string and objects

[Guest]

Forgive me here as I am not really sure what i am asking.

I have a range of anchors with "id" from "er0" to "er250"

now I have some script that positions a layer next to these anchors,
this script works well when i specify the variable like
var aPos = er234
this works perfectly

however the code that I am using is arriving as a string so I am
getting
var aPos = 'er234'
which fails, "is null or not an object"!

I am guessing here that
er234
is an object and
'er234'
is a string

how do i convert the string into an object??

 

[Jeremy J Starcher]

Forgive me here as I am not really sure what i am asking.

I have a range of anchors with "id" from "er0" to "er250"

now I have some script that positions a layer next to these anchors,
this script works well when i specify the variable like var aPos = er234
this works perfectly

however the code that I am using is arriving as a string so I am getting
var aPos = 'er234'
which fails, "is null or not an object"!

I am guessing here that
er234
is an object and
'er234'
is a string

how do i convert the string into an object??

Without seeing code, I can only take a few guesses.

If you are running in IE, that particular browser has the mis-feature of
creating variables for (IIRC) named and ID'd elements.

The cross-browser and correct way is to use document.getElementById
(string).

 

[Guest]

Without seeing code, I can only take a few guesses.

If you are running in IE, that particular browser has the mis-feature of
creating variables for (IIRC) named and ID'd elements.

The cross-browser and correct way is to use document.getElementById
(string).- Hide quoted text -

- Show quoted text -

here is the function
z arrives as a string
I have got the whole thing working with the "eval" statement, still
think there could be a beter way to turn the string nto an object

function conectAnalyse(z){
var el = eval(z)
if (el !== null) {
var xVal = el.offsetLeft;
var yVal = el.offsetTop;
var oP = el.offsetParent;
var pN = el.parentNode;

while (oP !== null){
xVal += oP.offsetLeft;
yVal += oP.offsetTop;

if (oP != document.body && oP != document.documentElement){
xVal -= oP.scrollLeft;
yVal -= oP.scrollTop;
}
//firefox adj
if (ff){
while (oP != pN && pN !== null){
xVal -= pN.scrollLeft;
yVal -= pN.scrollTop;
pN = pN.parentNode;
}
}
pN = oP.parentNode;
oP = oP.offsetParent;
}
}
document.getElementById("lnk").style.top = yVal + 12 + "px"

}

 

[Thomas 'PointedEars' Lahn]

[...]
however the code that I am using is arriving as a string so I am getting
var aPos = 'er234'
which fails, "is null or not an object"!
I am guessing here that
er234
is an object and
'er234'
is a string
how do i convert the string into an object??

Without seeing code, I can only take a few guesses.

If you are running in IE, that particular browser has the mis-feature of
creating variables for (IIRC) named and ID'd elements.

The cross-browser and correct way is to use document.getElementById
(string).

here is the function
z arrives as a string
I have got the whole thing working with the "eval" statement, still
think there could be a beter way to turn the string nto an object

function conectAnalyse(z){
var el = eval(z) ^^^^^^^^^^^^^^^^
if (el !== null) {
var xVal = el.offsetLeft;
[...]

This madness is the reason why it does not work in many non-IEs: there is no
`er234' etc. property, and the result of untested access to it is a
ReferenceError.

Do as Jeremy said, use the document.getElementById() method instead of
eval(). Also use document.all[...] as a fallback for the former (necessary
only for IE 4 and older).

Because of the universal wrapper method required for providing the fallback,
you are better off with

function conectAnalyse(el)
{
if (el)
{
...
}
...
}

Then you can simply pass the return value of the wrapper method for the `el'
argument. See also <http://PointedEars.de/scripts/dhtml.js>, dhtml.gEBI() &
friends.

Please get a real name.


