>>>I think it helped to a degree but not a large degree
>>>,... although some people feel SE's like tight pages [in terms of file
[quoted text clipped - 8 lines]
>
>That you know of, anyway.
When I've noticed an error I've always checked to see if it had an
impact. With small errors no noticeable impact, the one large error
resulted in the page not existing to search engines. So for those
pages it was very bad. It was the 80th page of a book, so not a big
deal to SEO, would be much more irritating for a visitor finding page
80 missing though!!
I recall another error I made on a site to over 1000 pages. I changed
a header from H3 to h4, but didn't change the </h3> to </h4> so
basically 90% of the page was after an open <h4>. Result was the SERPs
went up! I eventually fixed it and the SERPs continued to go up at
roughly the same rate (suggesting it made little to no difference).
>>Funny one I did that meant page 80 of every book resulted in a server
>>error because when I was creating the template I pasted a number in
[quoted text clipped - 3 lines]
>
>Hey - nothing disastrous then!
That wasn't a small error, that was a very big error. I think you'd
agree a server error is big, missing a > off a tag is small (the page
will still get indexed).
>>That site did better than a lot of very clean coded sites I've created
>>since. If Google considered this extra code in it's algo that site
[quoted text clipped - 10 lines]
>might have done had it been able to reassure itself it knew what it
>was indexing on the basis that the code was syntactically correct.
Yes, but I also know from experience search engines can to a degree
cope with small errors like these. If you did have <table for example
it might be treated as text. Not a big deal if you added a bit of text
saying <table. Same if you miss the same from a header <h1 you've
probably changed a header into a bit of body text, so it's not a major
disaster, but IF H1 headers confer extra benefit you just lost the
benefit. But the page will still be spidered.
>>My advice would be to validate your code so you know it's got no
>>mistakes just in case you do something that does result in a search
>>engine completely misunderstanding your site.
>
>which is my point, so how come you don't?
You ever tried validating 5000 pages in a day? Even 100 pages takes up
more time than I have.
>>You should add a Doctype
>>and charset unless the server adds one automatically (if in doubt add
[quoted text clipped - 3 lines]
>comes with no doc-type? I'm just asking here. I never heard of servers
>adding doc-types.
I didn't say the server adds a Doctype, you can have the charset dealt
with by the server.
>> This way not only will all the different versions of browsers be
>>able to view the site as intended,
>
>I really don't think it's that simple as they all display stuff
>different anyway without a zillion workarounds.
Yes, but if you miss your Doctype you just made it worse. Do you know
how your pages will be served to a visitor using a very old browser,
what happens if the browser guesses wrong? Why leave it to guess work
when a simple line of code will almost guarantee what browsers and
most likely search engine spiders will do.
>I'm sorely tempted to
>sally over to alt.html and tell'em "Hey! Dave's arse hanging in the
>breeze over here!".
I think they'd agree with me since most will see adding a Doctype as
the norm.
>Since you've already had your quota of nutters for this week I won't
>though.
So you'll be waiting until the end of the week then, thanks :-)
>> spiders will be able to understand
>>what your markup means.
>
>Not if it isn't syntactically correct they won't and hanging a
>doc-type and a charset on the walls won't help that.
If it's correctly coded (validates) then I believe it doesn't matter
how much markup you add to a page, search engines will only count
content and any markup they consider important (it's generally believe
headers falls under the above).
I'll quickly add I don't know for sure, it might be spiders completely
ignore the Doctype and charset.
I really should test it. I think I'll create two identical text pages,
but with one fill it with tables and fonts. Maybe font every word
until it hits 100KB!!
Hmmm, just noticed a flaw in your argument. The reason for not
including the Doctype, charset and removing as much markup as possible
is because you believe Google etc... counts everything in the code of
a page. Yes?
If you believe Google counts the above, wouldn't that suggest it might
have some value? If Google completely ignores it, it doesn't matter if
it's there or not, so as it helps with other things (like validating
and browsers displaying the page correctly) doesn't it make sense to
add a Doctype at least?
>BB
David

Signature
http://www.search-engine-optimization-services.co.uk/
C.W. - 30 Oct 2004 19:43 GMT
[snip]
>>>My advice would be to validate your code so you know it's got no
>>>mistakes just in case you do something that does result in a search
[quoted text clipped - 4 lines]
>You ever tried validating 5000 pages in a day? Even 100 pages takes up
>more time than I have.
I don't do all the pages in one day.
I validate the template that I will use for site ... I use HomeSite as
an HTML editor, even thouhg I hand code, and will click to validate
the page to make sure I don't have any errors in the new content added
to the template. Or I may use the Web Dev toolbar on Firefox to see if
the page validates HTML Strict with the new content added to the
already validated template.
But I will say "that's just me".
I also use DocType [doesn't add that much at all to the page and more
reasons for it than against it] and META keywords and descript on my
pages. Although I don't put a lot of value on META keywords out of the
two METAs. *shrug*
[snip]
Big Bill - 30 Oct 2004 21:17 GMT
>Hmmm, just noticed a flaw in your argument. The reason for not
>including the Doctype, charset and removing as much markup as possible
>is because you believe Google etc... counts everything in the code of
>a page. Yes?
I've been wondering about this for ages and have been working on the
basis of erring on the side of caution. I've had it inmind to ask you
and Will and separately the boys at Planet Oceran to run tests to
determine this for engines as I know you all enjoy that kind of stuff.
But, Google's going all weird again so we may well be seeing a new one
soon and MSN are threatening again to unleash their beta on a
permanent basis so now is not the time.
>If you believe Google counts the above, wouldn't that suggest it might
>have some value? If Google completely ignores it, it doesn't matter if
>it's there or not, so as it helps with other things (like validating
>and browsers displaying the page correctly) doesn't it make sense to
>add a Doctype at least?
I get no. 1 rankings without it so why bother?
But I think really we should be testing for this and every line of
code to determine what's diluting the S/N ratio and what isn't. I try
to allow for CSS as opposed to fonts etc. as a matter of course now
and I'm kind of umming and awwing about going to total css. Bar the
odd table. Maybe.
And I don't care about old browsers. If they're bleedin' Luddites
neither I nor my clients want to know them. It was different when I
was working for the Tibet people because then you had to think of
monks hiding out who only had access to Netscape 2 and the like, but
these days those days are the bad old days.
BB
www.kruse.co.uk
The home of SEO that's shiny!