IE5/6, but not IE4, doing unexpected

[Dr John Stockton]

In <URL:http://www.merlyn.demon.co.uk/estrdate.htm#T3> the button "Table
III Reversed" executes

function BCPTableIII() { var GN, DM, Ro, CY
document.write("\t\t\tCalculated Table III",
"<fieldset>\t\t\t THE GOLDEN NUMBERS\t\t SL P F M\n")
for (GN=19 ; GN ; GN--) document.write(Prfx(GN, 3, " "))
document.write("<hr>\n")
for (Ro=DM=21 ; Ro<=51 ; Ro++, DM++) {
for (GN=19 ; GN ; GN--) {
CY = (11*GN + DM + 6) % 30
if (Ro==49 && GN==11) { Ro++
document.write("\t\t\t\t ", RightCols(DM-1), "\n\t\t\t")}
document.write(Prfx(CY, 3, " ")) }
if (DM==50) DM--
document.write(RightCols(DM), Ro%5==0 ? "<hr>\n" : "\n" ) }
document.write("<\/fieldset>") }

in which the output of RightCols is not germane, and we have

function Prfx(Q, L, c) { var s = Q+"" // ??
// if (!c) var c = ' '
if (c.length>0) while (s.length<L) { s = c+s } ;
return s }

and the output (chopped top, bottom, and right) is in IE4 as expected

28 17 6 25 14
29 18 7 26 15
0 19 8 27 16
---------------
1 20 9 28 17
2 21 10 29 18
3 22 11 0 19


However in a later IE, the first or second space of a row is omitted in
rows where the first column of numbers is single-digit. There's no
problem with other single-digit entries, as you can see.

28 17 6 25 14
29 18 7 26 15
0 19 8 27 16
---------------
1 20 9 28 17
2 21 10 29 18
3 22 11 0 19


I cannot easily experiment, as the IE>4 machine is a mile away.

What's the explanation, and the suggested fix?

 

[cwdjrxyz]

In <URL:http://www.merlyn.demon.co.uk/estrdate.htm#T3> the button "Table
III Reversed" executes

function BCPTableIII() { var GN, DM, Ro, CY
document.write("\t\t\tCalculated Table III",
"<fieldset>\t\t\t THE GOLDEN NUMBERS\t\t SL P F M\n")
for (GN=19 ; GN ; GN--) document.write(Prfx(GN, 3, " "))
document.write("<hr>\n")
for (Ro=DM=21 ; Ro<=51 ; Ro++, DM++) {
for (GN=19 ; GN ; GN--) {
CY = (11*GN + DM + 6) % 30
if (Ro==49 && GN==11) { Ro++
document.write("\t\t\t\t ", RightCols(DM-1), "\n\t\t\t")}
document.write(Prfx(CY, 3, " ")) }
if (DM==50) DM--
document.write(RightCols(DM), Ro%5==0 ? "<hr>\n" : "\n" ) }
document.write("<\/fieldset>") }

in which the output of RightCols is not germane, and we have

function Prfx(Q, L, c) { var s = Q+"" // ??
// if (!c) var c = ' '
if (c.length>0) while (s.length<L) { s = c+s } ;
return s }

and the output (chopped top, bottom, and right) is in IE4 as expected

28 17 6 25 14
29 18 7 26 15
0 19 8 27 16
---------------
1 20 9 28 17
2 21 10 29 18
3 22 11 0 19


However in a later IE, the first or second space of a row is omitted in
rows where the first column of numbers is single-digit. There's no
problem with other single-digit entries, as you can see.

28 17 6 25 14
29 18 7 26 15
0 19 8 27 16
---------------
1 20 9 28 17
2 21 10 29 18
3 22 11 0 19


I cannot easily experiment, as the IE>4 machine is a mile away.

What's the explanation, and the suggested fix?

Since I have several browsers installed I checked your mentioned page
and button on several. My SBC/Yahoo DSL browser(just a modified IE6)
shows the placement problems you mention. The latest Opera displays the
table correctly. However the latest Mozilla family browsers(Mozilla,
Firefox, and Netscape) only give some of the top headings when you
click the button. The progress bar halts at about 1/2 way, and does not
move and output the bulk of the table, at least for the 3 minutes I
waited. Netscape 4.8 refuses to download your page at all and gives an
alert with the message "File include1.js end fault, RSVP!". When you OK
the alert, the browser goes down. This likely is of little importance
since the Netscape 4.8 is all but dead in most countries.

