Fortran - Standards compliance (HTML)

Author

Standards compliance (HTML)

Michael Prager

2004-07-19, 4:00 pm

This post is about standards compliance in general, and HTML in
particular. If you're not interested, please go no further.
However, I thought that since our Fortran users are well aware
of the Fortran standard, there might be some interest in this.

I am doing some Web programming on my off hours (a volunteer
job). As I finish a page, I validate the code against the HTML
standard; this can be done at

http://validator.w3.org/

It took me a few tries to get everything OK.

Then I tried my bookmarks, including several Fortran vendors,
government sites, shareware sites, and so on. After trying 50
or so, not a single one validated as correct (standard) HTML.

Finally, I searched Google for "computer science department." I
didn't have all day, so I tried only a few: U Washington,
Carnegie Mellon, Stanford, University of Illinois, Perdue,
University of California at Berkeley, University of North
Carolina, and University College London. Not ONE of their home
pages passed the test.

This has given me a new perspective on standards compliance.

--
Mike Prager, NOAA, Beaufort, NC
Address spam-trapped; remove color to reply.
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.

TimC

2004-07-19, 8:57 pm

On Mon, 19 Jul 2004 at 19:36 GMT, Michael Prager (aka Bruce)
was almost, but not quite, entirely unlike tea:
> http://validator.w3.org/
>
> It took me a few tries to get everything OK.
>
> Then I tried my bookmarks, including several Fortran vendors,
> government sites, shareware sites, and so on. After trying 50
> or so, not a single one validated as correct (standard) HTML.

It's telling, isn't it?

I guess I don't feel so guilty about my Fortran code now :)

> Finally, I searched Google for "computer science department." I
> didn't have all day, so I tried only a few: U Washington,
> Carnegie Mellon, Stanford, University of Illinois, Perdue,
> University of California at Berkeley, University of North
> Carolina, and University College London. Not ONE of their home
> pages passed the test.

We unofficially renamed our old department SchIT for a reason.

--
TimC -- http://astronomy.swin.edu.au/staff/tconnors/
"You can't trust any bugger further than you can
throw him, and there's nothing you can do about it,
so let's have a drink." -- Terry Pratchett

Gary L. Scott

2004-07-19, 8:57 pm

Michael Prager wrote:
>
> This post is about standards compliance in general, and HTML in
> particular. If you're not interested, please go no further.
> However, I thought that since our Fortran users are well aware
> of the Fortran standard, there might be some interest in this.
>
> I am doing some Web programming on my off hours (a volunteer
> job). As I finish a page, I validate the code against the HTML
> standard; this can be done at
>
> http://validator.w3.org/
>
> It took me a few tries to get everything OK.
>
> Then I tried my bookmarks, including several Fortran vendors,
> government sites, shareware sites, and so on. After trying 50
> or so, not a single one validated as correct (standard) HTML.
>
> Finally, I searched Google for "computer science department." I
> didn't have all day, so I tried only a few: U Washington,
> Carnegie Mellon, Stanford, University of Illinois, Perdue,
> University of California at Berkeley, University of North
> Carolina, and University College London. Not ONE of their home
> pages passed the test.
>
> This has given me a new perspective on standards compliance.

What was the predominant noncompliances? My guess is that since 95% of
all net traffic is from Windows computers, that it involves MS
extensions of some sort.

>
> --
> Mike Prager, NOAA, Beaufort, NC
> Address spam-trapped; remove color to reply.
> * Opinions expressed are personal and not represented otherwise.
> * Any use of tradenames does not constitute a NOAA endorsement.

--

Gary Scott
mailto:garyscott@ev1.net

Fortran Library: http://www.fortranlib.com

Support the Original G95 Project: http://www.g95.org
-OR-
Support the GNU GFortran Project: http://gcc.gnu.org/fortran/index.html

Why are there two? God only knows.

