Re: [Eliot-dev] Catalan language file and dictionary

[Joan Montané]

Hi Olivier

I've been testin Eliot with multichar tiles support. Works really
fine, like a charm. I've found only a two very little 'bugs'. Nothing
really important.

1.-  If you allow wrong words. Then they are market in red (thats
good), but the wrong word is displayed with using internal letters
(Q,W,Y).
2.- Some sorting problems on output listing at dictionary tools.
Usually words are sorted by internal letters, and Ç char is sorted at
last (it should be sorted behind C and D tiles)


2009/6/29 Olivier Teuliere <address@hidden>:
> On Sun, Jun 28, 2009 at 8:06 PM, Joan Montané<address@hidden> wrote:
>>> The only bug I'm aware of is that typing 'Q' or 'Y' transforms the
>>> string into 'QU' or 'NY' respectively, which should not happen, since
>>> the internal character is not supposed to be a valid input string for
>>> the tile. I know where the bug comes from, but I don't know how to fix
>>> it yet :-) However, it shouldn't prevent playing normally.
>>>
>
> Actually there is another little problem: if 'L.L' is present in a
> regular expression, it will always match the 'L·L' character, and not
> "L followed by any character followed by L".
> An easy workaround is to use "[L].L" instead...
>
About these other 2 minor bugs.
I think the first can be called easy "a feature", :-D
About the second bug... dont' worry. I think "L·L" tile will be the
unique scrabble world tile with an input containing a dot ".", so just
We must advise Catalan users.
.
>> Well, the main feature request is an option to validate/unvalidate
>> manually words, specially for duplicate mode (also for computer
>> player). The reason is that DANOSC (and maybe other languages)
>> dictionary it's not 100% perfect. We have detected some missing and
>> wrong words in DANOSC. It's a quick hack from Linux catalan dictionary
>> and it's hard to update properly.
>
> I don't think this is a "good" feature to add, for 2 main reasons:
> 1) Eliot is designed to work with a correct list of words. If some
> words are incorrect in a dictionary, the dictionary should be fixed,
> but this is not related to playing Scrabble.
> 2) Even if Eliot could maintain lists of words marked as "valid" or
> "invalid", you would have no guarantee, after playing thousands of
> games, that all words will have been flagged. To get a 100% correct
> dictionary, you would still need to do the painful work of checking
> the words one by one (which you are already doing).
>
I see. In fact, we must have a 100% correct wordlist. So we must work
hard on it, now.

> In my opinion, a better approach would be to generate, from time to
> time, a new dawg with an updated list of words. I wouldn't mind
> replacing the previous one on the website.
>
Yes, you are correct.

>> Another minor request is a master duplicate mode. I think in France
>> it's also played. This mode requiere save and load features, :-(
>>
>> In master duplicate a player, call it Master, planned a full duplicate
>> game (racks and played turns).
>> Then, the others players, play duplicate as normal, but the racks are
>> the racks planned by Master (they aren't random), and the choosed
>> played word is always de Master played word, neverless isn't the
>> better score. This mode is played in schools.
>
> Indeed, this requires save and load features, but they were already on
> my todo-list anyway :-)
> Once saved games are possible, I think it wouldn't be very hard to
> implement master duplicate mode (BTW, is this the official name for
> this type of game?).

Ummm. There isn't "official name". We call it just "duplicate" (both
modes) and, if needed, say "planned duplicate" for the "master
duplicate". I will check other Catalan users, so wait a two or three
days for a finish answer.


Joan Montané.