View Full Version : "the" items (artists, albums, songs?)
since i'm making changes to the database anyway, i wanted to see how people felt about those pesky "the" items. should they be added to the database with ", the" to sort correctly?
if so, what should be altered to this format? artist names? album titles? song titles?
lastly, should the ", the" addition use the same capitalization as the original tag, or should it default to ", the" ?
(note, this change will likely mean you need to delete your database and rescan. if not, you'll probably wind up with duplicates for all your "the" items.)
jmarshall
2005-03-02, 17:00
i think it should be everything (artists, albums + titles)
not sure if there's any other prefixes that need the same behaviour? can't think of any off the top of my head.
also, it should be doable without having to delete the database - i've added upgrade code to the video database, and this seems an appropriate time to add it to the music database as well. i can take care of this if you're busy with other stuff.
on a semi-related note, i've been thinking how to better do custom label formatting in the case where fields are missing. perhaps something like this:
[%n. ][%a - ]%t
would make sense? i don't think we'd need to offer the ability for brackets within brackets, so it shouldn't be too bad.
edit: actually, now that i see the system "in-print" it seems like all we need is to ignore all characters between a void field and the next non-void field - no need for brackets at all i don't think.
cheers,
jonathan
Gamester17
2005-03-02, 17:34
i have to say the hated words 'please make it optional' here, (as i and probly many more too prefer "the matrix" etc. instead of "matrix, the" etc. ...yes even in a a=>z database)
@gamester
right now i'm just targeting music, but i'm thinking about doing something similar to the video database next. and i knew you were gonna say to make it an option. but. it should be easy to make an option under library.
i was just thinking that we dont need to have the database store them in ", the" format, but just have the function that sets the label convert it to ", the" format. [is this what you were thinking, jmarshall?]
and now that i look at the sorting code, we can correct the sorting without even changing the label since the sorting uses a string copy. so its just a matter of removing the "the " from the beginning.
the question then becomes, do you want "the matrix" sorted by "m" even though the label says "the matrix"?
jmarshall
2005-03-02, 19:15
kraqh3d: yeah, i think it's best done in the sort routine as you say. no point having extra options unless they are absolutely necessary.
i think sorting the matrix in matrix seems most obvious to me. this is how every video store i've been to has sorted it, not to mention the phone book etc. etc.
cheers,
jonathan
so then do we have two options:
1) label "the *" as "*, the"
where this will actually change the label of the list item.
2) sort without "the " prefix
where this just sorts ignoring a "the" prefix
of course option 2 would not be selectable if option 1 is already selected.
jmarshall
2005-03-02, 23:27
my preference is no options and just implement number 2. i see little real reason for rearranging the label - they are supposed to read with the the in front, after all. the only reason to rearrange it is so that it sorts without the the. this should just be done at the sort level - no need for options then. no doubt someone will request that the "the" appears in the t's, thus forcing an option.
i implemented option 2 without a setting to disable it. it's only in the nav window for now. i'll commit it later tonight and we can guage the response.
Gamester17
2005-03-04, 11:02
where's my option god dammit?!, hihi :lol:
:)
i'm working on it. let's see if there's any interest in it or not first. what should the option even be called?
and should i later include it in the other music windows, or just leave it here?
my video list have serious problem
http://a1.myblog.com/xena/screenshot001.jpg
how to fix it?
6-24build
thanks
its correct. the sorting ignored a leading "the ". thats "the" plus a space. your last item does not have the necessary space. the space is necessary or movies like "they live" would sort wrong.
pervious version of xbmc no such problem
is there a way to fix it?
thankz
jmarshall
2005-06-30, 18:46
there isn't a problem.
as kraqh3d has already pointed out, all that is occuring is that titles with "the " at the beginning are being sorted by the second word, just like in practically every name based sorting method.
this has only just recently been done like this, which is why you didn't see it occuring previously.
example: if you wanted to look for "the matrix" in a long list of movie titles, would you think of "the" or would it be easier to think of "matrix".
the matrix a lot more easier for me (a-z )
that why i ask a custom fix for the filename sorting
Gamester17
2005-07-01, 09:48
please make it optional (i still prefer "the" listed first), ...pretty please with suger on top, pretty please with suger on top with whipped cream and a cherry :saint:
i still prefer "the" listed first
emefbiemef
2005-07-21, 02:20
yeah, me too. it's just pain in the eyes to find "the" in places that aren't expected.