PointedEars

 

[Guest]

spamme wrote:
[...]
however the code that I am using is arriving as a string so I am getting
var aPos = 'er234'
which fails, "is null or not an object"!
I am guessing here that
er234
is an object and
'er234'
is a string
how do i convert the string into an object??
Without seeing code, I can only take a few guesses.
If you are running in IE, that particular browser has the mis-feature of
creating variables for (IIRC) named and ID'd elements.
The cross-browser and correct way is to use document.getElementById
(string).

here is the function
z arrives as a string
I have got the whole thing working with the "eval" statement, still
think there could be a beter way to turn the string nto an object

function conectAnalyse(z){
   var el = eval(z)

        ^^^^^^^^^^^^^^^^

   if (el !== null) {
           var xVal = el.offsetLeft;
           [...]

This madness is the reason why it does not work in many non-IEs: there isno
`er234' etc. property, and the result of untested access to it is a
ReferenceError.

Do as Jeremy said, use the document.getElementById() method instead of
eval().  Also use document.all[...] as a fallback for the former (necessary
only for IE 4 and older).

Because of the universal wrapper method required for providing the fallback,
you are better off with

  function conectAnalyse(el)
  {
    if (el)
    {
      ...
    }
    ...
  }

Then you can simply pass the return value of the wrapper method for the `el'
argument.  See also <http://PointedEars.de/scripts/dhtml.js>, dhtml.gEBI() &
friends.

Please get a real name.

PointedEars- Hide quoted text -

- Show quoted text -

It all works well, thankyou.
But, is my understanding of it correct ?

function conectAnalyse(z){
var el = document.getElementById(z)
if(el){

If the browser supports getElementbyId, "el" becomes the correctly
named object and the code is sucessful if it dosnt support
getElementById then "el" remains null, the code does not run and no
error is created ?

Browser compatability is a bit of an unknown for me, I only have IE6
and FireFox to test on, is there any online utility that can show what
platforms script will run on ???

 

[Thomas 'PointedEars' Lahn]

Do as Jeremy said, use the document.getElementById() method instead of
eval(). Also use document.all[...] as a fallback for the former (necessary
only for IE 4 and older).

Because of the universal wrapper method required for providing the fallback,
you are better off with

function conectAnalyse(el)
{
if (el)
{
...
}
...
}

Then you can simply pass the return value of the wrapper method for the `el'
argument. See also <http://PointedEars.de/scripts/dhtml.js>, dhtml.gEBI() &
friends.

Please get a real name.
[...]

It all works well, thankyou.
But, is my understanding of it correct ?

Not quite. You should *pass* the return value of the wrapper method
to conectAnalyse() for the single argument of conectAnalyse():

conectAnalyse(gEBI("foo"));

function conectAnalyse(z){
var el = document.getElementById(z)
if(el){

Then you can make `el' the name of the argument, and lose the superfluous
`z' along with the first line of conectAnalyse().

If the browser supports getElementbyId, "el" becomes the correctly
named object and the code is sucessful if it dosnt support
getElementById then "el" remains null, the code does not run and no
error is created ?

No, if document.getElementById() is not supported, calling anyway throws a
TypeError exception (or worse). That is why you need a wrapper method that
determines if that feature is available an if not, uses another feature or
returns a false-value (optimally here: null). In its simplest, most
inefficient and most error-prone form:

function gEBI(id)
{
if (document.getElementById)
{
return document.getElementById(id);
}
else if (document.all)
{
return document.all[id];
}

return null;
}

I have already referred you to my dhtml.js, which uses a more sophisticated,
most efficient, and least error-prone approach. UTSL.

Browser compatability is a bit of an unknown for me, I only have IE6
and FireFox to test on, is there any online utility that can show what
platforms script will run on ???

I know of none.

Please trim parts you are not referring to (and reply using Google Groups'
"More options" Reply instead if you must use GG), and get a real name.


PointedEars

 

[Laurent vilday]

kangax :

Thomas 'PointedEars' Lahn :

I have already referred you to my dhtml.js, which uses a more
sophisticated,
most efficient, and least error-prone approach. UTSL.