Democracy is two wolves and a sheep, voting on what to eat for dinner...
Liberty is a well armed sheep contesting the vote. - Thomas Jefferson

James Van Buskirk

2004-07-20, 3:59 pm

"Gary L. Scott" <garyscott@ev1.net> wrote in message
news:40FC658B.721C8452@ev1.net...

> What was the predominant noncompliances? My guess is that since 95% of
> all net traffic is from Windows computers, that it involves MS
> extensions of some sort.

I tried the given validator on my web page and it rated my
web page as invalid for not referencing the validator's web
page!

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

Richard Edgar

2004-07-20, 3:59 pm

James Van Buskirk wrote:

> I tried the given validator on my web page and it rated my
> web page as invalid for not referencing the validator's web
> page!

Where's the page? I've never seen that error before.

Richard

David Ham

2004-07-20, 3:59 pm

On Tue, 20 Jul 2004 13:47:16 GMT
"James Van Buskirk" <not_valid@comcast.net> wrote:

> "Gary L. Scott" <garyscott@ev1.net> wrote in message
> news:40FC658B.721C8452@ev1.net...
>
>
> I tried the given validator on my web page and it rated my
> web page as invalid for not referencing the validator's web
> page!

If you are referring to http://home.comcast.net/~kmbtib/ then I don't
get that message but I do get a bunch of others. That is one of the more
spectacularly invalid pages I've seen. Not only is there no Document
Type Declaration, there is no <html> and no <body> for that matter
either.

David

>
> --
> write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
> 6.0134700243160014d-154/),(/'x'/)); end
>
>

Michael Prager

2004-07-20, 3:59 pm

"Gary L. Scott" <garyscott@ev1.net> wrote:

>What was the predominant noncompliances? My guess is that since 95% of
>all net traffic is from Windows computers, that it involves MS
>extensions of some sort.

My impression is that there is a basic ignorance of HTML. This
probably stems from two things: most browsers are forgiving, and
some Web authoring programs may not emit correct HTML. So
"whatever works" is considered good. Sound familiar?

As to errors, many pages do not declare the HTML level or
character encoding, or even have an HTML declaration, all of
which are required. Many include graphics withoug an alt
attribute (text description), which is required. Many values
that should have been in quotes were not.

For more information, try it yourself on a few pages!

--
Mike Prager, NOAA, Beaufort, NC
Address spam-trapped; remove color to reply.
* Opinions expressed are personal and not represented otherwise.
* Any use of tradenames does not constitute a NOAA endorsement.

TimC

2004-07-20, 3:59 pm

On Tue, 20 Jul 2004 at 14:00 GMT, Richard Edgar (aka Bruce)
was almost, but not quite, entirely unlike tea:
> James Van Buskirk wrote:
>
>
> Where's the page? I've never seen that error before.

I'd say it's simple misinterpretation of the error message.

--
TimC -- http://astronomy.swin.edu.au/staff/tconnors/
A mouse is a device used to focus xterms.

James Van Buskirk

2004-07-20, 3:59 pm

"Richard Edgar" <rge21@astro.su.se> wrote in message
news:cdj8hf$401$1@news.su.se...

> James Van Buskirk wrote:

[color=darkred]
> Where's the page? I've never seen that error before.

Just type up any old thing by hand in notepad and post it
as your index.html file. Here's an example of a page I
created just because some fool wanted to see a program of
mine so I copied a whole directory to my web site and
created an index.html file that gave links to all the files.
Took me 11 lines of Fortran to write that index.html file:

program zzz
implicit none
character(80) filename

open(10,file='zzz.txt',status='old')
open(11,file='index.html',status='replace')
do
read(10,'(a)',end=10) filename
write(11,'(5a)') '<a
href="',trim(filename),'">',trim(filename),'</a> '
end do
10 continue
end program zzz

Of course I had to execute the command:

dir /b > zzz.txt

at the Windoze command prompt before running the above.
This technique gives you quite a serviceable index.html file:

http://home.comcast.net/~kmbtib/Sieve30h/index.html

but the validator doesn't like it, one of the errors given
is "No DOCTYPE Found" and when you follow the links you
get to example DOCTYPE tags such as:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
"http://www.w3.org/TR/html4/strict.dtd">

Seems like a racket to me: if you want to have a valid web page,
you have to reference the validator's page right off the bat.