RockDawg
2005-08-21, 05:55
while i can't imagine why anyone would want the word "the" listed first, i say give it to them as long as it's optional.
i'm getting sick of this discussion so i added an option to video and music so that this can be enabled and disabled.
Gamester17
2005-08-21, 20:11
i added an option to video and music so that this can be enabled and disabled.thank you! :bowdown: @rockdawg, personaly i prefer 'the' first because the movie is called "the matrix", not "matrix" nor "matrix, the" :p
while putting in the option, i noticed something a little odd... the "the " sorting was also present in movies files view. now, i can understand title view since it's supposed to have "correct" information, but i didnt expect it in files view. i was considering removing it completely from there. what do u all think?
if it's in file view it doesn't work. i always use file view and if it was doing "the " sorting i would have been complaining here also :d
noumenon
2005-08-25, 18:19
the way that files are sorted in the file view in videos has recently been changed back to sorting by the word "the". i don't like this, and i don't really understand why it was switched back. i've got a bunch of foreign films on my computer and they come up with funky names from the imdb, which is why i use the file view. for example - "the good, the bad, and the ugly" comes up as "buono, il brutto, il cattivo, il". as i am not an italian speaker (nor korean, chinese, japanese etc.... i've got a lot of kung fu), this drives me nuts. so... i guess my question is - why was this changed back? and my suggestion is - can we sort without "the" or at least make it an option in the settings like it is for the other views?
thanks
try searching next time... upgrade your version. i removed it from files, and made it optional for title and library views.
noumenon
2005-08-26, 00:01
i apologize, i meant no disrespect. i'm running a version that i compiled tuesday (through revision 1.1424), and i don't see anything in the changelog about this matter that has changed since then (1.1419 looks like the last time).
it is in files that i'd like to have "the" not recognized, or across the board with the preference setting. i use the files view to access my video files for the reasons i explained earlier.
thanks
sorry. i misread your note. you are correct. the option does not extend to files view... with good reason.
files view sorts by filename. xbmc should not be making any assumption in files view. if you dont want an item sorted with "the ", remove it from your filename. you have direct control over how your files are named.
killerbee123
2005-09-05, 18:34
files view sorts by filename. xbmc should not be making any assumption in files view. if you dont want an item sorted with "the ", remove it from your filename. you have direct control over how your files are named.
i don't understand why this feature can't be extended to the 'files view'. i don't have internet hooked up to my xbox so 'files view' is my only option.
i believe my answer explained it adequately.
i agree with killerbee123 here - it seems very strange, inconsistent, and unintuitive that there is a "general" sorting option, but it only applies to certain views.
where are you getting this idea that xbmc should be making "no assumptions" in files view? it currently makes a ton of assumptions, i.e. the file name cleaning, extensions removing, and reg-ex matching and stacking. how is this any different?
if there is an option for this, why should it not also apply to files view?
if the answer is just that it makes more coding work to do, i'll be happy to put it on my plate - i am leaning more and more towards this sorting method, and i use file view almost exclusively.
additional gripe: why is this only applied to "the"? there are a number of other similar prefixes that it seems it would also make sense to ignore, in addition to "the". in particular, this is very english-specific, and if i am sorting a list of films in another language, this doesn't help much.
that brings me to the conclusion: obviously it doesn't make much sense to hard-code every possible "ignorable" word for every language. so it seems that this would be best handled the way we are doing stacking - via regex. why not have a configurable "sort-ignore" set of patterns - then when sorting, the sort order would be determined on a special set of cleaned up file names, based on these regex expressions. given that most of the code is already there to do this, it doesn't seem like it would be much work.
and again, i would have this sorting done at the display level, regardless of whether we are in a db view or files view - if the user has explicitly set that option, they will certainly be confused as to why it only works "sometimes".
