From: "Hal E. Fulton" Date: 2003-07-01T06:10:51+09:00 Subject: Re: Message catalogs (I18N) overnight hack... ----- Original Message ----- From: "Josef 'Jupp' Schugt" To: "ruby-talk ML" Sent: Monday, June 30, 2003 3:40 PM Subject: Re: Message catalogs (I18N) overnight hack... > > And another is to preserve readability and simplicity -- I dislike > > having to call a method every time I use a string. > > IMHO redefining a built-in command's effect on a built-in data type > makes code unreadable. I see your point, and someone else expressed the same opinion. But I stand by this principle. I like it personally, whether others do or not. Personally I don't want to sprinkle 'tputs' and such throughout my code. I want to modify my code as little as possible. Besides, a name like tputs will lead the reader subconsciously to think that something unusual is happening, when it is really (in effect) just a puts. Sure, it's doing a translation, but I want the programmer NOT to think about translation as he looks at the code. In other words, I want translation to be as transparent and behind-the-scenes as possible. It's not true AOP, but it has an AOP-like flavor to me. At any rate, the library may not ever be finished, as no one has expressed real interest in it (other than in changing the design). I don't want to make *fundamental* changes in my design without a reason that I agree with; if I did, it wouldn't be my project anymore. :) If someone likes the overall idea, but dislikes my concept, then they can write their own code. In fact, I have discovered there are similar libraries already in the RAA -- they differ from mine only in having a larger, more intrusive footprint in the code (and obviously they have more features). Cheers, Hal