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

James Van Buskirk

2004-07-20, 3:59 pm

"David Ham" <d.a.ham@citg.tudelft.nl> wrote in message
news:20040720161809.3a661629.d.a.ham@citg.tudelft.nl...

> If you are referring to http://home.comcast.net/~kmbtib/ then I don't
> get that message but I do get a bunch of others. That is one of the more
> spectacularly invalid pages I've seen. Not only is there no Document
> Type Declaration, there is no <html> and no <body> for that matter
> either.

Oooh! Oooh! Can I quote you?

"one of the more spectacular... pages I've seen."

-David Ham

Would look real good right on the top of the page in .

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

Richard Edgar

2004-07-20, 3:59 pm

James Van Buskirk wrote:

> but the validator doesn't like it, one of the errors given
> is "No DOCTYPE Found" and when you follow the links you
> get to example DOCTYPE tags such as:
>
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
> "http://www.w3.org/TR/html4/strict.dtd">
>
> Seems like a racket to me: if you want to have a valid web page,
> you have to reference the validator's page right off the bat.

A racket? How are they making money off it? In any case, I don't believe
that the URL is mandatory - it simply provides an opportunity for the
user agent to get the SGML description of HTML (in this case). I doubt
that any current user agents do this, since they have them built in.
IIRC, some _do_ use the presence/absence of the URL to alter how they
render the page. If the URL is absent, they assume that the page author
doesn't know what they're doing, and relax the rules accordingly.

The DOCTYPE declares what the file contains, and as such is rather
useful. Compare and contrast with '.f', '.f90', '.f95' etc.

Richard

David Ham

2004-07-20, 3:59 pm

On Tue, 20 Jul 2004 14:34:25 GMT
"James Van Buskirk" <not_valid@comcast.net> wrote:

> "Richard Edgar" <rge21@astro.su.se> wrote in message
> news:cdj8hf$401$1@news.su.se...
>
>
>
>
> Just type up any old thing by hand in notepad and post it
> as your index.html file. Here's an example of a page I
> created just because some fool wanted to see a program of
> mine so I copied a whole directory to my web site and
> created an index.html file that gave links to all the files.
> Took me 11 lines of Fortran to write that index.html file:
>
> program zzz
> implicit none
> character(80) filename
>
> open(10,file='zzz.txt',status='old')
> open(11,file='index.html',status='replace')
> do
> read(10,'(a)',end=10) filename
> write(11,'(5a)') '<a
> href="',trim(filename),'">',trim(filename),'</a> '
> end do
> 10 continue
> end program zzz
>
> Of course I had to execute the command:
>
> dir /b > zzz.txt
>
> at the Windoze command prompt before running the above.
> This technique gives you quite a serviceable index.html file:
>
> http://home.comcast.net/~kmbtib/Sieve30h/index.html
>
> but the validator doesn't like it, one of the errors given
> is "No DOCTYPE Found" and when you follow the links you
> get to example DOCTYPE tags such as:
>
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
> "http://www.w3.org/TR/html4/strict.dtd">
>
> Seems like a racket to me: if you want to have a valid web page,
> you have to reference the validator's page right off the bat.

OK, with respect to sgml and xml, "valid" is a technical term. It means
that a document is both well-formed (which has to do with basic issues
like tags being properly nested) and it has to conform to a Document
Type Definition. In order to be valid html, you have to conform to one
of the versions of the html DTD, that's what "valid html" means. The
Document Type Definition defines the given language (in this case,
html). The DTD defines which elements and attributes are allowed and
specifies which elements may contain which other elements (eg a <li> is
only valid in a list).

Now, the body responsible for the html recommendation (ie standard) is
the World Wide Web consortium (W3C). The canonical reference for the
html DTD is them. A DOCTYPE or Document Type Declaration is the way that
you specify which sgml or xml application (ie language) your document is
written in. It is therefore not a surprise that an html document's DTD
points to W3C.