just another note in case it wasn't clear:
i understand that what you are saying is that because i can just edit the file names, that should handle any special thing i want to do.
however, the really great thing about this feature (and i swear it was working like this in a previous build?) is that i could still have the file display name saying "the matrix", but have it sorted as "matrix". that's the problem with just doing it with the file name - i don't want it to actually display as "matrix", just to sort that way.
killerbee123
2005-09-25, 08:19
and i swear it was working like this in a previous build
that is exactly why i was confused as to why the feature was taken out. i also use file view exclusively and it's good to see that you're willing to code it back in. it will definitly make browsing through movies easier.
ok, so here's the question: should it be only for "the"? i was looking at a list of english-language films last night, and noticed "a beautiful mind", which is currently before "about a boy" and "amadeus". i can't decide whether that makes more sense or not - often when looking for this film, i think i go to the b's, looking for "beautiful mind".
killerbee123
2005-09-25, 17:47
i just checked a few online stores for movies, and they also ignore the letter a when sorting. i think it would make more sense to ignore a aswell.
let's revisit this issue...
"a" could be added as well. what other prefixes should be ignored? maybe, its best to allow the user to define the list of prefixes lin xboxmediacenter.xml. something like this:
<myvideoignoreprefixes>the|a</myvideoignoreprefixes>
<mymusicignoreprefixes>the</mymusicignoreprefixes>
then, it can be altered for various languages:
<myvideoignoreprefixes>the|a|le|les</myvideoignoreprefixes>
i still dont think that files view should follow these rules. i think this is up to the user to alter their file names accordingly. if the user want things sorted like that in files view, then the user should name their files according. and if you want things displayed and/or behave differently then how they are named, i feel you shouldnt be using files view. as its name implies, its about "files".
but, i really dont care if its added back as long as its a seperate option in settings. i'm sure i'm not the only user who doesnt want xbmc messing with files view. i keep my files named a certain way for a reason, and xbmc should not be making any assumptions on why they are named as they are. do any operating systems do this? absolutely not.
of course, this will mean the settings screen will likely need to be reorganized a bit, as we're running out of space. both music and video should probably have a "library" screen in settings, in addition to "files". (i start on this while we continue discussing this.)
thoughts, ideas, or comments?
why would the options screen need to change from what it currently has? i.e. one simple option - that seems to do it.
regarding the files view - i'll say this again: i would like "the matrix" sorted in the m's, but i don't want it do display as "matrix" or "matrix, the" - i want it to display as "the matrix". i don't see how your philosophy on files view can possibly cater to this - and i'm clearly not the only one who thinks like this.
if there is already an option available for this, i can't see what the problem is. if you don't want it to apply, turn it off. wouldn't it be the same for title view?
Gamester17
2005-09-26, 10:43
i really dont care if its added back as long as its a seperate option in settings.i like to second that, if implemented please make it a seperate option, thanks in advance.
if there is already an option available for this, i can't see what the problem is. if you don't want it to apply, turn it off. wouldn't it be the same for title view?no it should be a seperate option for "files", becuase i "title" once might want to sort it with 'the' removed but in "files" view it should be the real file-name.
ok, but could we make that option "optional"?
;)
but seriously, i think you guys have some strange idea about what the file view is about. you seem to think it is meant to mirror my files or something. most of us who use file view do so because (1) we don't have internet connection; or (2) titles view doesn't really give us exactly what we want, and it is too much trouble to edit it manually.
i would actually propose that we add a new view option: "files-clean" or something like that - because we are essentially talking about two different usage patterns. for you guys who normally use titles view and occasionally switch to files view, it makes sense that the view options for files view should be different, and maybe much more strictly representing what you have in your file system.
but for the rest of us, for whom files view is our primary view, we could care less whether we see an exact representation of the file system - we just want to see the most user-friendly view of files possible. that's why we've implemented all these other nifty features, like stacking, file name cleaning, etc.
i personally don't care if it's one option or two, but i think this is going the direction of being highly unintuitive for less-techie users - the average newbie user, who hasn't discovered the amazing powers of the titles view, is going to be very confused about why the option they selected doesn't work when they haven't realized the difference between files and titles view.
