Loading ...
Sorry, an error occurred while loading the content.

Re: [NH] Strange characters...

Expand Messages
  • loro
    ... I think you mean HTTP headers. That doesn t mean the host made that happen. ... Superscript 1 and 2, encoded the way they are, are not exclusive to
    Message 1 of 28 , Nov 9, 2008
    • 0 Attachment
      Axel Berger wrote:
      >loro wrote:
      > > but we didn't know it was the host's doing. I still don't.
      >
      >Where in Bob's source did you find UTF-8? If not there it must be in the
      >HTML headers. It has to be somewhere.

      I think you mean HTTP headers. That doesn't mean the host made that happen.

      > > That's another matter than the illegal windows characters you were
      > > referring to.
      >
      >No. The browser and the validator first look what encoding to expect and
      >then parse the code using that. So what is illegal and what is not
      >solely depends on that declaration. In another context those very same
      >characters may be perfectly legal, but that doesn't matter.

      Superscript 1 and 2, encoded the way they are, are not exclusive to
      cp-1252, and that was what you were talking about. They aren't
      "illegal", they just mean different things in ANSI and UTF-8. No
      validator refuses to parse Bob's page because of those two
      characters. They would have through an error had they been in the so
      called illegal range though.

      Lotta
    • loro
      ... This is so backwards. If Bob can use .htaccess, he should of course use it to make the server send the character encoding he prefers, which may very well
      Message 2 of 28 , Nov 9, 2008
      • 0 Attachment
        Axel Berger wrote:
        >Bob Gorman wrote:
        > > Something in the Head part of my html files?
        >
        >Yes, that too. First you need to shut up that server. You might need
        >help from your provider, but this line in .htaccess ought to be the
        >first step:
        >
        > AddDefaultCharset Off

        This is so backwards. If Bob can use .htaccess, he should of course
        use it to make the server send the character encoding he prefers,
        which may very well be UTF-8 for all we know. Why in the whole world
        not use it as it is intended instead of relying solely on a fallback
        mechanism like Meta?

        ><META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=US-ASCII">

        Except ASCII doesn't cover superscript 1 and 2, so that's hardly an
        improvement.

        Bob, just use the entities (¹ and ²) for now and be done
        with it. You can read up about character encoding when you feel up to
        it. I fear this will just confuse you and I'm sorry for my part in that.

        Lotta
      • Bob Gorman
        ... Yes, I did, & I m happy for now. ... I will. I obviously need to learn about character sets and this mysterious .htaccess, but it can wait till I get a
        Message 3 of 28 , Nov 9, 2008
        • 0 Attachment
          loro wrote:

          > Bob, just use the entities (¹ and ²) for now and be done
          > with it.

          Yes, I did, & I'm happy for now.

          > You can read up about character encoding when you feel up to
          > it.

          I will.
          I obviously need to learn about character sets and this mysterious
          .htaccess, but it can wait till I get a good night's sleep.

          I have 30+ web pages now and plan over time to double that, so I want to
          learn good practices now, to avoid excessive fix-up later.

          Thanks, and good night.

          Bob
        • Axel Berger
          ... Absolutely. It was way past midnight here, when I wrote that. ... Well, someone who has control over the server. And that usually is not the customer or
          Message 4 of 28 , Nov 9, 2008
          • 0 Attachment
            loro wrote:
            > I think you mean HTTP headers.

            Absolutely. It was way past midnight here, when I wrote that.

            > That doesn't mean the host made that happen.

            Well, someone who has control over the server. And that usually is not
            the customer or only to the very limited degree .htaccess allows.

            > Superscript 1 and 2, encoded the way they are, are not exclusive to
            > cp-1252, and that was what you were talking about. They aren't
            > "illegal", they just mean different things in ANSI and UTF-8.

            I have to admit to not being firm in UTF-8. I do know that (nearly?)
            everything that's in the upper 128 for other encodings is a two
            character sequence in UTF-8. And I have tried this: The validator said
            "illegal, no UTF-8" first and was satisfied when I overrode that with
            telling it "use cp-1252". I have not checked which characters were the
            offenders, but the ones that showed up wrong in the browser is a good
            guess IMHO.