Your page is fairly complicated and you use a style considerably
different from what I use, so I do notnow have time to attempt to
determine exactly what is going on. I now do nearly all positioning
with CSS , often using script to calculate the CSS positioning as I did
in the full year calendar display output for my perpetual calendar. In
cases that would require a document.write in javascript, I now have to
do that portion of the code in php on the server, as document.write can
not be processed in a true html 1.1 page served as
application/xhtml+xml, which is parsed as XML. A php include at the top
of the php page converts the code to html 4.01 strict for outmoded
browsers, such as the IE6, that will not accept the mentioned mime
type.

Of course, you only mentioned IE browsers, so what your page does on
other browsers may or may not be of interest to you.

 

[cwdjrxyz]

However the latest Mozilla family browsers(Mozilla,
Firefox, and Netscape) only give some of the top headings when you
click the button. The progress bar halts at about 1/2 way, and does not
move and output the bulk of the table, at least for the 3 minutes I
waited.

I have a screen capture for you at
http://www.cwdjr.info/temp/table3.jpg that shows exactly where output
of table 3 stops on the Firefox browser. Perhaps this will help you
narrow the portion of the code that may be causing the problem.

 

[cwdjrxyz]

The latest Opera displays the table correctly.

There is a screen capture using the latest Opera browser at
http://www.cwdjr.info/temp/table3Opera.jpg . Part of the table is
off-screen at the bottom. I see that I even managed to capture the
cursor . This view looks normal to me, but you know what you wanted.

 

[Dr John Stockton]

JRS: In article <[email protected]>
, dated Sat, 15 Oct 2005 22:14:14, seen in

Since I have several browsers installed I checked your mentioned page
and button on several. My SBC/Yahoo DSL browser(just a modified IE6)
shows the placement problems you mention. The latest Opera displays the
table correctly. However the latest Mozilla family browsers(Mozilla,
Firefox, and Netscape) only give some of the top headings when you
click the button. The progress bar halts at about 1/2 way, and does not
move and output the bulk of the table, at least for the 3 minutes I
waited.

None of the script seems likely to hang the system; the loop structures
are quite simple.

Netscape 4.8 refuses to download your page at all and gives an
alert with the message "File include1.js end fault, RSVP!".

Well, you've done the RSVP. That message means that include file
include.js is only partially loaded (if it had not loaded at all, you'd
have got a 'top fault').

The only buttons that should take more than a second or so on any PC
less than a decade old are the Timing ones, which have warnings, and the
tests when over a big range of years.

When you OK
the alert, the browser goes down. This likely is of little importance
since the Netscape 4.8 is all but dead in most countries.

Your page is fairly complicated and you use a style considerably
different from what I use, so I do notnow have time to attempt to
determine exactly what is going on. I now do nearly all positioning
with CSS , often using script to calculate the CSS positioning as I did
in the full year calendar display output for my perpetual calendar. In
cases that would require a document.write in javascript, I now have to
do that portion of the code in php on the server,

That means that the page cannot be executed independently of the Net,
which I want to avoid.

as document.write can
not be processed in a true html 1.1 page served as
application/xhtml+xml, which is parsed as XML. A php include at the top
of the php page converts the code to html 4.01 strict for outmoded
browsers, such as the IE6, that will not accept the mentioned mime
type.

Of course, you only mentioned IE browsers, so what your page does on
other browsers may or may not be of interest to you.

Well, I'd like it to be OK in all, or all after Version 4 level.

The output should look essentially, but not exactly, like that in the
Church of England Book of Common Prayer and in the (British) Calendar
Acts Annexes, and in Halsbury's Laws of England - but reversed left-to-
right. For those without these books (I only have the first), there's a
pure-HTML (non-reversed) version in <URL:
http://www.davros.org/misc/easter.html>, headed TABLE III.


I've now added code for a dot at the beginning of each line, and other
dots where necessary to maintain alignment. If that looks good in IE
5/6, I intend to replace them by   or actual spaces.

.. 28 17 6 25 14
.. 29 18 7 26 15
.. 0 19 8 27 16
----------------
.. 1 20 9 28 17
.. 2 21 10 29 18
.. 3 22 11 0 19

.... ...