A bit off-topic here, but looking at your dhtml.js [1]
[1] http://pointedears.de/scripts/dhtml.js

Don't get too worry about what you can found in this file. This dhtml.js
file (among most, if not all, of the others files at this german URL)
must *not* be used or even looked after.

It is :

- full of useless documentation (I think I get it, this stuff is totally
(C) 2002-2008 Lans, he can keep it anyway)

- provided with the viral GPL licence... sigh.

- full of typos (+ + in ligne 365, or using "," when ";" *is* needed in
line 1188, but many more all over every required files)

- relying sometimes on automatic semi colon insertion and is missing
some open and close braces. A *very* bad practice from a library
perspective.

- not coherent with it's use of dot notation, sometimes yes and
sometimes nope.

- providing *NOT* tested code (for instance the getFirstChild() function
is funny, it will not give the same result *at* *all* with different
user agent, the proof at the end [1])

- linking to many 404 (http://pointedears.de/scripts/JSdoc/ for instance)

- using a *LOT* of globals functions, which is once again not a very
good practice from a library perspective (globals from dhtml.js = DHTML,
DHTMLException, _addEventListener, _addEventListenerCapture,
_replaceEventListener, addOption, createElement, dhtml,
disableElementGroup, disableElements, display, getAbsPos, getAttr,
getCheckedRadio, getCont, getElem, getElementsByTabIndex, getFirstChild,
getParent, getStyleProperty, getTextContent, hasStyleProperty, hoverImg,
loadScript, removeOptions, selectRadioBtn, setAttr, setCont,
setStyleProperty, setTextContent, setValue, visibility, visible)

- full of very funny tests like this one for instance :
this.isIE4DOM = typeof document.all == "object" && !this.isOpera;
this.IE4DOM = 2;
this.DOM = this.supported && (this.isIE4DOM && this.IE4DOM);

- using "x == null" or "x == 0" when the author keeps arguing here it
should be "x === null" and "x === 0" (it is a funny "do what I say not
what I do" situation)

- ...

And so on all over every files, types.js and object.js, which are
required, got their load of funny javascript errors and typos too (the
not filtered "for in" are so funny to remember when I hear the author
yelling about it on others people).

Well in my opinion - which worth nothing I know it very well - noone
should use or even look at it, this utter crap *never* used in real
situation. And it is absolutly not a surprise to me !

[1] Here is the proof :

* IE :
- "YES"
- "YES"

* FF, Opéra and probably the rest of the world :
- "undefined"
- then got an error (o.firstChild is null in line 482 dhtml.js)

Doesn't looks very cross-browser to me.

The author of this weird javascript code is giving me good laugh
everytime he is sending someone to his library.

<!DOCTYPE
HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"
"http://www.w3.org/TR/html4/loose.dtd">
<html>
<head>
<title>Full of errors</title>
<script type="text/javascript"
src="http://pointedears.de/scripts/object.js"></script>
<script type="text/javascript"
src="http://pointedears.de/scripts/types.js"></script>
<script type="text/javascript"
src="http://pointedears.de/scripts/dhtml.js"></script>
</head>
<body>

<ul id="myul">
<li>YES</li>
</ul>

<script type="text/javascript">
var ul = dhtml.gEBI('myul');
var li = dhtml.getFirstChild(ul);
alert(li.innerHTML);
alert(dhtml.getCont(li));
</script>

</body>
</html>

 

[Thomas 'PointedEars' Lahn]

A bit off-topic here,

I don't think so.

but looking at your dhtml.js [1], why does `gEBCN` use `new RegExp("\\b"
+ s + "\\b");`? I'm sure you know that this matches wrong values (e.g.
"catching" class="foo-bar" elements when asked for class="foo" ones).

Good catch, I wasn't aware of that at the time. Will be changed in the next
release.

