[DesDev] weo,beo,leo,deo browser/monitor resolution stats

[.jeff]

beau,

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> From: Beau Hartshorne
>
> Should we be collecting "real world" browser/monitor
> resolution stats while we toss around designs for the
> new site?
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

possibly.  i'm not sure it'll really prove to be that useful though.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> Here's aardvark's article (watch the wrap -- can someone
> please tell me offlist how to make the evolt.org article
> links short?):
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

isaac already illustrated, but let me suggest one extra thing.

1)  remove the "www."
2)  remove the "index.html"
3)  swap out the article title for some short string
    that represents where the link to the article is
    coming from.  if i'm posting the link to another
    list i'll usually put the name of the list in
    place of the title.  presumably we could use that
    to track where traffic is coming from.

http://evolt.org/article/desdev/17/2295/

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> The JavaScript looks good to go, and I can rewrite that
> ASP stuff in PHP (useful for at least deo). I don't know
> my way around CFM (not yet anyway), so I wouldn't know
> how to make it work on the other evolt sites.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

again, i'm not in favor of bothering with doing this as i don't think the
stats will be very enlightening.  sure, we'll know resolution and window
sizes (for the browsers that can actually report it), but won't know font
sizes which is the other half of the equation.

i'd rather put effort in to designing the new site to support user-defined
font sizing at the very least.

><><><><><><><><><><><><><><><><><><><><><><><><><><><><><
> Another thing to think about would be adding a more
> robust browser sniffer so we can figure out how far we
> can go with xhtml and css. Since evolt is a community
> of web developers, I'd assume that on the whole we're
> ahead of the curve when it comes to screen size and
> browser version.
><><><><><><><><><><><><><><><><><><><><><><><><><><><><><

we can actually look to our traffic reports now for that information.

http://browsers.evolt.org/stats/stats.may2002.html

browsers:

71.44%: msie
12.83%: search engine spider
 9.21%: netscape
 2.40%: opera
 0.97%: netscape (compatible)
 0.36%: konqueror
 0.10%: avantgo
 0.09%: omniweb
 2.60%: everything else

operating systems:

75.39%: windows
13.20%: robots
 4.47%: macintosh
 3.63%: os unknown
 3.28%: unix
 0.03%: everything else

notice the large percentage of bots and spiders in the percentages above.
taking those figures completely out of the mix and adjusting the ratio for
the other browers and operating systems above, here's how the numbers would
look:

browsers:

81.95%: msie
10.56%: netscape
 2.75%: opera
 1.12%: netscape (compatible)
 0.41%: konqueror
 0.11%: avantgo
 0.10%: omniweb
 3.00%: everything else

operating systems:

86.85%: windows
 5.15%: macintosh
 4.16%: os unknown
 3.78%: unix
 0.06%: everything else

or, looking at the user agent report, of the top 15 reported (not counting
spiders and bots) about 80% are ie6/win.

i'm running some separate reports on all the sub-domains (weo, leo, beo,
deo) separately and will hopefully be able to give a better breakdown of
browser/os usage.

however, even without that i'd say we're more than safe going xhtml (which
does *not* mean table-less, all-div, 3-column bluerobot ripoff layout).

.jeff

http://evolt.org/
jeff at members.evolt.org
http://members.evolt.org/jeff/