mediatype>
<mediaids>
  <mediaid>
    <source>IMDB</source>
    <sourceid>tt7846383</sourceid>
  </mediaid>
  <mediaid>
    <source>AllMovie</source>
    <sourceid>286093</sourceid>
  </mediaid>
...........
etc

The important thing is that the name of the archive is totally IRRELEVANT, it is merely a placeholder for a human readable name so that you can find and download packs easily - IT DOES NOT need to match the name of the media file on your hard disc. The script takes care of matching the archive to your media file.

In the example above, say your movie filename was

I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi

then once you had run the script, it would create the files

Code:
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.poster.landscape
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.poster.portrait
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart.HiRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart1.HiRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.fanart.LoRes
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.widebanner
I,-robot-[2004]-ws-xvid-dvdrip-ac3.avi.movie.widebanner1
......
and so on

therefore packs could be created on one machine and easily deployed to another.

I like the idea of a 'pack'.

To my mind, this would simply roll up a bunch of artwork into a zip, with the possible option that the contained artwork wouldn't HAVE to contain the moviename / ID, this could be inherited from the archive name, or even from some embedded manifest file?

For the moment, I'd like to put this in the back burner, and come back to it later, I think it has some great merits.


Back to basics - AnalogKid - 2009-06-28

OK, let's take a step back and list the initial rational for all of this:



1) Today, XBMC allows users to use a number of ways of specifying artwork:
folder.jpg, movie.jpg, <movie>.jpg, movie.tbn etc etc
Some of these can ONLY work in folder structures, others will work in either.

This sounds mega flexible, but end users continuously struggle with getting art to work properly, and it's hard to give a response sometimes because so much depends on how they have their media set up.

2) An initial solution to the problem could have been to abandon ALL the variations, and simply mandate <movie>.jpg since this works in folder and flat systems.
This would remove ALL ambiguity about artwork, for sure! ;-)

3) However, since we knew that there was also a need for additional art types beyond just a fanart and a thumb, we may as well fix that too.... so we introduced a name scheme for more artwork types that was VERY easy to read and left no doubt about that the art did. Good so far right?

4) We could have stuck with <movie>.fanart.xxx.xx etc and insisted that the artwork be placed side by side with the actual media file as is required today!

ALL of the above has no effect on profiles, scapers etc etc... it's a relatively minor change.

5) The next step was "well, does the artwork HAVE to be side by side with the media?"
At this point we said... actually, it would be cool if it didn't have to be, this would make scanning the artwork faster, and could have some caching uses too (the user just copies all the artwork to the local machine for optimal speed).... basically nice if it wasn't necessary to be side by side.

It is at this point (5) where issues arise... where a user has a folder based movie structure and we are trying to represent it in a flat way, so that XBMC can tie a media file to it's associated artwork.
It really isn't a million miles away from what already happens... we just need to teach XBMC how to find the art in a new way.




Note:
At this point, we don't give a damn about additional scraping tools outside of XBMC, it is just convenient for them to finally have some naming consistency. But this is not a GOAL for us.
We also are not assuming anything about 'inbuilt' scrapers either, nor are we relying on them (some XBMC user don't have net connectivity)
Finally, we don't care about skins either.... but we are aware that many new skins are highlighting artwork limitations in XBMC.
Oh, and it's not a solution for file mode vs library mode, but it's primarily used when media and art are imported INTO the library.

The only goals are:
To make artwork naming TOTALLY unambiguous for end users, and to support new artwork types that many new skins are using.
To find a solution that is flexible enough for future artwork types
To NEVER favour any skin or platform.... it's an agnostic solution


Are we agreed on this so far?