I tried saving the generated HTML and using TIDY on it. There is
objection to <pre>...<fieldset> so I've changed it (in Table IIIa) to
<fieldset><pre> which could help. Alas, TIDY now complains about <hr>
in <pre>, but changing to </pre><hr><pre> spreads the layout vertically
- changing <hr> to a blank line is OK. Other Tables also dealt with.


Thanks for looking.

 

[cwdjrxyz]

I tried saving the generated HTML and using TIDY on it. There is
objection to <pre>...<fieldset> so I've changed it (in Table IIIa) to
<fieldset><pre> which could help. Alas, TIDY now complains about <hr>
in <pre>, but changing to </pre><hr><pre> spreads the layout vertically
- changing <hr> to a blank line is OK. Other Tables also dealt with.

Your revisions have helped on IE6 so that it now seems to display
correctly. However Netscape 4.8 still has the same problem, and the 3
Mozilla family browsers have the same problem, other than there are a
few dots in the few headings for table 3 generated when one version of
the button is clicked. Recent versions of Mozilla family browsers are
very strict compared with IE6, will not tolerate nearly as much older
code as IE6 will, and often will not attempt to correct something as
does IE6. It is beginning to appear that there may be two problems.
Your changes may have fixed the IE problem. However the Mozilla family
problem likely is caused by something quite different.

 

[Randy Webb]

Dr John Stockton said the following on 10/16/2005 5:24 PM:

JRS: In article <[email protected]>
, dated Sat, 15 Oct 2005 22:14:14, seen in


None of the script seems likely to hang the system; the loop structures
are quite simple.

But it does. In Firefox 1.0.7 the main page
<URL: http://www.merlyn.demon.co.uk/estrdate.htm#T3 >
gives an error in the JS console that
"Tx2 is not defined" line 1362.
Clicking on the "Table III Reversed" button opens a second page that
hangs and the Console reports:
"Prfx is not defined" line 1417.

 

[Dr John Stockton]

Earlier :

I have a screen capture for you at
http://www.cwdjr.info/temp/table3.jpg that shows exactly where output
of table 3 stops on the Firefox browser. Perhaps this will help you
narrow the portion of the code that may be causing the problem.

That's OK so far as it goes; the next line requires function Prfx which
I think your FireFox failed to get to work. Prfx should have been shown
in a box just above H4 "Page Index to Part of the Prayer Book", higher
in the page. Prfx does have a loop, but it should execute no more than
twice here.

function Prfx(Q, L, c) { var s = Q+"" // ??
// if (!c) var c = ' '
if (c.length>0) while (s.length<L) { s = c+s } ;
return s }



Earlier :

There is a screen capture using the latest Opera browser at
http://www.cwdjr.info/temp/table3Opera.jpg . Part of the table is
off-screen at the bottom. I see that I even managed to capture the
cursor . This view looks normal to me, but you know what you wanted.

That's just as it should be. Lower down, the line that starts 24 13
should have a step down near the middle, and the last line of data
should be
18 25 14 3 22 11 0 19 8 27 16 5 24 13 2 21 10 29 18 7 C Apr 18


JRS: In article <[email protected]>
, dated Sun, 16 Oct 2005 17:13:51, seen in

Your revisions have helped on IE6 so that it now seems to display
correctly. However Netscape 4.8 still has the same problem, and the 3
Mozilla family browsers have the same problem, other than there are a
few dots in the few headings for table 3 generated when one version of
the button is clicked. Recent versions of Mozilla family browsers are
very strict compared with IE6, will not tolerate nearly as much older
code as IE6 will, and often will not attempt to correct something as
does IE6. It is beginning to appear that there may be two problems.
Your changes may have fixed the IE problem. However the Mozilla family
problem likely is caused by something quite different.

That will call for testing modifications in Mozilla.



<FAQENTRY> It would be useful to have a list of the names of the browser
families, and their salient members in each case.

 

[Dr John Stockton]

JRS: In article <[email protected]>, dated
Sun, 16 Oct 2005 23:56:18, seen in Randy Webb

Dr John Stockton said the following on 10/16/2005 5:24 PM:

But it does. In Firefox 1.0.7 the main page
<URL: http://www.merlyn.demon.co.uk/estrdate.htm#T3 >
gives an error in the JS console that
"Tx2 is not defined" line 1362.
Clicking on the "Table III Reversed" button opens a second page that
hangs and the Console reports:
"Prfx is not defined" line 1417.

Giving an error message is not hanging the system.


