About the final issue. I thought about this one for a long time. My conclusion is basically that it should not happen. If you do so, the page diffs people will see will include both translation changes and the new content. Making it as two separate edits makes things much easier.
There is no real constraint on this in the back-end. It would be perfectly possible to create a translation bit when a certain checkbox is activated on the translation UI.
Re: Issue #???
alain_desilets Tue 05 of Feb, 2008 21:43 GMT
> About the final issue. I thought about this one for a long time. My conclusion is basically that it should not happen. If you do so, the page diffs people will see will include both translation changes and the new content. Making it as two separate edits makes things much easier.
>
> There is no real constraint on this in the back-end. It would be perfectly possible to create a translation bit when a certain checkbox is activated on the translation UI.
You mean Issue #11, right?
When you say that "it should not happen", what do you mean?
Do you mean that:
a) Users will not tend to do this?
or
b) Users should not be allowed to do this?
If you mean a), I disagree. I have done this many a time, eventhough I was aware that this would screw things up totally. In the heat of the action, I would just not remember that. I think "normal" users will be even more likely to do so.
If you mean b), then there are two question to ask:
1) How can we actually prevent users from doing it?
2) In doing so, will we be interrupting the user's natural flow too much?
One way to prevent users from doing this would be to provide a more consrainted translation UI. For example, we could have an array of English lines on the left and an array of French lines on the right (assuming translation from English to French). Only those French lines that correspond to changes that need to be brought into French would be editable.
This is the design that TraduWiki adopted.
Note that this solution is not full proof. If the user is translating a sentence that needs to be brought into French, and then decides that this sentence should be modified altogether, he will still be able to do it. But at least, it minimizes the chances of that happening.
Would an interface like that interrupt the flow of the end user too much? Probably not. If the user sees something that needs to be changed in a part of the text that does not need translation, he would see that he can't change it there, and would naturally tend to write himself a note.
In fact, this kind of interface might make it easier for the end user to locate the place in the French side where a particular English change must be translated.
There is no real constraint on this in the back-end. It would be perfectly possible to create a translation bit when a certain checkbox is activated on the translation UI.
>
> There is no real constraint on this in the back-end. It would be perfectly possible to create a translation bit when a certain checkbox is activated on the translation UI.
You mean Issue #11, right?
When you say that "it should not happen", what do you mean?
Do you mean that:
a) Users will not tend to do this?
or
b) Users should not be allowed to do this?
If you mean a), I disagree. I have done this many a time, eventhough I was aware that this would screw things up totally. In the heat of the action, I would just not remember that. I think "normal" users will be even more likely to do so.
If you mean b), then there are two question to ask:
1) How can we actually prevent users from doing it?
2) In doing so, will we be interrupting the user's natural flow too much?
One way to prevent users from doing this would be to provide a more consrainted translation UI. For example, we could have an array of English lines on the left and an array of French lines on the right (assuming translation from English to French). Only those French lines that correspond to changes that need to be brought into French would be editable.
This is the design that TraduWiki adopted.
Note that this solution is not full proof. If the user is translating a sentence that needs to be brought into French, and then decides that this sentence should be modified altogether, he will still be able to do it. But at least, it minimizes the chances of that happening.
Would an interface like that interrupt the flow of the end user too much? Probably not. If the user sees something that needs to be changed in a part of the text that does not need translation, he would see that he can't change it there, and would naturally tend to write himself a note.
In fact, this kind of interface might make it easier for the end user to locate the place in the French side where a particular English change must be translated.