HTML Forum

**Julia** · 05-13-2006, 02:25 PM

When I google my home page home-sweet-mexico, Google returns the following:

DOCTYPE HTML PUBLIC " - / / W 3 C / / DTD HTML 4 . 0 1 ... ... height : 8 0 px ; " alt = " Logo & quot ; home - sweet - mexico . com ... b > Live or Retire in Mexico store for home - sweet - mexico < / b > < / a ...
www.home-sweet-mexico.com/ - 35k - Cached - Similar pages
This is terrible because no human visitor will click on this!

I've just recently updated my site using Nvu. How can I fix this problem?

Julia

**mastercomputers** · 05-13-2006, 03:08 PM

Make sure to use:

<!DOCTYPE html PUBLIC ...

The "html" maybe case sensitive, since "html" is defining the root of your document <html>.

However I think the problem is the character encoding used, it seems by the additional spaces inserted between words shown above, is that it's not understanding that particular unicode encoding correctly, making the HTML elements actually appear as spaced out characters/words.

There might be other reasons too, but to be honest, I've never had this problem before.

Cheers,

MC

**Syndac** · 05-13-2006, 03:27 PM

Try viewing your source. I see nothing but a bunch of non-ascii characters that I can't even copy/paste. If I can't view it, chances are google can't either.

**mastercomputers** · 05-13-2006, 04:01 PM

Well Syndac proves that it's character encoding, so you should probably not use UTF-LE but either UTF-8 or UTF-16 depending on what character set you actually require.

Cheers,

MC

**chrishirst** · 05-13-2006, 05:11 PM

I would agree with the character set encoding causing the problem.

the other two pages that are indexed have no problem, they are set to ISO-8559-1

**Julia** · 05-14-2006, 04:22 PM

Thank you all for your helpful replies.

I just viewed my source inside the Nvu and deleted the EL. It went from 16EL to just 16. Then I re-published my page.

Would anyone mind explaining what I just changed? I can read html, but only when someone explains what it means. I really don't know anything and I'm just learning as I go along.

And the million dollar question: How will I know that Google has re-spidered my page so that I can tell if the problem is fixed or not?

Julia

**mastercomputers** · 05-14-2006, 06:17 PM

The character encoding is like a map of characters used to identify what your document characters could contain, it makes it easier for your browser's viewer to understand what they should use to read it. Usually unicode character range is to accept double byte characters and more, things like Japanese and Chinese characters. If any User Agent has a problem with reading the encoding, then you would see unreadable characters, which this seems to have been your case.

UTF-LE is the UTF-16 Little Endian, which basically is defining the byte order of representing those characters, you have UTF-LE (little endian) and UTF-BE (big endian). They are equivalent to UTF-16 except for byte ordering, which to me is quite pointless as I have not found a situation in which you must define the byte ordering like this. UTF-8 is what I mainly use as I still keep mainly to basic characters, it's just XML documents including family documents like XHTML default with using Unicode.

I'm not sure about User Agent support for UTF-16 but UTF-8 is quite widely supported and it's what I would recommend, usually people who know what character set to use would not have problems in choosing the right one. It just happens that Firefox supports a large range of character sets, so it was able to render your site, it seems IE does not support UTF-LE at all, which would explain the unknown characters being shown. It would seem that google also does not understand UTF-LE correctly.

Cheers,

MC

**mgraphic** · 05-14-2006, 08:16 PM

Google has your entir home page cached as text:
http://72.14.207.104/search?q=cache:...s&ct=clnk&cd=1

**mgraphic** · 05-14-2006, 08:20 PM

When I looked at your source code, every single characture is on a newline resulting in 14,694 lines.

**Julia** · 05-15-2006, 12:11 PM

Thank you. I am learning a ton.

MC, would you explain the character set ISO-8859-1? Some of my pages have that character set. Also, let me be sure that I got it right: you recommend using UTF-8, correct?

mgraphic, I opened the google page with the cache of my site. It appears to only show my home page. How can I see the other pages in my site?

It also shows it as continuous html (I don't see 14 thousand lines). How did you see that?

So to see if Google has crawled it again I just scan the html to see the character set, right?

Are you suggesting that I need to take some action so that each characture isn't on a new line? How do you look at source code for a page online?

How do you find the cache for a page? That is really useful.

