2008-08-28, 23:34
(sorry, again my answer is probably too long 
)
I do not understand what you mean... <genre> already is a multiple instance field! like
<genre>Action</genre><genre>Comedy</genre>
Yes, that is what I was saying, I intend to do it when the "scraper for dummies" wiki entry is finished, maybe next week.
The particular reasons I was giving were only to justify that file mode has in specific configurations some advantages over library mode; if I ask if it is possible to have metadata or fanart in file mode the response is not "use library mode" like if I just where lazy or ignorant. In fact I'm seeing that many people prefers file mode also but misses the metadata etc like I do.
It is not that I do not like it, it is that, simply, some information from different sites should not be mixed; for example, if you sort by rating and a site rates 1 to 5 and another 0 to 10, the sorting has no meaning. What's the solution? using the same scraper always? not possible nor even desirable. Creating thousands of nfo files and then editing all by hand? simply overwhelming.
One solution to that is to categorize things (more or less automatically) and then you can compare apples to apples and microsofts to microsofts
right now the "categorization" is that the files are in the same directory, but since metadata only exist in library mode, that categorization is lost or the metadata is lost.
For reasons like that is why many people suggest emulating library mode under file mode with nfo files, probably a good solution, but I think I can do the opposite, emulate file mode under library mode, and so *now*, not after atlantis, have the cake and eat the cake (or like we say in spanish, swim and keep the clothes). I intend to do it using <genre> field as tag (instead of the actual genres scraped from internet) populating them automatically at scraper level. A much better and elegant solution would be to have something like that supported by XBMC but that would only be possible after atlantis is out, and even then all must be designed, programmed and actually implemented... maybe a Christmas present
Well, that's one of many reasons why I do not like other media systems! most if not all are even less flexible than XBMC's library mode.
The turning off the servers was just an example, is simply to show the need to inform the user that data may be unavailable, that is usually clear in file mode! but when it *is* available, is good to have the metadata in the library. Even the wiki sees that need and suggest using fake files to maintain offline media into the library... I find that solution a real, big, enormous chapuza (do not know how to translate that, in spanish is very colorful... maybe "botched job"? it means a somewhat working but really inelegant solution)
Yes! that is exactly what I was proposing in the previous message, and what I intend to simulate using <genre>. But having both <genre> and <tag> would be much better... but that could only be after atlantis, of course
)kraqh3d Wrote:@pko
i agree that the nfo xml should support multiple <genre> tags for simplicity.
I do not understand what you mean... <genre> already is a multiple instance field! like
<genre>Action</genre><genre>Comedy</genre>
kraqh3d Wrote:If you dont like how something is done, and you feel you can do it better, in a manner that suits you, then I welcome you to do so.
Yes, that is what I was saying, I intend to do it when the "scraper for dummies" wiki entry is finished, maybe next week.
kraqh3d Wrote:and to respond to your original comments...
The particular reasons I was giving were only to justify that file mode has in specific configurations some advantages over library mode; if I ask if it is possible to have metadata or fanart in file mode the response is not "use library mode" like if I just where lazy or ignorant. In fact I'm seeing that many people prefers file mode also but misses the metadata etc like I do.
kraqh3d Wrote:1) if you dont like the scraped data, use nfo files to categorize things however you want. you can use whatever genre strings you want. there is no limit.
It is not that I do not like it, it is that, simply, some information from different sites should not be mixed; for example, if you sort by rating and a site rates 1 to 5 and another 0 to 10, the sorting has no meaning. What's the solution? using the same scraper always? not possible nor even desirable. Creating thousands of nfo files and then editing all by hand? simply overwhelming.
One solution to that is to categorize things (more or less automatically) and then you can compare apples to apples and microsofts to microsofts
right now the "categorization" is that the files are in the same directory, but since metadata only exist in library mode, that categorization is lost or the metadata is lost.For reasons like that is why many people suggest emulating library mode under file mode with nfo files, probably a good solution, but I think I can do the opposite, emulate file mode under library mode, and so *now*, not after atlantis, have the cake and eat the cake (or like we say in spanish, swim and keep the clothes). I intend to do it using <genre> field as tag (instead of the actual genres scraped from internet) populating them automatically at scraper level. A much better and elegant solution would be to have something like that supported by XBMC but that would only be possible after atlantis is out, and even then all must be designed, programmed and actually implemented... maybe a Christmas present
kraqh3d Wrote:the point of the library is to consolidate all your media regardless of where its located. upnp servers so the exact same thing. they do not discriminate. and honestly, its your choose to leave the fileservers offline.
Well, that's one of many reasons why I do not like other media systems! most if not all are even less flexible than XBMC's library mode.
The turning off the servers was just an example, is simply to show the need to inform the user that data may be unavailable, that is usually clear in file mode! but when it *is* available, is good to have the metadata in the library. Even the wiki sees that need and suggest using fake files to maintain offline media into the library... I find that solution a real, big, enormous chapuza (do not know how to translate that, in spanish is very colorful... maybe "botched job"? it means a somewhat working but really inelegant solution)
kraqh3d Wrote:but i have had an idea regarding this problem for a while. [...] its a very simple thing -- generic tagging.
Yes! that is exactly what I was proposing in the previous message, and what I intend to simulate using <genre>. But having both <genre> and <tag> would be much better... but that could only be after atlantis, of course