Tx2 is the textarea in the Table II section. Its only use is - well,
should be - shown in the bottom like of the box immediately above H5
"Computing Table III". That line was
Tx2.value = St }
and now is
document.getElementById("Tx2").value = St }
which might help.


Prfx is from an include file; if it had not been defined, that should
have become apparent earlier, judging from IE.

If IE opens a new page or window, functions provided to the old page by
include files still work. Is that not the case in FireFox?

 

[cwdjrxyz]

Tx2 is the textarea in the Table II section. Its only use is - well,
should be - shown in the bottom like of the box immediately above H5
"Computing Table III". That line was
Tx2.value = St }
and now is
document.getElementById("Tx2").value = St }
which might help.


Prfx is from an include file; if it had not been defined, that should
have become apparent earlier, judging from IE.

If IE opens a new page or window, functions provided to the old page by
include files still work. Is that not the case in FireFox?

Unfortunately your recent quoted change has made no difference for the
latest versions of Firefox, Mozilla, or Netscape browsers. Also the old
Netscape 4.8 does the same as I reported earlier. The most recent
version of Netscape has a loading indicator that moves an arrow in a
circle as long as the page is still loading. After you click on the
button for table 3, the first few items of the heading appear. However,
after that, the loading indicator keeps spinning and nothing happens.

 

[RobG]

JRS: In article <[email protected]>, dated
Sun, 16 Oct 2005 23:56:18, seen in Randy Webb




Giving an error message is not hanging the system.


Tx2 is the textarea in the Table II section. Its only use is - well,
should be - shown in the bottom like of the box immediately above H5
"Computing Table III". That line was
Tx2.value = St }
and now is
document.getElementById("Tx2").value = St }
which might help.

No, because 'Tx2' is the textarea' name, not it's id:

Prfx is from an include file; if it had not been defined, that should
have become apparent earlier, judging from IE.

If IE opens a new page or window, functions provided to the old page by
include files still work. Is that not the case in FireFox?

Err, yes and no. Global stuff persists in Firefox for one
document.write() call, after that it's hosed. Use of a single
document.write() should fix the problem - test below.

The same effect is achieved by preceding document.write() with
document.open().



[include.js]

function getExtValue(){
return 'The external value';
}

<html><title></title>

<script type="text/javascript" src="test.js"></script>
<script type="text/javascript">

var internalValue = "Here is the in-page value";

function newPageA()  // This will fail in Firefox but not IE
{
document.write('<html><title></title>');
document.write( internalValue );
document.write('<br>');
document.write( getExtValue() + '</html>' );
document.close();
}

function newPageB()  // Works in Firefox and IE
{
//  document.open();  // This will cause Firefox to fail
document.write(
'<html><title></title>'
+ internalValue + '<br>'
+ getExtValue() + '</html>' );
document.close();
}

</script>

<input type="button" value="Write new page A" onclick="newPageA();">
<input type="button" value="Write new page B" onclick="newPageB();">

</html>

 

[Dr John Stockton]

JRS: In article <Y8%[email protected]>, dated Tue, 18
Oct 2005 04:48:56, seen in RobG

No, because 'Tx2' is the textarea' name, not it's id:

<textarea name=Tx2 rows=12 cols=67>?</textarea>

Replaced name=txt by ID=Tx2 .

Err, yes and no. Global stuff persists in Firefox for one
document.write() call, after that it's hosed. Use of a single
document.write() should fix the problem - test below.

So it only persists up to providing the argument(s) for document.write?


I've changed all such instances in that page, I think; I suppose I now
have to look across the whole site, and I don't see an automated way to
search. That's 264 + 3 buttons to look at ... - though most will not be
affected.

The same effect is achieved by preceding document.write() with
document.open().

The good effect or the bad? I imagine you mean the good.


At Computing Table III there's a button "Pop Code Up" : if
document.open is good, it should be OK, if not then ???. Relevant code
is visible in <URL:http://www.merlyn.demon.co.uk/js-nclds.htm#Inc1>, end
of section, in "PopBtn & PopThis", with 3 test cases.

Thanks.

 

[cwdjrxyz]

I've changed all such instances in that page, I think; I suppose I now
have to look across the whole site, and I don't see an automated way to
search. That's 264 + 3 buttons to look at ... - though most will not be
affected.