W3C also provide the validator that the OP used. It's not the only one
on the web but, since it's provided by the standards body itself, it's
rather widely trusted. See, it's not a racket.

Incidentally, to make it produce valid html you could modify your
program as follows:

program zzz
implicit none
character(80) filename

open(10,file='zzz.txt',status='old')
open(11,file='index.html',status='replace')
write(11, *) '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"'
write(11, *) '"http://www.w3.org/TR/html4/strict.dtd">'

write(11, *) '<html>'
write(11, *) '<head>'
write(11, *) '<title>Directory listing</title>'
write(11, *) '</head>'
write(11, *) '<body>'
write(11, *) '<ul>'
do
read(10,'(a)',end=10) filename
write(11,'(5a)') '<li><a href="',trim(filename),'">',trim(filename),'</a> '
end do
10 continue
write(11, *) '</ul>'
write(11, *) '</body>'
write(11, *) '</html>'
end program zzz

>
> --
> write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
> 6.0134700243160014d-154/),(/'x'/)); end
>
>

TimC

2004-07-20, 3:59 pm

On Tue, 20 Jul 2004 at 14:34 GMT, James Van Buskirk (aka Bruce)
was almost, but not quite, entirely unlike tea:
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"
> "http://www.w3.org/TR/html4/strict.dtd">
>
> Seems like a racket to me: if you want to have a valid web page,
> you have to reference the validator's page right off the bat.

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">

WFM.

--
TimC -- http://astronomy.swin.edu.au/staff/tconnors/
"Mango sorbet is clearcut proof that we have progressed beyond the
bare needs of survival and have progressed to the transcendant."
- Marc Goodman in talk.bizarre

David Ham

2004-07-20, 3:59 pm

On Tue, 20 Jul 2004 15:29:25 GMT
TimC <tconnors@no.astro.spam.swin.accepted.edu.here.au> wrote:

> On Tue, 20 Jul 2004 at 14:34 GMT, James Van Buskirk (aka Bruce)
> was almost, but not quite, entirely unlike tea:
>
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
>
> WFM.

In a feat of stunning hypocracy, I note that your work page isn't valid
either :-). (My excuse is that I'm not allowed to fix the mistakes on my
work page. My personal page is valid).

David

>
> --
> TimC -- http://astronomy.swin.edu.au/staff/tconnors/
> "Mango sorbet is clearcut proof that we have progressed beyond the
> bare needs of survival and have progressed to the transcendant."
> - Marc Goodman in talk.bizarre

James Van Buskirk

2004-07-20, 3:59 pm

"David Ham" <d.a.ham@citg.tudelft.nl> wrote in message
news:20040720171452.5c727ee5.d.a.ham@citg.tudelft.nl...

> Incidentally, to make it produce valid html you could modify your
> program as follows:

Actually, your program didn't produce valid html so I
modified it as follows:

! File: zzz.f90
! Public domain 2004 James Van Buskirk, David Ham, Tim Connors
! Before running this program, create a file called zzz.txt
! which has all the desired filenames, one per line, as by
! typing:
! dir /b > zzz.txt
! at the windows command prompt.
program zzz
implicit none
character(80) filename

open(10,file='zzz.txt',status='old')
open(11,file='index.html',status='replace')
write(11,'(a)') '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01
Transitional//EN">'
write(11,'(a)') '<html>'
write(11,'(a)') '<head>'
write(11,'(a)') '<META HTTP-EQUIV="Content-Type" CONTENT="text/html;
charset=utf-8">'
write(11,'(a)')
'<title>документ</title>'
write(11,'(a)') '</head>'
write(11,'(a)') '<body>'
write(11,'(a)') ''
do
read(10,'(a)',end=10) filename
write(11,'(5a)') '<a
href="',trim(filename),'">',trim(filename),'</a> '
end do
10 continue
write(11,'(a)') ''
write(11,'(a)') '</body>'
write(11,'(a)') '</html>'
end program zzz
! End of file: zzz.f90

The result is at:

http://home.comcast.net/~kmbtib/Sieve30h/index.html

BTW, what browsers fail to load my pages properly even if
they were nonstandard?

