M
Michael
Does anyone have a link to any accurate stats for the current browser market
share?
share?
K
kchayka
Michael said:
no
N
Nico Schuyt
Michael said:
http://www.w3schools.com/browsers/browsers_stats.asp
but ... http://j3e.de/statistics_lie.html
Nico
J
John W.
B
brucie
no such thing, just lots of guesses.
W
Whitecrest
brucie01 said:
Sometimes and educated guess is the way to go.
L
Leif K-Brooks
Whitecrest said:
How is the guess educated, exactly?
M
Mark Parnell
Sometime around Wed, 10 Dec 2003 02:46:51 GMT, Leif K-Brooks is reported to
have stated:
By fake UA strings, of course.
have stated:
By fake UA strings, of course.
W
Whitecrest
Take a second and thing about it. If you can not figure it out, head
back to remedial statistics.
B
brucie
garbage in ---> garbage out
although in the case of internet statistics it seems more like
garbage in ---> gospel out
so there! put that in your pipe and smoke it!
D
DU
Michael said:
Rather than asking such question, you should have explained which
purpose would serve such info (if it was truly accurate). When I design
a webpage or do a website, I set myself a minimum common denominator. In
the last 2 years, I always set myself the goal of supporting browsers
and other web-aware applications/devices which could support well (did
not say entirely nor perfectly)
- HTML 4.01 elements and attributes,
- CSS1 properties and
- DOM 1 methods and attributes.
Roughly 96% (and still climbing) of all browsers in use out there do
support well HTML 4.01, CSS1, DOM1. (If 96% of all browsers in use out
there support HTML 4.01, CSS1, DOM1, then why would it be useful,
relevant to know the respective browser market shares? If I don't need
cross-browser code for 96% of browsers - different versions, different
browser manufacturers - then why would it matter to know their
respective browser market share?) As for the remaining 4%, I write very
clearly on my sites that people should upgrade their browsers. I test my
sites without javascript and then without CSS to see if content still is
accessible and if basic functionalities of the sites are rendered. I
test my sites with text browsers and check if content still degrade
(still can be accessed) in other user agents: MSN-TV viewer, Simply Web
2000 (text-to-speech browser and a rather good one), Delorie services
(colors, Lynx viewer) and that's it. With experience, I know better how
to make a site furthermore accessible and usable: that is what really
matters here.
In the final analysis, the real issue is about browsers and web-aware
applications, devices complying with web standards and source code of
webpages using, conforming to web standards. Simple as that.
FWIW, 85% of users out there use MSIE 5+.
5% - 8% of users out there use a Mozilla-based browser.
2% of users out there use an Opera based software or application.
Give or take 10% on these figures. That's what I think and I don't try
to convince anyone with these figures. They represent trends, rough
convergences among all stats available out there.
DU
B
Bertilo Wennergren
DU said:
No they don't. They support well those parts of HTML 4.01, CSS1 and DOM1
that web authors expect to work reasonably well. There are lots of other
parts of e.g. HTML 4.01 that don't work at all, except in some modern
minor browsers (Mozilla, Opera, Lynx...). But people tend to just forget
about those parts, since Explorer doesn't have any support for them.
Some of those things have been around since HTML 2.
There is also CSS2 with lots of interesting stuff that web authors
might want to use. A quick look at the browser statistics, and a
quick look at what Explorer does not support, can then be quite
sobering.
And there are lots of things in HTML 4.01 that don't work for them.
D
DU
Bertilo said:
Lots of other parts?
In the post I wrote, I clearly and distinctively mentioned this:
"(...) support well (did not say entirely nor perfectly)". Next time, I
would appreciate if you could explicitly mention that you snipped my
post and where you did it.
Nowhere in my post I am speaking of CSS2. So your comment here does not
really "interact" with my post. Of course CSS2 has useful and
interesting stuff. But my post was aiming at more basic, fundamental
matters. We're still at a web global picture where a lot people are
using table designs, nested tables, <font>, <img src="spacer.gif">,
insane amounts of document.write(), eval(), setTimeout(exp,1),
innerHTML, badly designed pages, etc.. All of this can be removed and
replaced by CSS1 properties and sound, better DOM 1 methods inside a
better designed page, with respect for user system resources.
According to W3C, it is estimated that 99% of all webpages out there (10
billions indexed or so) would fail validation. How many webpages out
there do trigger MSIE 6 for windows into standards compliant rendering
mode where this corrects CSS1 box model, where this reduces a lot the
need for cross-browser code?
I'm talking about big numbers here and a global picture.
Lots of things? Can you be more specific? How many? Which ones? Your
claim here; not mine.
I see a lot of green cells in this page:
http://www.robinlionheart.com/stds/html4/results.xhtml
and those who are not green do not need to be used often or can have a
suitable workaround.
The more complex and/or requiring a webpage is, the more you'll reach
the limits of a browser: I'm sure we'll agree on this. But for a very
wide majority of webpages out there (say 90%), you don't need CSS2
properties, rarely seen HTML 4.01 attributes, etc. That is what my post
basically was trying to say. And I'm not alone thinking and saying this
out loud. E. Meyer, R. Lionhearth, P.P. Koch, www.westciv.com and sites
who do reviews and test browsers support and will support my opinion
here: roughly 96% of all browsers in use out there do support very well
commonly used HTML 4.01 elements and attributes, CSS1 properties, DOM1
methods.
This page
http://devedge.netscape.com/toolbox/tools/2001/feature-detection/
gives MSIE 6 for windows good score for DOM 1 HTML, CSS1, DOM Core 1.
(N.B.: border-spacing is a CSS2 property, not a CSS1 property). And
again here, I don't claim that MSIE 6's support for DOM 1 is impeccable,
flawless, perfect: I do claim it support well, even very well DOM 1
methods and attributes.
I remember seeing a webdesigner testing 12 different webpage designs
with 6 recent browsers (I can't find that page anymore): if I remember
right, all of them except 1 (ICab, I think) achieved 95%+ of accurate
rendering. I'm trying to find that page with the results now. It's a
recent test, not more than 3 months old.
DU
T
Toby A Inkster
DU said:
The <link /> element springs to mind. As does <abbr />.
Support for @cite of <q/> and <blockquote/> is also lacking.
D
DU
Toby said:
Yes, this one missing is rather annoying; it's an HTML 2.0 element on
top of that.
You can workaround this with a
<span title="Description of abbreviation" style="cursor:help,
Agreed. But this can be worked around too and is not as fatal, damaging
or major as using 500 in a page markup instead of using properly
CSS1 margin and padding properties. Or as damaging as nested tables,
over-constrained tables, 100 document.write()'s, improper nesting
everywhere, etc. The big problem on the web is no longer with
browsers/browser manufacturers and their compliance to HTML 4.01, CSS1,
DOM1 but rather the poorly authored webpage, outdated markup code (like
<font>, <center>, <layer>), disconnected-disfunctional webpage designs,
"if (document.all) ... if(document.layers)" code in millions of pages.
If a very large majority of browsers now support
document.getElementById, then testing for/supporting document.all and
document.layers is no longer needed.
It's not enough to have people upgrade their browsers; upgrading the
code (and coding techniques) that these browsers support is also needed.
DU