[ruby-talk:00439] Re: New feature for Ruby?
From:
Clemens Hintze <c.hintze@...>
Date:
1999-07-09 04:44:10 UTC
List:
ruby-talk #439
On Fri, 09 Jul 1999, you wrote: >Hi > >In message "[ruby-talk:00431] Re: New feature for Ruby?" > on 99/07/07, Clemens Hintze <c.hintze@gmx.net> writes: [...] > >Oh! Why do you know that Ich war fleissig Schuler nicht??? As you >guess, I can't understand German. But please let me know when you >write. I'll try to read that. Wow, how many Japanese have learned German once? ;-) Now I know more Japanese be able to speak German (perhaps only a little bit) than I know Germans being able to speak Japanese. :-) But on every case I will send you such an article, if I ever have the opportunity to write one! :-) [...] >>It seems only mine :-) I think, that Sequence would also be okay, >>perhaps. As I have told you, the name is coming from the Smalltalk >>world not from any logical or even mathematical insight! I have thought again, now I think, you was right to insist. The best would be, that we would change Range to fulfill all these requirements. But if I cannot convince matz, I would now take `Sequence' as name. Not only reminds `Sequence' me as "... descrete series, i.e., enumerable and one dimensional.", but also it seems to imply a direction (up or down). > >My apologies if you feel that I say same things again and again <bow>. Oops! I was not feeling, that you say the same again and again. I was feeling, that `I' saying the same again and again! :-) But that was not as I think you are not able to understand english, only I fear, that my Pidgin english was not clear enough to transfer my opinion very clear. :-( You know, one time I let pass a good opportunity, as I could not make clear my desires. After some time, it was resolved by another guy. Than I have thanked them, and asked why it was now implemented and not six month before? The answer was something like: "Oh! That was what you have meant! Sorry we totally have misunderstood you." and I have even not recognized, that I was misunderstood :-} To not repeat that sat feeling, I tend to repeat my opinion over and over again, until a decision is made AND reasons are clear and logical for me (if the decision was `no', where I hope it was `yes', of course) ;-) >I remember your naming is from Smalltalk. Though it is not bad, I'd >like to name to be more plausible or more intelligible. I believe the >name is very important; The most part of design, I believe, is >choosing the name. A suitable name leads us what feature should be I agree with you! >supported. I use mathematics to name in many case because just I do >not have the sense of word in English. I'm not intending ``the world, >be logical'' :-) Whereas I feel that you're right, but it seems most in the computing world is not mathematic, for good or for bad. :-) But in our case here, I second you now! [...] >Me too. Your opinion is very persuasive. However, >Regarding Integer only, int.step(to, step) can make it: I already have recognized that. But I want that feature for Floats and others too. I dont like, that I have to choose Integer#step, String#upto and <class>#<whatever> to perfrom nearly the same task. So I would insist of adapting class Range or write a new class Sequence. [...] >Is it esoteric? Saying esoteric, I mean that I have to use different ways to archieve the same goal with different kind of classes :-) [...] >>It really hurts me, that the mighty Ruby have use some esoteric >>methods to archive the same goal :-( > >Perhaps, I don't understand the above sentence, maybe. I meant, such crab like BASIC can use _one_ construct `FOR A=1 TO 9 STEP 2' to do all what I desire here. But Ruby has to use Integer#step, String#upto (as step doesn't exist now) and for Float there exist nothing of these methods. > > >By the way, I introduce tips to redefine `new' here. You can use it >in order to modify the Range to have some properties (e.g., stepsize) >by specifying them as options. Sorry! I don't very like it (the solution, not your trick ;-). It would also modify Range, but would not introduce all of the features I want. If we decide to touch class Range, I would like to let Range behave like current class Interval. I would code it in C. That would make it a cleaner solution. What do you think? But nevertheless thanks for your lessons :-)))) \cle