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

TimC

2004-07-21, 3:57 am

On Tue, 20 Jul 2004 at 15:45 GMT, David Ham (aka Bruce)
was almost, but not quite, entirely unlike tea:
> On Tue, 20 Jul 2004 15:29:25 GMT
> TimC <tconnors@no.astro.spam.swin.accepted.edu.here.au> wrote:
>
> In a feat of stunning hypocracy, I note that your work page isn't valid
> either :-). (My excuse is that I'm not allowed to fix the mistakes on my
> work page. My personal page is valid).

I noticed someone checked my page :)

I'll fix my html after I fix my fortran (I do put the w3c logo on
pages that I do get around to checking).

--
TimC -- http://astronomy.swin.edu.au/staff/tconnors/
If I'd known computer science was going to be like this, I'd never have
given up being a rock 'n' roll star. -- G. Hirst

David Ham

2004-07-21, 8:59 am

On Tue, 20 Jul 2004 19:40:05 GMT
"James Van Buskirk" <not_valid@comcast.net> wrote:

> "David Ham" <d.a.ham@citg.tudelft.nl> wrote in message
> news:20040720171452.5c727ee5.d.a.ham@citg.tudelft.nl...
>
>
> Actually, your program didn't produce valid html so I
> modified it as follows:

Umm, I think it did. I checked. The message which the system produced
was
that the page was tentatively valid subject to a correct character
encoding. Character encoding is usually specified by the web server in
the http headers so I didn't fix it. Providing the information in a meta
element is, of course, also OK.

>
> ! File: zzz.f90
> ! Public domain 2004 James Van Buskirk, David Ham, Tim Connors
> ! Before running this program, create a file called zzz.txt
> ! which has all the desired filenames, one per line, as by
> ! typing:
> ! dir /b > zzz.txt
> ! at the windows command prompt.
> program zzz
> implicit none
> character(80) filename
>
> open(10,file='zzz.txt',status='old')
> open(11,file='index.html',status='replace')
> write(11,'(a)') '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01
> Transitional//EN">'
^
The transitional DTD is deprecated (by W3C) for new work. It's there for
legacy purposes and contains many elements which deal with
presentational issues in ways which are no longer considered a good
idea.

> write(11,'(a)') '<html>'
> write(11,'(a)') '<head>'
> write(11,'(a)') '<META HTTP-EQUIV="Content-Type"
> CONTENT="text/html;
> charset=utf-8">'
> write(11,'(a)')
> '<title>документ</title>'

^
That bit is quite

, even if I can't read it.

> write(11,'(a)') '</head>'
> write(11,'(a)') '<body>'
> write(11,'(a)') ''
^
Font is one of those deprecated tags. We now have stylesheets for this.
For a little page like this one, the simplest thing to do is chuck a
style attribute in the body tag. eg :

<body style="font-family:'Courier New',sans-serif">

One of the advantages of this approach is that it does graceful
fallover. If a browser doesn't have Courier New, it will use another
sans-serif font rather than falling back to the default (which is often
some form of Roman). As an example, my computer (running Linux) doesn't
have a "Courier New" font although it does have plenty of sans-serif
fonts. Using the element, it produces output in roman but using
the style element it falls back to another sans-serif font producing a
result which is something like what you asked for. Oh, incidentially, to
do this you may (depending on the web server settings) need to put a:

<META http-equiv="Content-Style-Type" content="text/css">

in the <head> of the document.

