RE: Roundtripping Solved

From: Lars Kristan (lars.kristan@hermes.si)
Date: Mon Dec 20 2004 - 11:10:13 CST

  • Next message: Philippe Verdy: "Re: [hebrew] Re: atnah hafukh"

    Mike Ayers wrote:

    > Wrong. What you are proposing will only encourage
    > people to try to make UTF-8 and legacy encodings live
    > together indefinitely, which will make for far more pain than
    > a hard conversion would.

    I prefer evolution to revolution. If you ask me, there is no way you can
    convert everything. Some users won't go for it. Especially in environments
    where Latin 1 is all they need. First you have to force all users on a
    system to change everything. Filenames first, then for some files also the
    contents. Then you need to force all users in a network to do the same. Some
    will resist, especially after hearing about the painful process. You'd
    expect that the hard way would be quicker, but I am not so sure. Maybe it is
    quicker for a single system, but you need to interact with other systems.

    I would much rather switch to UTF-8 and rename files when I find the time.
    And perhaps just those one that I regularly use. But if I will have problems
    accessing the files, I won't even be able to rename them.

    And then you have hierarchical storage systems. That'll be havoc. Need
    something from archives? A restore perhaps? Hopefully you'll only need a
    single file, so you can specify a new name while restoring. Otherwise some
    files won't restore at all, or will restore, but you won't be able to access
    them. According to Murphy, you will be in a real hurry, and there will be
    nobody in the IT department on a Saturday to help you either.

    It is so easy to put "responsibility for the filename on the file's creator"
    as you once wrote. Creators of many of the files are no longer around. There
    are tons of documents nobody owns. And even with user's own files, do you
    think they'll accept that responsibility? The blame will be on IT department
    and/or the software.

    But maybe you are right. Maybe think this is a bigger problem than it
    actually is. All we need is someone from the list to tell us all about how
    they switched to UTF-8 on their UNIX system.

    > Incidentally, what you want to do is, in the general
    > case, impossible anyway. Look closer and think broader, and
    > it should be obvious.

    I am trying to think broad. All the cases. All the scenarios. All the users.
    Not just those that speak English and never had any need for a letter
    outside of ASCII. Or those who know enough about computers that have always
    had a feeling it would be better to stick with ASCII names even though they
    didn't need to. And that constitutes the majority of people on this mailing
    list. Including me.

    And, I don't see what is impossible about it.

    Lars



    This archive was generated by hypermail 2.1.5 : Mon Dec 20 2004 - 11:16:03 CST