On a side note, is there a reason why native
`document.getElementsByClassName` is not used in clients where it is
available (and behaves as expected)?

document.gEBCN() is a fairly new feature, the implementation of the WHATWG
Web Applications 1.0 (HTML 5) Working Draft. It was not implemented at the
time the online version of the library was last updated (see the build
version). I became aware of it only a month or so ago, through postings
here. However, given that I have not done sufficient testing of it yet, and
the Working Draft status of its specification, I don't think it will be used
in the next release.


PointedEars

 

[Guest]

pointedEars wrote

Also use document.all[...] as a fallback for the former (necessary
only for IE 4 and older).

is this really nesisary, as far as i can tell not many people must use
this browser. In fact looking at the stats for my site, I have had NO
visitors with this browser over the past 12 months. (unless its the
unknown)

Looking at the list below which shows visits to my site over the past
12 months, is there any point in not just assuming that all browsers
are "getElementById" browsers. Seems to me that i would have to put in
a lot of effort to accomodate someone who may never visit.

Internet Explorer 6.x------157088
Internet Explorer 7.x-------42384
web bots--------------------15544
Firefox 2.0x-----------------7976
Mozilla/5.0------------------4952
Internet Explorer 8.x--------2136
(unknown) ---------------------912
Mozilla/4.0-------------------440
Firefox 3.1x------------------320
Fedora/2.0.0.18---------------280
Opera 9.x ---------------------264
Charlotte/1.1-----------------194
Firefox 1.5x------------------176
Konqueror 2.x-----------------144
Firefox 1.0x-------------------24
Opera 6.x ----------------------18
voyager/2.0--------------------16
Opera 7.x ----------------------13
PycURL/7.18.0-------------------5
SonyEricssonK800i---------------3
Charlotte/1.0t------------------2
Internet Explorer 5.5-----------2

the webbots are largely made up by yahoo slurp, presumably they
download all the pictures as well, what a waste of bandwidth, I may
configure some php to just let webots have a one byte image.

 

[Thomas 'PointedEars' Lahn]

pointedEars wrote
Also use document.all[...] as a fallback for the former (necessary
only for IE 4 and older).

is this really nesisary,
Depends.

as far as i can tell not many people must use this browser.

It is not the browser but the layout engine of it that counts. How can
you know in which applications the MSHTML Browser Component (MSHTML.dll),
version 4.0 is used?

In fact looking at the stats for my site, I have had NO
visitors with this browser over the past 12 months. (unless its the
unknown)

Looking at the list below which shows visits to my site over the past
12 months, is there any point in not just assuming that all browsers
are "getElementById" browsers. Seems to me that i would have to put in
a lot of effort to accomodate someone who may never visit.

Internet Explorer 6.x------157088
Internet Explorer 7.x-------42384

To begin with, your "statistics" contradicts a current, more reliable market
analysis which says that IE 7 has greater a market share than IE 6 by now:

<http://marketshare.hitslink.com/>
<http://marketshare.hitslink.com/browser-market-share.aspx?qprid=3>
<http://marketshare.hitslink.com/bro...&qpdt=1&qpct=4&qptimeframe=M&qpsp=108&qpnp=11>

If you really have more visitors with IE 6 than IE 7 (which is doubtful),
maybe it is because you have optimized your code for IE 6 and ignored the
changes in IE 7.

And, of course, all IEs have "Mozilla/4.0" in their User-Agent header.

web bots--------------------15544
Firefox 2.0x-----------------7976
Mozilla/5.0------------------4952

Firefox 2.0.x *is* a Mozilla/5.0 browser.

Internet Explorer 8.x--------2136

Still beta.

(unknown) ---------------------912
Pardon?

Mozilla/4.0-------------------440

That would be Netscape 4.x and might be other browsers that have
"Mozilla/4.0" in their User-Agent header but cannot be identified as a
specific browser.

Firefox 3.1x------------------320

Firefox 3.1.x is beta, but a Mozilla/5.0 browser.

Fedora/2.0.0.18---------------280
Opera 9.x ---------------------264

Opera can have "Mozilla/4.0" in its UA header.