            Axel
          • loro
            ... But you suggested Bob would use .htaccess. I d say declaring the charset is one of the most common uses people make of .htaccess. ... You are absolutely
            Message 5 of 28 , Nov 9, 2008
            • 0 Attachment
              Axel Berger wrote:
              > > That doesn't mean the host made that happen.
              >
              >Well, someone who has control over the server. And that usually is not
              >the customer or only to the very limited degree .htaccess allows.

              But you suggested Bob would use .htaccess. I'd say declaring the
              charset is one of the most common uses people make of .htaccess.

              >UTF-8. And I have tried this: The validator said
              >"illegal, no UTF-8" first and was satisfied when I overrode that with
              >telling it "use cp-1252".

              You are absolutely right. The W3C validator does do that now (while
              the WDG one does not). My bad.

              Lotta
            • Axel Berger
              ... We do know, Bob told us. He uses NoteTab and writes in his native Windows charset. Apart from that you re right of course, and I already said so. If Bob
              Message 6 of 28 , Nov 9, 2008
              • 0 Attachment
                loro wrote:
                > This is so backwards. If Bob can use .htaccess, he should of course
                > use it to make the server send the character encoding he prefers,
                > which may very well be UTF-8 for all we know.

                We do know, Bob told us. He uses NoteTab and writes in his native
                Windows charset.
                Apart from that you're right of course, and I already said so.
                If Bob can make the server send the correct HTTP headers, that's best.
                Only his provider can tell him that. The two providers I'm using (one is
                my own choice and with the other I'm webmaster for someone else) don't,
                but at least allow me to stop them sending the wrong ones.

                > Bob, just use the entities (¹ and ²) for now and be done
                > with it.

                That is a possibility. It is the one I use on my own site for maximum
                backwards compatibility and there I declare US-ASCII in step with what
                I'm actually doing.

                For the other site I made easy maintainabilty by others the priority and
                declare cp-1252, meaning that they can just type whatever their Windows
                computer allows them and need not bother about encoding.
                Unless you want to restrict yourself to the lowest common denominator on
                ideological grounds, like I do, that's the best choice. It means in
                essence "whatever you can type and display correctly in NoteTab, the
                server and browser will accept and display correctly too."

                > You can read up about character encoding when you feel up to it.

                Bob, there really is not much to it. Most computers use a 255 character
                alphabet - I'm ignoring extensions like UTF for the moment. In all these
                the first 127 characters are identical and standardized by ASCII. The
                top 128 ones, your ä ö ü é ê € µ and so on, can be all over the place.
                This used to be more of a problem when the Macs, Ataris, Amigas DOS with
                cp-437, DOS with cp-850 and so on all had sizeable market shares. As
                long as you are using Windows and don't switch to cyrillic, greek,
                hebrew or something like that, everything you type and display will be
                encoded as cp-1252 (of which terms like AnsiNew and others are synonyms,
                but not ANSI, Latin-1 or ISO 8859-1). So if you go and tell that to the
                browsers rendering your pages, you'll be fine. If you don't, they or the
                server have to guess and may guess wrong. That's all there is to it.

                Axel
              • loro
                ... Axel, it s only the so called illegal range that s unique to the windows codepage. The rest, as the superscript characters at hand, are not. ... Trial and
                Message 7 of 28 , Nov 10, 2008
                • 0 Attachment
                  >We do know, Bob told us. He uses NoteTab and writes in his native
                  >Windows charset.

                  Axel, it's only the so called illegal range that's unique to the
                  windows codepage. The rest, as the superscript characters at hand, are not.

                  >If Bob can make the server send the correct HTTP headers, that's best.
                  >Only his provider can tell him that.

                  Trial and error works pretty well too. ;-)

                  >The two providers I'm using (one is
                  >my own choice and with the other I'm webmaster for someone else) don't,
                  >but at least allow me to stop them sending the wrong ones.

                  Do they let you use .htaccess but they don't let you use it to
                  declare the character encoding? That sounds strange and unusual indeed.


                  >encoded as cp-1252 (of which terms like AnsiNew and others are synonyms,
                  >but not ANSI, Latin-1 or ISO 8859-1). So if you go and tell that to the
                  >browsers rendering your pages, you'll be fine.

                  So he will with an iso latin charset.

                  I'll be quiet now. This doesn't lead anywhere and has very little to
                  do with Bob's question. Again, I'm sorry for this bickering. It
                  really wasn't my intention but that's how it turned out. I just
                  wanted Bob to get an answer to his question.

                  Lotta
                Your message has been successfully submitted and would be delivered to recipients shortly.