From: Nikolai Weibull Date: 2011-03-17T19:44:46+09:00 Subject: Re: The future of the character-encodings library On Thu, Mar 17, 2011 at 00:56, Eric Hodel wrote: > On Mar 16, 2011, at 3:15 PM, Nikolai Weibull wrote: >> Eric, could you please reply to all in the future? > No.  I don't know two of the email addresses in your To header so I can't judge if my response is topical for them. But I do and I made the judgment call for you. > The third appears to be a mailing list to which I am not subscribed.  I don't wish to fend off possible "you must subscribe" bounces. That is a valid point. I should have cross-posted my request for help instead. >> I have “skip” set for this mailing list > I think it means that you don't want to see messages from this mailing list. Correct. > If this is true why did you post to it? Because I wanted this to reach as many (interested) people as possible. If I’m going to make a big change here I want as many to know about it as possible. I know that people have used the library in the past, especially in back-ends, which makes it a lot harder to know how many users I actually have. I have, believe it or not, even been paid (minute amounts) to work on this library. I figured that perhaps there were some hidden users that I didn’t know about that were still using it and I therefore posted to the most public Ruby forum that I know of. >> as, as you point out below, it’s rather high in noise. > I don't see where I made this assertion. You implicitly made (I thought at the time, see below) it by saying that the library in question doesn’t have that many users and, as such, my posting wasn’t relevant to the majority of the readers of this list. This low level of relevancy is something that I have judged to be the case for many topics on this list. >> It makes it rather hard to stitch things together when I can’t easily reply to your reply. > I don't see why I should be inconvenienced to make it easier for you to see responses you do not want to see. The inconvenience that you would have to endure by pressing Reply to all and removing the char-encodings list from the Cc list must surely not be as great as that which you have put me through by not including me in the Cc list so that I would receive your response to the posting that I made (that I, of course, do want). Either way, this is a moot point, as I’ve now set noskip. (I was hoping that either those that replied would include me or that the mailing list software would be intelligent enough to not skip replies to my postings. I was wrong.) >> Eric Hodel wrote: >>> On Mar 16, 2011, at 6:26 AM, Nikolai Weibull wrote: >>>> Could any of the users of this library please provide me with some input on this point. >>>> >>>> I’m looking forward to receiving your input! >> >>> There don't appear to be many users of character-encodings: >>> >>> https://rubygems.org/gems/character-encodings >> >> I don’t see how this is relevant, but thank you for pointing out my >> failure in selling and maintaining my library. > I was attempting to suggest that since there aren't many downloads for your gem maybe there's no need for you to continue to maintain it in its current form (if at all). Then, for my sake, please say so. A short – easily interpreted as snide – remark like that can easily be misinterpreted. > Some of the functionality of your gem has been taken up by ruby 1.9.  Anyone seriously considering handling encodings other than US ASCII should move to 1.9.  I would rebuild character-encodings atop 1.9 if I were in the maintainer and had such a need. To what need are you referring? The whole point of the library was to provide UTF-8 support for 1.8. I now want to shift focus to both providing support for UTF-8 for those of us stuck with 1.8 (due to 1.9’s horrendous I/O and require performance on Windows) and as an extension to 1.9’s built-in Unicode support. Looking at 1.9 it is now (because it sure wasn’t in 2006 when I began developing this library) clear that Ruby won’t be supporting a lot of features that would be desirable. You can, for example, not easily perform collation, normalization, or character-class lookup. Even such a thing as String#upcase doesn’t seem to be able do the right thing. I might be doing something wrong, but # -*- coding: utf-8 -*- puts "äbc".upcase prints “äBC”, not “ÄBC”. > Due to the low number of downloads you have an excellent opportunity to throw out your existing API and rebuild your library to integrate well with the encoding features of ruby 1.9. I know that this type of behavior is popular in the Ruby community, but I wanted to give my, albeit few, users a chance to have their say on this matter. > I don't see why you would consider a low number of downloads to be any failure on your part.  I simply made a statement of fact. There are many statements of fact that one can make that are often best not made. As I already noted above, you need to contextualize such a statement so that it’s not open for interpretation. Had you written something along the lines of “Judging from the statistics at rubygems.org, perhaps you can get away with your proposed changes without too many users becoming upset?”, I would have known what you were trying to get at. As you wrote it it only stands as a pointless remark. > I have many, many gems that nobody uses and I no longer maintain. I actually wish to continue maintaining this library and I actually do have active users. > It would be ridiculous for me to attempt to attach any judgements to such a fact. Am I ridiculous for not sharing your level of detachment from your work? I don’t know if you actually looked at the source code, but it’s actually quite a few lines of (sometimes rather complex) code, and for me to throw it away without at least considering its future utility is not something that I could easily do.