Charlotte/1.1-----------------194
Firefox 1.5x------------------176

Firefox 1.5.x is a Mozilla/5.0 browser.

Konqueror 2.x-----------------144
Firefox 1.0x-------------------24

Firefox 1.0.x is a Mozilla/5.0 browser.

Opera 6.x ----------------------18
voyager/2.0--------------------16
Opera 7.x ----------------------13

See above.

PycURL/7.18.0-------------------5
SonyEricssonK800i---------------3
Charlotte/1.0t------------------2
Internet Explorer 5.5-----------2

the webbots are largely made up by yahoo slurp, presumably they
download all the pictures as well, what a waste of bandwidth, I may
configure some php to just let webots have a one byte image.

Your "statistics" of the past are obviously flawed, and also bear no meaning
regarding present or future use. STFW.

<http://PointedEars.de/scripts/test/whatami>


PointedEars

 

[dhtml]

kangax :

Thomas 'PointedEars' Lahn :

I have already referred you to my dhtml.js, which uses a more
sophisticated,
most efficient, and least error-prone approach. UTSL.

A bit off-topic here, but looking at your dhtml.js [1]
[1] http://pointedears.de/scripts/dhtml.js

Don't get too worry about what you can found in this file. This dhtml.js
file (among most, if not all, of the others files at this german URL)
must *not* be used or even looked after.

It is :

It is crap. Anyone with half a clue who visited that site already noticed.

You pointed out the obvious shortcomings in the crap that Thomas
attempts to pass off as code. This speaks to the nature of Thomas.

Thomas Lahn is hyper-critical of others and unable (or unwilling) to see
his own shortcomings.

Thomas Lahn tries to make the appearance that he is knowledgeable. He
does this by imparting knowledge he has accrued over the years and by
talking down others in his very frequent posts.

It just shows how worthless Thomas Lahn thinks he really is.

More from a post from 2006:
http://groups.google.com/group/comp.lang.javascript/msg/c81bd16ffd734a20?dmode=source

Garrett Smith

 

[Guest]

To begin with, your "statistics" contradicts a current, more reliable market
analysis which says that IE 7 has greater a market share than IE 6 by now:


PointedEars

Just looking at the past month IE6 62% IE7 9%

Maybe its more to do with the people that are looking at my site,
hillwalkers don't go in for new browsers who knows

If you really have more visitors with IE 6 than IE 7 (which is doubtful),
maybe it is because you have optimized your code for IE 6 and ignored the
changes in IE 7

I doubt people are taking one look, declaring this boy has made this
for IE6 and switch off in disgust!

 

[Guest]

That's easier done with a robots.txt file.

O' Explain more, that sounds interesting

stuart

 

[Thomas 'PointedEars' Lahn]

Internet Explorer 6.x------157088
Internet Explorer 7.x-------42384

To begin with, your "statistics" contradicts a current, more reliable market
analysis which says that IE 7 has greater a market share than IE 6 by now:
[...]

Just looking at the past month IE6 62% IE7 9%

Maybe its more to do with the people that are looking at my site,
hillwalkers don't go in for new browsers who knows

If your numbers were based on fact, for *now* that might be so. But imagine
what could happen if you designed your site usable for visitors with IE 7
too, let alone for visitors with other user agents. They don't need to get
all the eye-candy, I'm talking about pure accessibility (in double sense).

I doubt people are taking one look, declaring this boy has made this
for IE6 and switch off in disgust!

You miss the point. Visitors will seldom be experienced enough *why* it
does not work, only *that* it does not work. And they will blame you first,
not their program (of which they may not do anything about, especially in
the IE 6 vs. 7 case) and go on to the competition. That's why it's
(commercial) suicide to optimize a Web site for one browser or a subset of
browsers without apparent need.