> do
> read(10,'(a)',end=10) filename
> write(11,'(5a)') '<a
> href="',trim(filename),'">',trim(filename),'</a> '

Making a list by using removes the information that this is a list.
This is not so important in this case but on a more complex web page,
unusual browsers (like screen readers and browsers in very small
windows) can use that sort of information to work out how to render a
page. Indeed, plain text directly inside the body is invalid in the
strict DTD because text should always occur inside some form of
structural markup. To generate a list which is not indented and has no
bullets, you can use:

<ul style="list-style-type:none;padding-left:0pt">

> end do
> 10 continue
> write(11,'(a)') ''
> write(11,'(a)') '</body>'
> write(11,'(a)') '</html>'
> end program zzz
> ! End of file: zzz.f90
>
> The result is at:
>
> http://home.comcast.net/~kmbtib/Sieve30h/index.html
>
> BTW, what browsers fail to load my pages properly even if
> they were nonstandard?

For pages this simple, probably none. Browsers tend to be pretty
tolerant - they need to be because there's so much broken html out
there. The more you put on a page, the more significant it becomes. On
the other hand, it's like using nonstandard Fortran - there are lots of
extensions which almost always work but the processor is more than
entitled to start WWIII on encountering one if it feels like it.

So, a version of the program which produces what I think is the output
you want but conforming to the strict DTD is:

! File: zzz.f90
! Public domain 2004 James Van Buskirk, David Ham, Tim Connors
! Before running this program, create a file called zzz.txt
! which has all the desired filenames, one per line, as by
! typing:
! dir /b > zzz.txt
! at the windows command prompt.
program zzz
implicit none
character(80) filename

open(10,file='zzz.txt',status='old')
open(11,file='index.html',status='replace')
write(11, '(a)') '<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN"'
write(11, '(a)') '"http://www.w3.org/TR/html4/strict.dtd">'

write(11, '(a)') '<html>'
write(11, '(a)') '<head>'
write(11, '(a)') '<META HTTP-EQUIV="Content-Type" &
&CONTENT="text/html;charset=utf-8">'
write(11, '(a)') '<META http-equiv="Content-Style-Type" &
&content="text/css">'
write(11, '(a)') '<title>документ</title>'
write(11, '(a)') '</head>'
write(11, '(a)') '<body style="font-family:''Courier New'',sans-serif">'
write(11, '(a)') '<ul style="list-style-type:none;padding-left:0pt">'
do
read(10,'(a)',end=10) filename
write(11,'(5a)') '<li><a href="',trim(filename),'">',trim(filename),'&
&</a></li>'
end do
10 continue
write(11, '(a)') '</ul>'
write(11, '(a)') '</body>'
write(11, '(a)') '</html>'
end program zzz

Regards,

David

James Van Buskirk

2004-07-21, 3:58 pm

"David Ham" <d.a.ham@citg.tudelft.nl> wrote in message
news:20040721112535.59e902b3.d.a.ham@citg.tudelft.nl...

> On Tue, 20 Jul 2004 19:40:05 GMT
> "James Van Buskirk" <not_valid@comcast.net> wrote:

> Umm, I think it did. I checked. The message which the system produced
> was
> that the page was tentatively valid subject to a correct character
> encoding. Character encoding is usually specified by the web server in
> the http headers so I didn't fix it. Providing the information in a meta
> element is, of course, also OK.

The validator said that I was supposed to put a file whose name
started with a period in my directory to fix the error but my
ISP doesn't permit that.

> ^
> The transitional DTD is deprecated (by W3C) for new work.

Jeez, this is like java... everything is deprecated.

>
> ^
> That bit is quite

, even if I can't read it.

I was kind of curious about the above. MSIE displays this title
just fine, but am I really supposed to terminate each number with
a semicolon? As in:

'<title>документ</title>'

> ^
> Font is one of those deprecated tags. We now have stylesheets for this.
> For a little page like this one, the simplest thing to do is chuck a
> style attribute in the body tag. eg :
>
> <body style="font-family:'Courier New',sans-serif">
>
> One of the advantages of this approach is that it does graceful
> fallover.

The 'Doze default for everything is Arial, which is proportional,
sans-serif. If possible, I would prefer to fall back on
monospaced, serif font, like Courier (which I think comes with
a Mac.) The serifs need to be there to distinguish between 1Il
and I consider monospaced fonts to be much more readable, perhaps
due to habituation to punch card and line printer output.