Your table 3 now views the same on the 3 mentioned Mozilla family
browsers and the IE6. I have a screen capture at
http://www.cwdjr.info/temp/table3mod.jpg . This was made on the
SBC/Yahoo DSL browser(IE6 based). Near the bottom of the table on, on
the mentioned browsers, there is a displacement that I thought you
should see before you spend much time converting your pages. Is this
displacement what you intended?

 

[RobG]

JRS: In article <Y8%[email protected]>, dated Tue, 18
Oct 2005 04:48:56, seen in RobG



Replaced name=txt by ID=Tx2 .





So it only persists up to providing the argument(s) for document.write?


I've changed all such instances in that page, I think; I suppose I now
have to look across the whole site, and I don't see an automated way to
search. That's 264 + 3 buttons to look at ... - though most will not be
affected.




The good effect or the bad? I imagine you mean the good.

No, the bad - sorry for the ambiguity.

document.open() seems to remove everything of the current page from
Firefox's memory, the same as multiple calls to document.write() seem to
do. Don't use document.open() unless you intend flushing the memory.

At Computing Table III there's a button "Pop Code Up" : if
document.open is good, it should be OK, if not then ???. Relevant code
is visible in <URL:http://www.merlyn.demon.co.uk/js-nclds.htm#Inc1>, end
of section, in "PopBtn & PopThis", with 3 test cases.

Yes, they seem to popup OK.

 

[Dr John Stockton]

JRS: In article <[email protected]>,
dated Wed, 19 Oct 2005 11:21:48, seen in
I've done those - where a button gives a new page, it should be done by
a single final document.write[ln]().

Your table 3 now views the same on the 3 mentioned Mozilla family
browsers and the IE6. I have a screen capture at
http://www.cwdjr.info/temp/table3mod.jpg . This was made on the
SBC/Yahoo DSL browser(IE6 based). Near the bottom of the table on, on
the mentioned browsers, there is a displacement that I thought you
should see before you spend much time converting your pages. Is this
displacement what you intended?

It is right that the row which starts 24 should have white-space after
7, but the space should be one space longer. In the next row, the 26
should be under the 25; you show three spaces too many. With those
changes, the months would align vertically. Should be

21 10 29 18 7 26 15 4 23 12 1 20 9 28 17 6 25 14 3 G Apr 15
22 11 0 19 8 27 16 5 24 13 2 21 10 29 18 7 26 15 4 A Apr 16
23 12 1 20 9 28 17 6 25 14 3 22 11 0 19 8 27 16 5 B Apr 17
24 13 2 21 10 29 18 7 B Apr 17
26 15 4 23 12 1 20 9 28 17 6 C Apr 18
25 14 3 22 11 0 19 8 27 16 5 24 13 2 21 10 29 18 7 C Apr 18

I've made a new page showing the calculated tables by copy'n'paste -
<URL:http://www.merlyn.demon.co.uk/estrtbls.htm> - and I've removed tabs
from that copy of that table.

Table IIIa has replaced Table III, and lost its terminal a.

Again thanks.

 

[cwdjrxyz]

I've made a new page showing the calculated tables by copy'n'paste -
<URL:http://www.merlyn.demon.co.uk/estrtbls.htm> - and I've removed tabs
from that copy of that table.

Table IIIa has replaced Table III, and lost its terminal a.

It may be time to open the Champagne. Your new table 3 appears correct
on all of the browsers I mentioned in my posts to this thread,
including the old Netscape 4.8. Just to give you a final look on a
Mozilla family browser, I captured the new table 3 with the new
Netscape 8 browser. See
http://www.cwdjr.info/temp/table3newNetscape8.jpg . I know the text is
a bit small, but I had to reduce the size of it considerably to capture
the whole table on a single screen shot.

 

[Dr John Stockton]

JRS: In article <[email protected]>
, dated Thu, 20 Oct 2005 17:12:29, seen in

It may be time to open the Champagne. Your new table 3 appears correct
on all of the browsers I mentioned in my posts to this thread,
including the old Netscape 4.8. Just to give you a final look on a
Mozilla family browser, I captured the new table 3 with the new
Netscape 8 browser. See
http://www.cwdjr.info/temp/table3newNetscape8.jpg . I know the text is
a bit small, but I had to reduce the size of it considerably to capture
the whole table on a single screen shot.

Alas, that's a shot of estrtbls.htm, which contains no javascript
(though it was made from javascript output).

That page was made for reference, to show what the button-generated
Tables from estrdate.htm should really look like. That you see it right
is only to be expected.