it's like saying, "ok, let's have 2 options for subtitles - one for vobs, and another for divx!" for some users, that might actually make sense, but i don't think the majority of users would feel that way.
files view is a view of your *files*, no? therefore i think that how you have your files names should have priority. if you want them displayed in a particular way, you should rename them. i feel very strongly about this.
gamester and i seem to be on the same page. all i'm saying is that files and library should have unique options which enable the same feature so i can have it enabled in library, and disabled in files. i actually use the two views differently. i cant be the only one.
and i really dont see how this makes it "less" intuitive. there will be two seperate config sections - files, and library. a few options should already be in this missing library section.
and you mention the stacking and cleaning functions. these two options produce more questions than anything else. i dont understand why so many users think that xbmc should do the work. why cant they take a few minutes and clean their filenames?
but, i'm getting off-track now... let's put the number of options aside. if this becomes too much of a discussion point, it can be made a hidden option that needs to be set manually in settings.xml.
i think the more important question is how to extend this existing feature. today, it only works for the english "the". how should it be made configurable?
my opinion is that we should use a regexp expression, or if we want to be very fancy, a list of regexp expressions. the code is already there and works well for stacking - so it could easily be extended to work for sorting too. that would also allow us to go with english for the default config, and then foreign languages could be added later. this is the most flexible way, and would handle anything we might not anticipate that could come up later.
what i want to make clear, at least for the version of this functionality that we enable on files view, is that what i'm expecting here is something that modifies only the sort order, and not the actual names that are displayed. so again, i expect to see "the matrix", not "matrix, the" - but i want it displayed with the m's. a simple "([the ])..." regexp would take care of this case.
understood. it will not change the label of the item, just remove prefixes for sorting purposes.
yes, while i agree that a regexp would work, i think its unnecessary complexity. a simple pipe seperated string would suffice. it shouldnt need to do any fancy transformations, just remove certain starting words. i had an idea on how this would work and would be more than happy to share it with you.
btw, ([the ]) would not do what you wanted. the correct regexp is ^(the ).
your original regexp would potentially remove all t's, h's, e's and spaces. see what i mean about complexity? do we want to give users the ability to arbitrarily remove the second character with ^.(.) ?
edit... fixed some typos... its hard to type this stuff while im *supposed* to be working :)
ha, tell me about it!
just don't tell my boss...
ok, sorry, i should have said "pseudo-regexp", obviously wasn't going for accuracy. yeah i agree that it is more complex, but...
isn't it exactly the same issue we have in other places, i.e. stacking? i was on vacation for a few weeks while the stacking was changed to regexp, but when i came back, after i checked it out and started using it, i liked it. before that, stacking was just like what you are talking about - a pipe-delimited list of possible suffixes.
the main reason that regexp had to be introduced for stacking was due to the wide variety of file names that come in from various "sources" out there. i know your philosophy is that users should be responsible for renaming all of their files (and personally, i do) - but introducing the regexp definitely opened the doors to many more users to be able to use stacking without having to mess around with understanding how it works under the covers.
given that we are now essentially talking about a natural language problem, the potential for complexity seems even larger - before it was just the limited variety of file-naming formats for volume numbers - now we are talking about something that may need to be flexible enough to handle many complex natural language scenarios that we can't possibly imagine.
even though it introduced some problems initially, i think the strategy we've taken with stacking is working well now - basically, establish a good base set of regexp expressions that suit the needs of most users, and make sure those work well for most files. then on the exceptional cases, offer assistance.
again, agreed that this leads to more complexity, and not quite convinced myself one way or another. but i'm pretty sure that if we don't use regexp now, we'll regret it later.