Well, thanks for your comments. We managed to get some Fortran
code posted to this thread in any case.

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

Jan Vorbrüggen

2004-07-21, 3:58 pm

> The validator said that I was supposed to put a file whose name
> started with a period in my directory to fix the error but my
> ISP doesn't permit that.

Ugh...make a Unixism such as dot-files part of a standard? No thanks!

> I was kind of curious about the above. MSIE displays this title
> just fine, but am I really supposed to terminate each number with
> a semicolon? As in:
>
> '<title>документ</title>'

Yes please. Anything else is broken from the point of view of the parser.
All browsers I use do _not_ support this MSIE stupidity.

Jan

Richard Edgar

2004-07-21, 3:58 pm

James Van Buskirk wrote:

>
> Jeez, this is like java... everything is deprecated.

That hardly stops you using the transitional DTD.

> but am I really supposed to terminate each number with
> a semicolon? As in:
>
> '<title>документ</title>'

Yes.

>
> The 'Doze default for everything is Arial, which is proportional,
> sans-serif. If possible, I would prefer to fall back on
> monospaced, serif font, like Courier (which I think comes with
> a Mac.) The serifs need to be there to distinguish between 1Il
> and I consider monospaced fonts to be much more readable, perhaps
> due to habituation to punch card and line printer output.

If you think that it should be a monospaced font, why don't you have
<body style="font-family: monospace;">

Richard

Richard Edgar

2004-07-21, 3:58 pm

Jan Vorbrüggen wrote:
>
> Ugh...make a Unixism such as dot-files part of a standard? No thanks!

Actually, I can't find this suggestion from the Validator links.
<http://validator.w3.org/docs/help.html#faq-charset>
Simply says that this information is usually provided by the web server.

Richard

David Ham

2004-07-21, 3:58 pm

On Wed, 21 Jul 2004 15:18:30 GMT
"James Van Buskirk" <not_valid@comcast.net> wrote:

> "David Ham" <d.a.ham@citg.tudelft.nl> wrote in message
> news:20040721112535.59e902b3.d.a.ham@citg.tudelft.nl...
>
>
>
> The validator said that I was supposed to put a file whose name
> started with a period in my directory to fix the error but my
> ISP doesn't permit that.
>

I suspect that you managed to click through to something which told you
how to fix it in apache using the .htaccess file. apache is fairly
unixish even when running on Windows. If your ISP uses something else
(eg IIS) then that won't apply. Anyway, using the meta element is
always fine.

>
> Jeez, this is like java... everything is deprecated.
>

Yes. I think early versions of html were done in a fairly haphazard
way and a lot of structural markup and presentational stuff got mixed
up. html 4.0 was an attempt to clean up html and to line up with current
thinking in markup. As a result, there's an awful lot of older stuff
which is deprecated.

>
> I was kind of curious about the above. MSIE displays this title
> just fine, but am I really supposed to terminate each number with
> a semicolon? As in:
>
> '<title>документ</tit
> le>'
>
>
> The 'Doze default for everything is Arial, which is proportional,
> sans-serif. If possible, I would prefer to fall back on
> monospaced, serif font, like Courier (which I think comes with
> a Mac.) The serifs need to be there to distinguish between 1Il
> and I consider monospaced fonts to be much more readable, perhaps
> due to habituation to punch card and line printer output.

Ah, my bad. I thought courier was sans-serif. Anyway, the comment about
fall back still applies.

>
> Well, thanks for your comments. We managed to get some Fortran
> code posted to this thread in any case.
>

:-).

David

> --
> write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
> 6.0134700243160014d-154/),(/'x'/)); end
>
>

James Van Buskirk

2004-07-21, 3:58 pm

"Richard Edgar" <rge21@astro.su.se> wrote in message
news:cdm3id$1vg$1@news.su.se...

> James Van Buskirk wrote:

'<title>документ</title>'
[color=darkred]
> Yes.

OK, that's fixed.

