[theforum] movement for redesign

Morgan Kelsey morgan at morgankelsey.com
Wed Aug 13 15:30:44 CDT 2003


dear y'all,

> > At the risk of being dim, don't we do that already with
> > http://evolt.org and http://www.evolt.org?
> ><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
>
> there's a general assumption by most seo experts that google makes an
> exception for domain.com and www.domain.com.
>
> ><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

i'm going to disagree.
it sees them as different, and assigns pagerank accordingly is my
understanding.

it's best to try to focus on one or the other, otherwise the pagerank is
split between both.

i believe there is speculation that they may be trying to change this.
(there's lots of speculation)


but this is a foggy area, and i'm going to get back to judah's idea of using
both www. and work. addresses along the way....

firstly, i agree that we need a plan of action before attention to details
of this sort.

however, this seems to be an area where many people have different
opinions/understandings/feelings, but i think on the whole alot of people
 are in the dark as to what we're talking about with regards to
google and pagerank.

so....i'm going to try to be educational, because i think it's really
important.

i'm not trying to be evangelical about this or beat a dead horse into the
ground.

i'm going to try to state fact only, and leave my own opinions about google
out of it.

i'd just like people to have a clear understanding of some things before a
vote happens, and want to investigate some things here while the topic is
fresh.

for the i-wanna-be-SEO dept.
everyone industry-wide agrees that http://searchenginewatch.com/ has
reliable info. if you want to learn more, start there.

for the wtf is pagerank dept:
http://www.google.com/technology/index.html

for the how-do-you-know-the-PR (PR = pagerank) dept:
http://toolbar.google.com/
heh, subdomain....


ok, here we go:

the current evolt URL system allows for anything to be in the slot following
"article/" in the URL.

the aardvark's "entities" article:
(how come he get's the "the" anyway?)

1. the "deep" link, the master one, whatever:
http://www.evolt.org/article/A_Simple_Character_Entity_Chart/17/21234/index.html

2. also, available in the short form:
http://www.evolt.org/entities/

3. also, currently in the top 5:
http://www.evolt.org/article/rating/17/21234/index.html

4. also available as $foo:
http://www.evolt.org/article/what_sound_does_an_aardvark_make_when_eating/17/21234/index.html

current page rank for those:
1. PR6
2. PR0
3. PR0
4. PR0

now #3, is a little puzzling, because it is in the top 5. other articles in
the top5 have a PR of 7.
it only recently got there though, and that may be why its showing a zero.
if it stays in the
top5, it will likely become a 7 in the #1 and #3 forms, which is what the
other top 4
currently have.

this is *new* behavior.

up until a couple of months ago, every top 5 article had an 8, while the
long versions had 5s and 6s.

issac's usable forms article has been on top forever now, the link from the
top 5 bar leads to the version with /rating/ in the URL slot, that page has
a PR7. all the articles in the top 5 have PR7 (except the newest entry)

the long version of those links also have a 7.


so, it looks like they've figured out those are the same.


but looks can be deceiving.

it is a fact that the pagerank curve is logarhythmic.

a PR7 is a broad number.
i've seen guesstimates thats its between 5 and 10 times harder to achieve an
8 than a 7.

so, i think we have a high and low seven here.


looking a little deeper we'll see why:

the #6 ranked article is the evolt rebuild article:
http://www.evolt.org/article/rating/20/5816/index.html
PR3

long form:
http://www.evolt.org/article/Inside_the_evolt_org_Rebuild_The_HTML_and_CSS/20/5816/index.html
PR6

so what's up here?
where's that PR3 on the rating link coming from?

it used to be in the top5.

this is how we can tell that google doesn't see the pages as the same.

almost every page on evolt.org has the top 5 sidebar, thus giving pagerank
to
the rating version of the top 5 articles.

those articles would rank higher at google, if the link from the side bar
was to the long form (or some form, just one and not many)

there's another link on every page, the one to the ratings page itself:
(the [....more] in the top 5 ratings sidebar)
http://www.evolt.org/rating/index.html

the ratings page has a 7.

notice that is also re-lists the top 5.

the links on the ratings page point at the _long_version_
(keep in mind the links in the top 5 bar go to the /article/rating/ version)

#9, is my debugger article, which was in the top 5 until recently:
http://www.evolt.org/article/12_Steps_to_a_Bigger_Better_Debugger/17/41393/index.html
PR5

now, there is no active link anywhere for the ratings one.
but honk rating in the URL:
http://www.evolt.org/article/rating/17/41393/index.html
PR7

google hasn't updated those backlinks yet. when the PR for the current #5
article goes up, i'd expect the debugger's to go down.

ok, one more and i'll rest my case.

the next page in the ratings section:
http://www.evolt.org/rating/all/0/11/10/index.html
has PR5
(remember http://www.evolt.org/rating/index.html has a 7)
page 3 also has a PR5
page 4 has PR4
page 5 has PR4
page 6 has PR3

the articles those links go to also have descending PR.

keep in mind, incoming links from other sites play a factor here as well.
its possible some pages deep in the ratings have tons of other links coming
in. if those incomin links have high PR, that article will benefit. (you can
see the backlinks google counts with the google toolbar as well)

the first article on page 2, yet another mr. the aardvark
(the freaky armored raccoon that he is)
http://www.evolt.org/article/Liquid_Design_for_the_Web/20/15177/index.html
PR6

http://www.evolt.org/article/rating/20/15177/index.html
PR0
(may have never been in the top5, or its been long enough for the backlinks
to be updated)

so articles that are on the ratings pages, have higher rankings because
every page links to the ratings page, which in turn links to itself (pg 2 3
, etc, wich have less and less PR), and links to other articles.

i think this is good, really. the large number of evolt pages helps
democratically raise it's highest rated brethren in the search engines.



so, yes, we do aliasing now, and it works.

i think the variable URL slot is really cool, and i understand
and appreciate the many reasons for both starting it and maintaining it.

but--future evolt pages would have higher rankings if that stopped.


google clearly sees the addresses as two different pages (just like the www.
or no
www), which as jeff points out raises the spam flag. someone from google may
well have checked out evolt and flagged them as OK, but i doubt that too.

the thing is, the spam protection is built right into the google system.

google doesn't care because we're only hurting ourselves.

we are currently dividing the page rank that many pages would have between
multiple pages.

the pizza analogy is a good one:
we're inviting the URLs for dinner and putting a pagerank pizza on the
table. they divide the pizza up amongst themselves. if we had only one of
them over, it would get all the pizza.


so,

should we care?

are we risking catering to something that can change without our knowing?

i see it almost as a spider-usability issue--
spiders can't use our site well, they can't tell some pages are the same.

stupid spiders.


should we sacrifice a portion of our human usuability for the sake of our
spider friends?

i say yes to all 3.
i respect and appreciate those who disagree, its possible that SEO is not
what's most important to evolt.

i'm one of the newbies still around here too....i joined 2/9/2002.

there's alot of folks who were much more active then, that i don't see
around as much now.

they're busy making money instead of writing long, stupid emails. ;o)


i'm afraid .jeff is right though, if we don't attract fresh meat we'll
starve. google is where our meat comes from. (its how i found evolt too)


i think a single-URL system for future articles is the way to go.
existing links, shouldn't be messed with.


[zip]
ok, i'm done
[/zip]

[ouch /]


Morgan
http://morgankelsey.com




More information about the theforum mailing list