The Table III generated by pressing the button in estrdate.htm#T3 was
yesterday as I would wish in IE 5/6, so the original problem has
departed.

Does anyone see anything still wrong in <URL:http://www.merlyn.demon.co.
uk/estrdate.htm> in any browser (including non-javascript errors)?

Thanks for looking.

 

[cwdjrxyz]

JRS: In article <[email protected]>
, dated Thu, 20 Oct 2005 17:12:29, seen in

Alas, that's a shot of estrtbls.htm, which contains no javascript
(though it was made from javascript output).

That page was made for reference, to show what the button-generated
Tables from estrdate.htm should really look like. That you see it right
is only to be expected.

The Table III generated by pressing the button in estrdate.htm#T3 was
yesterday as I would wish in IE 5/6, so the original problem has
departed.

Does anyone see anything still wrong in <URL:http://www.merlyn.demon.co.
uk/estrdate.htm> in any browser (including non-javascript errors)?

Using the URL you provided in the quote just above, I generated table 3
on all of my browsers. The Mozilla family(Firefox, Mozilla, and
Netscape) and Opera, all most recent versions, give what you want.
However there still seems to be a problem near the bottom of table 3
when viewed on my SBC/Yahoo DSL browser. This is supposed to just be an
IE6 browser with additional "window dressing" and a few added functions
such as Yahoo mail. The DSL browser has taken over IE6 on my computer.
I would expect to see the same as on a "nude" IE6 browser, but you
might want to have someone else check to be certain. A screen capture
is at http://www.cwdjr.info/temp/table3SbcYahoo.jpg . Also the URL you
give will not display on the old Netscape 4.8 at all and gives the same
error message I mentioned in earlier posts.

 

[Dr John Stockton]

JRS: In article <[email protected]>
, dated Sat, 22 Oct 2005 13:57:22, seen in

Using the URL you provided in the quote just above, I generated table 3
on all of my browsers. The Mozilla family(Firefox, Mozilla, and
Netscape) and Opera, all most recent versions, give what you want.
However there still seems to be a problem near the bottom of table 3
when viewed on my SBC/Yahoo DSL browser. This is supposed to just be an
IE6 browser with additional "window dressing" and a few added functions
such as Yahoo mail. The DSL browser has taken over IE6 on my computer.
I would expect to see the same as on a "nude" IE6 browser, but you
might want to have someone else check to be certain. A screen capture
is at http://www.cwdjr.info/temp/table3SbcYahoo.jpg .

There appears to be an error, in that browser, in the size of horizontal
tab. That, if verified, could be taken up with the supplier.

I've changed estrdate.htm TableIII to use spaces instead of tabs; but
it's not practical to do so everywhere.

Also the URL you
give will not display on the old Netscape 4.8 at all and gives the same
error message I mentioned in earlier posts.

Netscape 4.8, giving "File include1.js end fault, RSVP!", is evidently
reading the start, but not the end, of <URL:http://www.merlyn.demon.co.
uk/include1.js>.

That can only be tracked down by someone who knows Netscape 4.8.

I don't much care, but if someone tells me a simple and harmless change,
I can apply it.

Thanks.

 

[cwdjrxyz]

There appears to be an error, in that browser, in the size of horizontal
tab. That, if verified, could be taken up with the supplier.

I've changed estrdate.htm TableIII to use spaces instead of tabs; but
it's not practical to do so everywhere.

The Big-8 newsgroup management is attempting to legitimise its questionable
practices while retaining its elitist hegemony. Read <URL:news:news.groups>.

Using http://www.merlyn.demon.co.uk/estrdate.htm and evaluation of
table 3 now gives the same correct results on all of my browsers except
Netscape 4.8.

However, I nearly wrote you a very different message. When I first
started checking, I was getting an RSVP error on all of the Netscape
family browsers. Then I received a connection refused popup. Then a
connection was very slow. Then, very suddenly, everything started
working correctly again. There may have been issues at your server, at
a router, or somewhere else between your server and me. One thing to
keep in mind is that I am using an extremely fast connection. I
download at 2.55 Mbps from a test site and at 2.3 Mbps from my server.
I downloaded a program from Finland at about 2 Mbps. The point is, that
if a site links to other things, they need to be available much sooner
than on a dialup connection. I have no idea if this would have any
bearing on your pages, but I thought I should mention it.