RockDawg
2005-09-26, 23:07
this is a really good discussion about how to handle file names and i would like to make a suggestion. *i would like there to be a sortname field in the mymovie db. *the reasoning for this is that i like to have sequels displayed in order (rocky i, rocky ii, etc), but movies like lord of the rings gives problems since the movies aren't numbered and sorting alphabetically leaves them in the wrong order. *this would also give free reign to any kind of sorting a user would want (including ignoring "the", "a", "an", or whatever).
maybe the best solution would be to use a settings option (on/off) along with the regexp or pipes for files view only. *then any changes that would be desired for title view users could be done by simply editing the sortname field in the db. *this would minimize the number of options for the user and provide more flexibility to those who use title view.
just my $0.02
gglaze,
it was necessary for stacking because it was the only way to describe where the volume descriptor was located. in this case, you should only be truncating characters from the front of a string (aka prefixes). what more do you need to do?
rockdawg,
this is a little different thing but yes this has brought up before, and it it will be added to the video database. there are some other changes to the video database coming soon so i want to bundle them all together. the "sortname" will be a hidden column in the database that advanced users will be able to take advantage of.
RockDawg
2005-09-27, 06:11
just curious. *why have sortname be hidden and how advanced will a user need to be?
any word on what some of the other changes will be, or should i just wait and see?
edit: i almost forgot... thanks a ton for adding the sortname field! now, if we could just get a seperate db for tv shows all would be right with the world. ;)
there was another discussion on sortname, brought about by tomkun iirc. search for it. the consensus was to make it hidden. you would need to open the db on your computer and fill in the field for those movies which need it. (the indy series was the example used in the thread.)
mercury was working on something that allows a user to flag a movie was "watched" which would be used to filter the display list. since db changes require a db upgrade, i was going to bundle adding this new field in with that.
a little later on, the video db windows will become a video library window like the music. there was a seperate discussion on that as well. (unlike the music database, it will likely not be a multi-tiered filtering system.)
RockDawg
2005-09-27, 17:48
yes, i posted in that thread, but i didn't see your last post saying you may implement it. also, i misunderstood what you meant by hidden. in this thread i suggested using a sortname field and editting the db so when you said "hidden" and for advanced users, i thought you meant hidden in the db and for advanced db users. now, i see you mean hidden as in not displayed in xbmc and advanced as in needing to edit the db. gotcha now. sorry for the brain fart.
the "watched" tag sounds great too. that was another feature i missed coming over from meedio. i had seen the patch for it, but since i don't compile my own builds (use t3ch builds) i wasn't able to have it.
i'll have to check out the music library view (and that thread if i can find it). currently i mostly use the files there. either way, it sounds like it will only make things nicer. great job to all you devs!
sorry to have drug this off topic.
noumenon
2005-10-20, 19:08
i just wanted to say that i think it would be great if sorting without "the" could be implemented in file view, even if it is in some xml file. that is fine by me.
i'll just say that the reason i don't like the library view is that unlike music i can't change the metadata on my video files. i've gone through and custom edited the id3 tags on all of my music (16,000+ tracks... that took forever :bomb: ), but i can't do this with my video files, i have to rely on the imdb (which is especially terrible for stuff like tv shows). as a result i want to sort things manually via file view (and thus organize tv shows in folders, etc), but i also want the niceites of library sorting (like leaving out the "the"). so, i guess you could say that i want to have my cake and eat it too (what the hell else is the point of cake, anyway?). :)
anyway, i would be super fantastic happy if ignoring "the" could be implemented for file view, even if it isn't in the gui, and i think that there is enough of a contingent of other users that would agree with me to make this a desireable feature.
thanks
Raoul Duke
2005-10-21, 11:11
i agree completely with noumenon's post above. i used an older version of xbmc before the "ignoring "the"" was excluded in file view and now i really miss that function. for the same reasons noumenon gave i think it would be great if this could once again be implemented for file view (although i understand the points brought forth by kragh3d). to make it optional solves the issues for all of us and would make me (and noumenon) very happy :). whether it would be in the gui or xml doesn't matter.
about ignoring "a" when sorting, i personally don't think it should be ignored in the sorting, but i'm sure someone would disagree. however, being able to sort for example rocky, rocky 2 etc. in the right order would be a nice feature too. but that would just be icing on the (aforementioned) cake. ;)
i just want to stress once again that this feature is really missed! i actually signed up for the forum just to make this comment. keep up the good work with xbmc! it's appriciated!
thank you
/the duke