PointedEars, with a fitting random signature

 

[Thomas 'PointedEars' Lahn]

kangax :

Thomas 'PointedEars' Lahn :
I have already referred you to my dhtml.js, which uses a more
sophisticated,
most efficient, and least error-prone approach. UTSL.
A bit off-topic here, but looking at your dhtml.js [1]
[1] http://pointedears.de/scripts/dhtml.js

Don't get too worry about what you can found in this file. This dhtml.js
file (among most, if not all, of the others files at this german URL)
must *not* be used or even looked after.

It is :

It is crap. Anyone with half a clue who visited that site already noticed.

Both my Web site and my script libraries are in part a bit outdated for
sure. (As you know, I am still working on the next release of dhtml.js,
personal and business life made it rather hard for me to keep pace last
year.) The method I use in dhtml.js and recommended to
(e-mail address removed) for solving (t)his problem, is still a good one, though.

You pointed out the obvious shortcomings in the crap that Thomas
attempts to pass off as code. This speaks to the nature of Thomas.

Thomas Lahn is hyper-critical of others and unable (or unwilling) to see
his own shortcomings.

As you can see in e.g. <I am *always*
open to *constructive* criticism. The rest goes to /dev/null, and so can
you. FOAD.


PointedEars

 

[Guest]

You miss the point.  Visitors will seldom be experienced enough *why* it
does not work, only *that* it does not work.  And they will blame you first,
not their program (of which they may not do anything about, especially in
the IE 6 vs. 7 case)

You have me paranoid now, maybe my stuff don't work in IE7 ! I only
have IE6 myself so wouldnt know!
Maybe somebody could go and try it for me. The only place where there
is any real scripting is on http://www.corbetteer.co.uk/munros/map/game/munrogame.php
and the associated leaderboard, it is on the leaderboard where you
have helped me in positioning the little green bar that links the
score to the table when you click on "analyse" (and very proud of it
I am) So thank You. I am sure you will blow a fuse if you look at my
scripting, it will undoubtably be full of bad practise and browser
incompatability! but hey, it works (just) and I have learned loads
writing it

thanks again

 

[Thomas 'PointedEars' Lahn]

You have me paranoid now,

However, that wasn't my intention. You should only be aware of the issue.

maybe my stuff don't work in IE7 ! I only have IE6 myself so wouldnt know!

(It's getting slightly off-topic, but back on-topic later

For successful Web authoring, I recommend you have at least a handful of
different user agents handy. Most of them can be downloaded for free, for
various operating systems, so that shouldn't be much of a problem.

If you have Windows XP or later, I suggest you update to IE 7 [1]; then you
can set up virtualization (e.g. Microsoft Virtual PC which can be downloaded
for free [2]) where you can still run IE 6 (IE 6 Standalone as provided by
the evolt.org archive [3], with IE 7 installed, seems to work for detecting
problems with script language features and the DOM, but it does not provide
the full layout engine [4]).

For another example, I am mostly on Debian GNU/Linux again [5], and even
though IE is rather Windows software, I can run the standalone versions 4.0
to 6.0 from the evolt.org archive through Wine, the Windows emulator [6], so
that I don't have to reboot to Windows XP.[7]

QEMU [8] is another free open-source possibility for emulation. (I had it
run my Linux system, on the other partition, in Windows XP before.[9])

Maybe somebody could go and try it for me. The only place where there
is any real scripting is on http://www.corbetteer.co.uk/munros/map/game/munrogame.php

I don't think native JScript objects will be much of a problem [10], but DOM
scripting, CSS, and standards compliance as a whole. In our company, we
frequently have to use stylesheet workarounds powered by
Conditional-Comments and style expressions for IE 6 in Web sites that are
otherwise standards-compliant and display properly in IE 7; sometimes
vice-versa.[11]

Also observe the changes in the security model in IE 7, especially if you
use SSL/TLS-encrypted HTTP connections as it can happen that a secure Web
site that displays fine in IE 6 won't display in IE 7; instead, a warning
will be displayed.[12]

BTW, in order to use your real name when posting with Google Groups, you
only have to subscribe to this newsgroup and update your profile
accordingly. BTDT.