So, I have more and more questions as we go along, but I need to learn all of these things.

Thank you both for your help!

**Minaki** · 12:33 PM

I think the reason you say 'Every character is displayed on a new line' is due to how whatever editor you're using displays the 2nd byte of Unicode characters.

(Correct me if I'm wrong here...)
Unicode characters are made up of 2 bytes, as opposed to one. (This makes sense for other languages such as Chinese and Japanese since they have a hell of a lot more characters than we do!) So when you open a Unicode document in an editor that doesn't support Unicode, it interprits the 2nd byte of Unicode as a normal byte. In Notepad on my system, it displays your code with a space in between every character.

**mgraphic** · 05-15-2006, 12:55 PM

The encoding - that HAS to be it. ?

**mastercomputers** · 10:29 AM

Hi Julia,

ISO-8859-1 is quite a popular character sets too and usually the default character language for many English programs, operating systems, etc.

It's easier to refer this to the Western (West European) character set, which you could imagine is only the West European characters, including a few accented characters, so there's no 2 byte characters like that used in the Japanese/Chinese language.

If you are however working with XML/XHTML, you're expected to use Unicode's UTF and yes I do recommend UTF-8.

Just a little history to explain character sets. We first had ASCII which had 128 characters because of the limited characters it was then extended and now held 255 characters, this is known as Extended ASCII. Now you can then imagine that no matter how many characters they came up with, it was never enough for other people/countries and soon you had different parts of the world coming up with their own set of characters for their own language, in which some of them became ISO standards. This was when they realised it was getting out of hand, so Unicode came to the rescue, in which they were to create a character set that was to hold every mentionable character, though this has not yet been reached. It is however proving quite effective being able to just use one of it's character set and have it support numerous characters from around the world, no need for different encodings, just one encoding for everything, this is what makes things standard.

As for mgraphic, that's due to the extra byte to form a 2 byte character. Our normal set of characters are only 1 byte per character, but in Unicode they are 2 bytes and to achieve this, another byte is added to the character, which most likely is NULL, \x00 or however you want to represent it, which does not alter the character you'd get, but does increase document size, since you're now using 2 bytes per character instead of 1. So it seems your editor interprets NULL characters like a newline character (\x0A), many old editors would do this because you're not suppose to be able to type a NULL character so it's just something that it never expected to see in the document. NULL in some programming languages also meant the end of a string.

If you want to save document size, then use ISO character sets, they're standard but not recommended anymore. You can also save document size by using on the fly compression.

Cheers,

MC

**Julia** · 05-16-2006, 12:44 PM

MC,
THANKS! It is all very clear now. So I think I'll just go into the html source for my page and change it to UTF-8 on your recommendation.

Thanks again.
Julia

PS How long does it usuall take Google to re-spider?

**chrishirst** · 05-16-2006, 03:25 PM

well if you post a proper website and NOT an click through affiliate link maybe we would look !!!!!!

**mgraphic** · 05-16-2006, 03:41 PM

Get 'em Chris, get 'em

**mastercomputers** · 03:28 AM

Hey Julia,

Even though you've changed the meta tag, that's a start, however, you must also change how the document is saved. Which means setting it's encoding to UTF-8 too when being saved, otherwise the document is still considered UTF-LE and that auto-detection done by the browsers will usually choose the appropriate character set to use, rather than relying on your meta tag.

In NVU you select Save and Change Character Encoding in the File Menu and you select UTF-8. Once you've done this, then ask users of IE to view your site and source code and see if it displays correctly.

As for how long google takes to respider, can not easily be said, some sites google visits each day, others can be weeks/months, etc. Depends on many factors related to your website and traffic that google directs to that site.

Cheers,

MC

**Julia** · 05-19-2006, 05:06 PM

Chris,
I don't understand your comment. I do have a proper website. It can be found at www.home-sweet-mexico.com
Julia

**chrishirst** · 05-19-2006, 06:19 PM

Julia, the comment isn't about your site. There was an affiliate link dropping spammer with a useless post that has since been deleted in the thread.

The comment was aimed at them. Apologies for the inconvenience

**Julia** · 05-20-2006, 02:33 PM

MC, Thanks! I'm glad you thought to tell me about changing it in Nvu. I sure wouldn't have.
Julia

HTML Forum

Find a Professional HTML Freelancer!

FREE Outsourcing eBook!