rickx Wrote:well, allthough I appreciate the use of java being on mac or linux, I have quiet some things to criticize here:
- first of all, this app is still in a very immature phase, I would say alpha more than beta...but the point is: it changes an existing collection without any warning. In my case that meant messing things up more than adding missing informations. But I'll come to that later.
- regarding movies, it changes the file name, then adds other files (nfo, tbn, fanart) using that new name. That's what I think is not ok:
- the matching of the imdb is not well done (I have worse results than using xbmc).
- the choice of renaming the file is not good either: having movies in different languages, I prefer to keep the translated movie names: why make everything english? When you add wrong matches to translations, you may guess the mess...
- the choice of adding fanart and picture should probably be a choice as well. default should add the nfo only. or make that a choice...
- the existing subtitle files are "orphaned". they're not renamed and so will be lost most of the times.
- all the movies consisting of 2 parts (2 cds) are not recognized as such and instead treated as the same film twice. For "Clerks" and "Clerks 2" both in two parts I now have 4 times the same "Clerks" movie...as example.
- the browser in os x is buggy, so I can not correct the wrong matches - the error was already posted here.
So now what I would change:
- Add multipart movie regexp - should be default to avoid the mentioned error.
- do definitely a better imdb match handling: it might be imdb's fault, but the first match is the correct one on a rather low rate. One option could be that of sorting the search results, another to look for exact matches only. Yet another that of using scraperxml (--> search).
- consequently do all the automated steps only on the exact matches. or make this a choice.
- all the renaming and adding: I'd suggest to store all those infos in a tmp folder in the "batch-phase". Then show the results to the user and apply them only if approved (possibly giving choice to check what to apply/not apply...but that could become complicated)
- pre-existing files must be respected: exisiting srt/sub etc. must be taken into account
- an advanced option could be that of moving a all-files-in-one-folder structure to a each-movie-in-its-subfolder and/or back.
- a new feature which is missing is the already mentioned studio tag, probably the one tag making nfo files most useful as of now. the nicest feature of such an implementation would be that of opening the movies to get infos from the movie itself like resolution (--> sd, 720, 1080 etc.), audio + subs etc. Only the studio name itself must be grabbed from the filename (unless that can be fetched from imdb as well...). There are lots of threads about this.
I made screenshots of the batch results. That's because they're inserted line by line in a list control instead of concatenated in a big text field (so I could have copied/pasted the whole thing out). I will not post them here (20 files), but if you're interested, I may email them.
some examples of my matches:
leg dich nicht mit zohan an.mpg (german title) --> matched as "vento di primavera (1958)" (italian...)
renaissance.mpg --> star trek - first contact (1996) [wtf*k%&£$!!]
the nines (2008) --> dogville (2003)
and what about the rocker matched as The Hurt Locker (2008) ??!!!
p.s. the vlc player chosen as movie player does not work (or the play function anyway...). And I'm using 0.44.
Thanks for the input. It would have been nice to see them on the code's forum but it is worthwhile anyhow.
Let me make a couple of comments:
- The title matching between XBMC and MR is identical. The lookup is a simple IMDB search. Try 'Battle of Algiers' or 'Don't let me die on Sunday'. Both will give you the same results. MR only goes against IMDB.COM not IMDB.DE.
- If you would read the readme you would see that it does renames. It distinguishes it from other scrapers, good or bad. I wanted something that would correct my spelling errors and generally look at a library and know what is in it without having to resort to anything. If you think that is not the right approach, either do not use it or wait until it supports matching without rename. If you had looked at the code's forums you would have seen a post asking for how many folks wanted this feature. So far only a single person asked for it.
- Adding the thumbnail and fanart in addition to the nfo substantially speeds XBMC up. But it's a good suggestion to make it optional. Why did you not post it under suggestions at the site?
- You can rename the file to what you want. Once there is an nfo for it, MR will not touch it anymore. The Rename is in MR 0.44+ There is a popup selection.
- I think you repeated the fanart and thumbnail part twice.