HTH

PointedEars
___________
[1] <http://windowsupdate.microsoft.com/>
[2]
<http://www.microsoft.com/downloads/...familyid=04d26402-3199-48a3-afa2-2dc0b40a73b6>
[3] <http://browsers.evolt.org/?ie/>
[4] e.g. the workaround for transparent PNGs does not work then
[5] <http://www.debian.org/>
[6] <http://www.winehq.org/>
[7] <http://www.tatanka.com.br/ies4linux/page/Main_Page>
[8] <http://bellard.org/qemu/>
[9] <http://www.h7.dion.ne.jp/~qemu-win/>
[10] <http://pointedears.de/es-matrix/>
[11] <http://msdn.microsoft.com/en-us/library/ms537512.aspx>
[12] <http://blogs.msdn.com/ie/archive/2005/12/07/501075.aspx> pp.

 

[SteveYoungTbird]

For successful Web authoring, I recommend you have at least a handful of
different user agents handy. Most of them can be downloaded for free, for
various operating systems, so that shouldn't be much of a problem.

If you have Windows XP or later, I suggest you update to IE 7 [1]; then you
can set up virtualization (e.g. Microsoft Virtual PC which can be downloaded
for free [2]) where you can still run IE 6 (IE 6 Standalone as provided by
the evolt.org archive [3], with IE 7 installed, seems to work for detecting
problems with script language features and the DOM, but it does not provide
the full layout engine [4]).

For another example, I am mostly on Debian GNU/Linux again [5], and even
though IE is rather Windows software, I can run the standalone versions 4.0
to 6.0 from the evolt.org archive through Wine, the Windows emulator [6], so
that I don't have to reboot to Windows XP.[7]

IEs4Linux at http://www.tatanka.com.br/ies4linux/page/Installation have
Internet Explorer 6, 5.5, 5 and a Beta version of IE7 that run on Wine.
I have all including IE7 installed on my Ubuntu Hardy system and have
not had any problems testing javascript code with them. I wouldn't use
the to surf with mind you.

Regards, Steve.

 

[RobG]

'PointedEars'  wrote:




Just looking at the past month IE6 62% IE7 9%

Maybe its more to do with the people that are looking at my site,
hillwalkers don't go in for new browsers  who knows

It is also a bit suspect that some reasonably popular browsers are not
represented at all, such as Safari and Chrome, which should have 1% to
5% representation. Perhaps they are in the "Unkonwn" basket.

I doubt people are taking one look, declaring this boy has made this
for IE6 and switch off in disgust!

So do I, more likely your statistics software is miscounting
"visits". You might find the following article interesting:

<URL: http://virtuelvis.com/archives/2005/05/statistics-nonsense >

Incidentally, your sites seems to work in Safari 3.2.1, check your log
to see if my visit figures in your stats. It is pretty much unusable
in Mobile Safari and I suspect any other mobile browser.

Put your script into an external script file, then validate at:

<URL: http://validator.w3.org/ >

(putting the script in an external file will get rid of many spurious
validation errors so you can find the real ones).

 

[Jeremy J Starcher]

IEs4Linux at http://www.tatanka.com.br/ies4linux/page/Installation have
Internet Explorer 6, 5.5, 5 and a Beta version of IE7 that run on Wine.
I have all including IE7 installed on my Ubuntu Hardy system and have
not had any problems testing javascript code with them. I wouldn't use
the to surf with mind you.

I'm a big fan of IEs4Linux, even have their easter-egg version installed
(Hint: Use the Source, Luke).

However, I have run across one instance where it was using slightly
different DLLs than "Real" install of IE6. I /think/ the real version of
IE6 was behind a patch level for something, but its been long enough I
don't recall.

Its a valuable tool, but sometimes nothing beats a virtual machine and a
retail CD.