what tweye said. And perhaps the 'wider search' would also find things that are almost the same? like the IMDB search....

damn i had a reply ready but was going to reboot, basically, I've got:

searchMovie(String searchString)

which will match exact (case insensitive), so adam ant will get:
Adam Ant
Adam Ant, The
Crappy Adam Ant
etc.

I guess its a good idea to add in another search, perhaps:
searchMovie(String searchString, true)

which will parse out spaces, so will give:
Adam Ant
Adam Ant, The
Crappy Adam Ant
adamant
etc.

Then i guess:
searchMovie(String searchString, false)

can give:
Adam and the special kind of death Ant
Ants, Adam

Anyway, posting to say, I've linked up Movie with edition now:
<img src="http://members.lycos.co.uk/horrorhasahome/uploads/akira.jpg" width="800">

cool
John's idea about fuzzy string matching is great, even tho i think it's not a trivial task. :-o

Well i guess the first two searches might be good, ignoring and including spaces, a fussy search though will be difficult.

It might just be easier splitting the search string by token then doing a recursive search... for example:
"Search for me"

Tokenise by " ", gives Object[] = Search, for, me

Then use the existing search recursively:
.*Object[1].*
.*Object[2].*
.*Object[3].*

That will do a search and return a collection for everything with search in it, then you do a new search on the returned collection for for, then you repeat with me.

That will get everything in the database with search for me somewhere in it.

.* = any character zero or more times.

Problem is that time complexity isn't so hot, but luckily its a vastly reducing array size each time.

Infact, the way i've done it that's just given me an idea!

All I need to do is tokenise the input string then i can apply the filter 3 times and then only do one search call.

But to prioritise words that aren't matched is a hard thing to do, as said, so I don't think its going to be easy/practical to do a search that will match things that aren't close matches. Unless i play with the function and return a collection of collections, with the first element = exact match... bah! Let's just put that on the maybe pile...

just for the sake of conversation...

you could create a dictionary of all the words available in your database, i.e. you tokenize every title string and put them in a collection, which will result in a smaller set cause some common words are likely to be repeated. you do this operation after every import from external file or after every manual insert of a movie.
you make this big fat collection Serializable, so you can save/load it on a file for future use.
then if the search doesnt produce any result, you could search for the most likely word in the collection with this handy already-made java algorithm and present the highest matching words to the user in a "maybe you were searching for.." fashion (google style).

so, basic search functions remain untouched and it's up to the user re-do the search with the similar words.

Sounds great in theory, but when you take into account the time complexity of searching a dictionary of so many words, and the fact that its searching a list of movies you actually already own (rather than taking a stab at looking for a movie you once saw on tv), I'm not quite sure implementing such advanced search methods are necessary, at least not for a while anyway.

However, given the java code for the edit algorithm is there (bookmarked, that's going to come in handy ), I'll flag it as a phase two search solution

hey just spotted EMM cheats now when reading filespecs, it uses a thirdparty freeware app called mediainfo.exe

Another screenshot (large again [sorry ] so clickable):
http://members.lycos.co.uk/horrorhasaho ... -again.jpg

That's sorted by imdb rating. Really playing around, I have to get stuck into some tedious gui construction using swing.

One more since I just added a letter filter

I have to say, gui design in java is tedious as hell!

Here's todays screenshot:

Some bits, clicking on # -Z filters the list to that letter. clicking the x on the tab closes the tab now, the tick and cross for seen is also clickable.

Probably moving onto the media tab next.

I've been batting on with the Media tab too, I also support 1027x768, it shrinks down the images to fit, whereas in 1600x1200 and proper resolutions, it keeps the screenshot large:


Ignore the misshaped middle tab, i haven't tweaked the layout yet.

The whole bottoms tab is wrapped in a JScrollPane so if it goes oversized, you get the scroller (so it will still all run on 800x600).

The tabs are full closeable, and can be closed by pressing the x.

cool! looking good! How have you added all your titles to it though? Imported from EMM, or added each one by one?

PS: There's nothing in genre...?

imported from EMM,
oh and there is something in genre, you have a selector for the main genre, and then a textarea for any subgenres, when your adding media it will have one box and some up, down, add and remove buttons to juggle them around

ahh... now I notice. hehe.

Okay, at this stage, I think i'm definitely going to end up using this over EMM myself, but then again that was the point

been messing around with the structure of media storage (got it now), anyway, the bonus is that you will, ultimately I hope, be able to compare your rips from any region, in any framerate and determine which is longer.

Simple, but hopefully it will be very handy

had to have a break for a couple of days, but I might get it to a gui test stage in a couple of days (then again i got a bit of work to do), it won't have any functionality, but it will help find any gui glitches/problems on other OSs and resolutions.

okay, no more coursework... back to actively developing this again. Hooray!