EHM The Blue Line

Thinking about what I said, would it be possible to set up the search function so that Nicklas Backstrom would return Nicklas Bäckström as a result. Basically the search engine would identify ö as o and vice versa. Would make searching easier.

It wouldn't help with Göteborg/Gothenburg or Fyodorov/Fedorov, but not much can be done there. A wildcard character could help a little.

This is assuming the program will (eventually) have a player/staff search function.

Perhaps we could implement search tags. But then there would be extra work required for names needing extra tags.

dabo wrote:Perhaps we could implement search tags. But then there would be extra work required for names needing extra tags.

So by default if we search for Petr Čajánek for instance we'll have to actually type "Čajánek" if we want to find him? I'm very fond of the European spelling and wouldn't want to only have the North American standard, but I reckon it'd pose some problems for those of us whose keyboards lack some characters, me included.

A simple way might be to simply add two extra fields in the database for North American standard, one for first name and one for lastname. This way those who prefer Euro names will have them, those who prefer NA names can have them by ticking a box that tells the program to use those two fields rather than the default ones when returning player names and when searching for names, the search program would scan both the Euro and NA fields to find a matching name and return results according to the user's spelling preferences. I think it may not be too difficult to do (but well, with programming and me, it's always very easy until I get to write actual codes ). That's extra work for the researchers to give the players both versions of their names and to have namesets come in two versions when necessary, but it might make everyone happy in the end.

That extra field might even offer the possibility for those who feel like having Russian names written in cyrillic or Japanese names in kanji to edit the db accordingly, assuming the game recognizes those characters.

By the way, I have an extra question about the db. Will the researchers have to add the fruit of their researching one by one in some order or can multiple partial dbs be merged into one to form the main db? Which opens a second question: will users be able to make db improvements that can be downloaded and merged with the official db? THAT's not easy-looking coding though


- YZG

dabo wrote:

YZG wrote:So by default if we search for Petr Čajánek for instance we'll have to actually type "Čajánek" if we want to find him?

If we used tags "Cajanek" would be one of them so no need for any special keyboard. But this is something me and jatahati will have to discuss and see what the best way is.

YZG wrote:By the way, I have an extra question about the db. Will the researchers have to add the fruit of their researching one by one in some order or can multiple partial dbs be merged into one to form the main db? Which opens a second question: will users be able to make db improvements that can be downloaded and merged with the official db? THAT's not easy-looking coding though

I assume multiple users will be able to edit the database at the same time. There will be instructions when everything is set up and ready.

There should not be any technical difficulties to create such searcher that will automatically find Čajánek when typing Cajanek. But if db uses russian spellings where names totally differ in both length and letter with american spelling, it's a almost impossible to make it work. Only possibilities I come up with is already mentioned tagging or making two different spellings for a name. If both ways are well requested by gamers, then in my opinion best approach would be game setting (for every player to chose) for either US spelling or International spelling. It would of course put lot of load on researchers to try and find out both spellings. But then again I think that this is not something that should be done in first versions. We will obviously make things so that It might be easy to implement later on.

Well, in EHM some of the search work both ways, i think swedish a,e,o etc work even if you're searching for example Backstrom

So maybe one might do tags for letters only, which would be simpler...

dabo wrote:Perhaps we could implement search tags. But then there would be extra work required for names needing extra tags.

That's a pretty good idea. We could add additional name tags for specific players where it makes sense.

The alternate name db would be a lot more work, I'd go with the tags myself.

But couldn't one write a routine like "if the user types C, show c, č, ć", if the user types A, show a, å, ä, and so on"? It would be easier than doing 2 namesets

Alessandro wrote:But couldn't one write a routine like "if the user types C, show c, č, ć", if the user types A, show a, å, ä, and so on"? It would be easier than doing 2 namesets

This is exatly what I meant on my post, we will do this somewhere on the way to the search.

empach wrote:Thinking about what I said, would it be possible to set up the search function so that Nicklas Backstrom would return Nicklas Bäckström as a result. Basically the search engine would identify ö as o and vice versa. Would make searching easier.

It wouldn't help with Göteborg/Gothenburg or Fyodorov/Fedorov, but not much can be done there. A wildcard character could help a little.

This is assuming the program will (eventually) have a player/staff search function.

A wildcard character (usually * or ?) would be a huge help in the search feature. I've always had problems searching for players in EHM because you don't get the spelling exact. Also, I hate that EHM searches for EVERYTHING at once, not just names. It searches teams too. I want to pick what I search for.

Oh and I like the idea of leaving a spot in the DB for a North American/English language translation for names. That way it can show it's original by default but still have the option of showing the modified spelling. Even if these names aren't all translated for the original DB, I think it would make sense to have ability to do it in place and have a spot in the DB for it so you don't have to re-do it later.