> If you think that it should be a monospaced font, why don't you have
> <body style="font-family: monospace;">

But how do you specify both monospaced and serif? It seems the
the browser is supposed to scan the font families until it finds
something it likes, so if it sees monospace it may present a sans-
serif font while if it sees serif first it may serve up a proportional
font. Is HTML really this badly broken?

--
write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
6.0134700243160014d-154/),(/'x'/)); end

David Ham

2004-07-21, 3:58 pm

On Wed, 21 Jul 2004 16:57:16 GMT
"James Van Buskirk" <not_valid@comcast.net> wrote:

> "Richard Edgar" <rge21@astro.su.se> wrote in message
> news:cdm3id$1vg$1@news.su.se...
>
>
> '<title>документ</tit
> le>'
>
>
> OK, that's fixed.
>
>
> But how do you specify both monospaced and serif? It seems the
> the browser is supposed to scan the font families until it finds
> something it likes, so if it sees monospace it may present a sans-
> serif font while if it sees serif first it may serve up a proportional
> font. Is HTML really this badly broken?

Well, strictly this is CSS not HTML but the point stands. The important
thing to realise is that generic families are fall-back options. You can
specify as many alternative font names as you like before getting to the
generic one. The idea is that the generic families provide 5 font names
which every browser will recognise and map to something reasonable. So
you can do:

<body style="font-family: 'Courier New',Courier,monospace;">

and monospace will only get used if Courier New and Courier are
both absent on the viewing system.

David

>
> --
> write(*,*) transfer((/17.392111325966148d0,6.5794487871554595D-85, &
> 6.0134700243160014d-154/),(/'x'/)); end
>
>

Gary L. Scott

2004-07-21, 8:57 pm

David Ham wrote:
>
> On Wed, 21 Jul 2004 15:18:30 GMT
> "James Van Buskirk" <not_valid@comcast.net> wrote:
>
>
> I suspect that you managed to click through to something which told you
> how to fix it in apache using the .htaccess file. apache is fairly
> unixish even when running on Windows. If your ISP uses something else
> (eg IIS) then that won't apply. Anyway, using the meta element is
> always fine.
>
>
> Yes. I think early versions of html were done in a fairly haphazard
> way and a lot of structural markup and presentational stuff got mixed
> up. html 4.0 was an attempt to clean up html and to line up with current
> thinking in markup.

HTML didn't even line up with 1970's era thinking about markup. GML
although proprietary, was much better thought out. Unfortunately for
it, it was too flexible.
[color=darkred]
> As a result, there's an awful lot of older stuff
> which is deprecated.
>
>
> Ah, my bad. I thought courier was sans-serif. Anyway, the comment about
> fall back still applies.
>
>
> :-).
>
> David
>

--

Gary Scott
mailto:garyscott@ev1.net

Fortran Library: http://www.fortranlib.com

Support the Original G95 Project: http://www.g95.org
-OR-
Support the GNU GFortran Project: http://gcc.gnu.org/fortran/index.html

Why are there two? God only knows.

Democracy is two wolves and a sheep, voting on what to eat for dinner...
Liberty is a well armed sheep contesting the vote. - Thomas Jefferson

Richard Edgar

2004-07-22, 8:56 am

James Van Buskirk wrote:

>
> But how do you specify both monospaced and serif? It seems the
> the browser is supposed to scan the font families until it finds
> something it likes, so if it sees monospace it may present a sans-
> serif font while if it sees serif first it may serve up a proportional
> font. Is HTML really this badly broken?

The end user is supposed to assign whatever font mappings they choose.
If I want, I can set my monospaced font to be Zapf Dingbats, and there's
nothing the webpage can do about it. It's _my_ computer, and any web
page author who thinks otherwise requires an introduction to a
<plank width="2" height="4" />.

I don't think that (X)HTML or CSS are broken. They are, however,
frequently misunderstood. (X)HTML is meant to describe the logical
structure of a page. CSS is a method of _suggesting_ a layout.

Richard

Sponsored Links