Return to MSX Snatcher Translation main page
MSX Snatcher Translation Problems
Note: This article was written at the point of January 2002. Since then,
a lot of additional problems were identified, and most were solved.
Text display routine isn't really found. Snatcher engine should contain somewhere the "pointers" for display routine to gain certain text. The display routine does not seem to check appropriateness of text data. (See Jan 13, 2002 entry.)
Currently, Frits is working on this matter. One idea is to apply numbers to all of the messages, and let the new routine picks up referent translation line whenever it receives an order to display which message, instead of receiving a command "pick up message from address xxxx in this disk", which is the case in the original game (am I right, Frits?).
The size problem is not very serious. Their entire size is actually smaller than Japanese version. However, some messages are excessively longer than original.
One idea is "numbering" system mentioned earlier, and other is "dictionary method" which showed enormous compression rate in Metal Gear 2 translation.
It's the mystery why same texts appear more than once within one disk. If this is nonetheless due to developer's careless mistake, it's best to delete them and use for storage purpose.
As the point of fact, people other than Metal Gear, who is the bigmouth know-it-all, do not employ too much vocabulary.
Japanese language does not have "space" concept. This means Snatcher does not have "line break" function which automatically breaks line to prevent one word displayed over two lines. There are a few spaces in Snatcher, but looks they are cleverly placed to avoid ugly look.
Automatic "line break" feature needs to be implemented.
Some people, for instance the Chief, moves his/her mouth while speaking. It looks to me their mouth shapes are correlated to what they are speaking. If that is the case, translated version should synchronize mouth movement to texts as well.
Speaking Chief, who uses four mouth patterns when he speaks.
The texts displayed inside frames (Metal Gear display, GAUDI and Gibson's PC) seem to be governed by different routine from the one for displaying speech messages.
|Metal during analyses||Gibson's PC|
Snatcher requires text input several times. In original version, whatever you type is treated as a kana input. Take a look.
Use keys to input name of the person
|you want to search|
|Ko Ji Ma Hi De O*_ _ _ _ _ _ _|
The characters "ji" and "de" have accents which look like double quotations, don't they? In Japanese kana writing system, adding dakuten (double quotation) or handakuten (a little circle) changes pronunciation of the target kana. In this case, what is pronounced "shi" without dakuten comes to be pronounced "ji".
Of course this concept doesn't exist in any language beside Japanese, so should be deleted.
Of course, a routine to modify kana input to alphanumeric input should be devised as first concern.
At least two important texts are written in form of graphic.
|Find home!||Pachinko minus "pa" equal a totally obscene word|
They should be redrawn unless you prefer Blade Runner-ish Japanese style. Frits know graphic decompression routine, so redrawing is possible.
In original, length restriction applies to menu and people's names.
Since this restriction is more because of screen size rather than text routine problem. However, I noticed one exception regarding names. It's the "bum" found out of Ivan's house. This is what happens when a name length exceeds 6.
Whatever these restrictions can be eliminated or not, I prepared shortened versions for both menus and names.