- Moving into it's own folder. Great suggestion. Sort of contradicts 'leave movies alone' but why not? As you or somebody else mentioned the Aeon mod more or less requires it. I will need to do it if I want to use it myself.
- There is a studio tag. It contains the studio at the moment which is what it initially (and for other skins) was meant to do. It does not (yet) support the XBMC mod. I figured it is a good idea to cover standard XBMC first is a priority. I answered elsewhere that I will try to get it into the drop second week of June.
- Movie codec analysis etc. is already in the works. There are only so many hours in a day. I wanted to do it one way but then stumbled upon something more elegant so it took me a couple more days to sort out.
- Title examples you name: Please key them into IMDB.COM and you will see that the results match exactly what IMDB gives you. THERE IS NO MULTILANGUAGE SUPPORT IN IT, YET, (i.e. wenn jemand die Zeit nehmen moechte es zu uebersetzen (ich hab noch nicht einmal ein Deutsches Keyboard)). That is at the very, very, very bottom of my list. If you only want it to scratch IMDB.DE instead of IMDB.COM it can be done (although I am not sure about the tags, might have to check).
- ScraperXML - You have to be kidding. You must be one of the very few that can code that. I looked at it and ditched the idea. I could have clone the XBMC code but it just wasn't the way I thought it should work. You do understand that the other day someone asked for an app to edit advancedsettings.xml. I believe there are is a 95 to 5% ration of unskilled vs. skilled users. The unskilled users are probably so intimidated by the advanced heads that they rarely post.
'Only exact matches' can be done, no problem.
- Giving the use the option to accept IMDB response: Absolutely on the list. There already is an option for that in the profile. It just doesn't do it, yet. I had a thousand movies to process and didn't want to be prompted. When processing only new ones that makes sense. Cleaning up the mismatches from 1,000 took half an hour. Having been prompted would have taken days. But, as I said, it's halfway in. I started off with what Media Companion does and then wanted to add a feature set I liked.
- How do I figure out that IMDB hands back a translation instead of the English title? It bothers me, too, but I see no way of doing it. With all my movies it did it it exactly three times. It picked maybe 5% incorrect titles. I thought that was pretty good for my (English) collection.
- Stacking: The very version I posted contains the ability to specify stacking criteria. Since I am leaving town in an hour I did not get to implement more. It's the next thing on the list.
- Subtitle support (actually arbitrary like named files) will be in the next drop.
- The help (granted, outdated) mentions that the movie player must be able to accept the file name as an argument. The MR board posts the question whether you want an integrated player. One 'nice to have response' that is why it's not in. If more folks verbalized a need, it would probably become more of an issue. I initially had integrated VLC player support (using the VLC tool kit) but then ditched it since it entailed all sorts of considerations. The player I am looking at integrating is the Java media framework (which theoretically should be able to play all Directshow stuff, not sure about Linux and Mac. I can test Linux but I am in the dark about the Mac. Maybe I can buy a used small box somewhere. Oh no, I remember now, there is no such thing as an affordable Mac. Ok, just a dig. Sometimes I can't help it.)
Browser support under Mac - I have been told it works, all I can say. Since your response comes in this forum (which I normally do not track for bugs/enhancements) I have a hard time figuring out whether it really does. You could have responded to the statement 'It works' by the Mac user on the MR forum.
Alpha vs. beta: It's pretty stable and has few bugs but misses features (which it always will). Ergo my choice. Notwithstanding that I will remove the alpha/beta label altogether. You make the choice.
Don't get me wrong: I really appreciate the input. It probably doesn't help much *here* so I will copy the whole thing to the MR forums. (which would also have allowed you just to paste the screenshots in). A little less confrontational would have been nice (aber Deutsche sind nicht immer so taktvoll, weiss ich selber als Deutscher).
So, if you can work with me in the spirit of cooperation you will likely get all your features as most of them are either already done or on the list. I am a pretty fast coder but cross-platform support adds another degree of complexity. This is the only piece of code that does it, so be patient.
Again, thanks for the input.