Re: Accelerator I18N [was: I don't wont the f* underlinings ...]

Hrvoje> IMO translation should be completely orthogonal from Hrvoje> program development. This is like saying documentation should be completely orthogonal from program development. I don't buy that, either. Hrvoje> All of these are fine with me. See the "courtesy of the Hrvoje> maintainer" part above. If "developing program" is the the goal, "courtesy of the maintainer" is fine. But the official point of the coup d'etat last fall was "serving users". See our de facto policy statement, _Architecting XEmacs_. If we are going to serve users, we need to do something about translations. My bad, in some sense---but I was caught completely unaware by Ben's changes. Hrvoje> Then we need to change how the translations are organized. Hrvoje> The whole point of gettext is to make development and Hrvoje> translation as orthogonal and unaware of each other as it Hrvoje> is at all possible. If every single change to the XEmacs It is as orthogonal as possible, subject to the constraint that maintaining the menus in at least one language be not so unpleasant that nobody would touch it. I doubt that Ben would have attacked the menu revision issue with such gusto if we had made things completely orthogonal by using catgets. The language that we, and GNU gettext, have chosen as the range of the message indexing functions just happens to be English, instead of the obvious natural-language-neutral choice, the integers. Hrvoje> UI requires a consideration for Japanese, something Hrvoje> somewhere is terribly wrong. No, it doesn't require consideration for "Japanese". It requires consideration for the fact that a single change to the UI implies more work for _all_ the translators. This is why we document and formalize (== make hard to change) external APIs: because there are many users whose own code must be coordinated with any API change. You yourself have often expressed sympathy for the "OK, let's forget about synching with the FSF" position precisely because rms and Sons, Inc, don't bother to consult with us on those API changes. But the response of a translator is not to fork. It's to _quit_. We want translators to be happy and think they're important. Hrvoje> I'll have to look at the menu translations. I thought Hrvoje> they were done through a gettext()-like mechanism. The X widget resources + massage_resource_name() technique we use is homomorphic to gettext(). -- University of Tsukuba Tennodai 1-1-1 Tsukuba 305-8573 JAPAN Institute of Policy and Planning Sciences Tel/fax: +81 (298) 53-5091 _________________ _________________ _________________ _________________ What are those straight lines for? "XEmacs rules."