Conditional processing for complete browser support?
Jim Cochrane
years, I'm only now studying it in depth, along with CSS, and trying to
learn how to design web pages properly. I didn't find my question
covered in the web authoring FAQ (http://www.htmlhelp.org/faq/html/) nor
with a quick google search of this newsgroup. (It may have been covered
here, but if so it's hard to find.)
My question is: Is it possible and wise to use conditional processing to
implement a web page using CSS such that the page displays reasonably
well in all or almost all currently used browsers (without making the
page overly simple or basic)?
By conditional processing, I mean sending different html data to the
user's browser depending on what browser he is using. From my experience
so far it seems to me that sometimes this is the only way to produce a
good-looking result in multiple browsers - that the ideal of using one
set of html and css instructions for all browsers is not always possible.
If the answer is not no (e.g., yes, or sometimes) then, of course, the
next question is how to do this - e.g., use Javascript? But I'd like to
start with just the first question, since I suspect this is not a simple
issue, that it's probably controversial, and that reasonable people will
have different views about it. Hopefully the resulting discussion will
be useful to others, too.
And if this issue has already been covered in detail, I'd appreciate it if
someone could point me to the discussion with a URL or newsgroup thread or
whatever.
Thanks very much!
--
Jim Cochrane; j...@dimensional.com
[When responding by email, include the term non-spam in the subject line to
get through my spam filter.]
Spartanicus
>From my experience
>so far it seems to me that sometimes this is the only way to produce a
>good-looking result in multiple browsers - that the ideal of using one
>set of html and css instructions for all browsers is not always possible.
Serving different html should never be needed, serving different css is
also not needed, but due to the many shortcomings in IE you need to
restrict yourself with regards to using certain css features.
CSS bugs and issues should be dealt with by exploiting UAs css bugs to
serve different css. Usually it suffices to distinguish between
Netscape4, IE5.x, IE6, and standard compliant renderers.
The common approach to deal with NS4 is to hide all css from it by
exploiting the fact that it doesn't support @import, there's Tantek's
voice-family hack to distinguish between IE5.x and 6, and there is the
child selector trick to distinguish IE from standard compliant
renderers.
There are quite a few more bugs to exploit, but those listed above
normally suffice. A list of exploitable bugs:
http://centricle.com/ref/css/filters/?whitebg
All of this can be incorporated into the same stylesheets, no need to
serve a different one (which btw would cause far more trouble than it
solves).
>If the answer is not no (e.g., yes, or sometimes) then, of course, the
>next question is how to do this - e.g., use Javascript?
Not for this.
--
Spartanicus
Martin Leese
...
> My question is: Is it possible and wise to use conditional processing to
> implement a web page using CSS such that the page displays reasonably
> well in all or almost all currently used browsers (without making the
> page overly simple or basic)?
Visit http://www.richinstyle/
They have bug lists for all major browsers. Some bugs
are minor, others totally screw up the page. The site
also describes tricks to hide style sheets from
particular browsers.
I found the site invaluable. I chose to use two style
sheets, a main one that will work in all browsers and a
small one containing stuff which breaks in Netscape 4.
--
Regards,
Martin Leese
E-mail: ple...@see.Web.for.e-mail.INVALID
Web: http://members.tripod.com/martin_leese/
Stephen Poley
<ple...@see.Web.for.e-mail.INVALID> wrote:
>Jim Cochrane wrote:
>...
>> My question is: Is it possible and wise to use conditional processing to
>> implement a web page using CSS such that the page displays reasonably
>> well in all or almost all currently used browsers (without making the
>> page overly simple or basic)?
>
>Visit http://www.richinstyle/
Actually http://www.richinstyle.com/
--
Stephen Poley
Karl Smith
> My question is: Is it possible and wise to use conditional processing to
> implement a web page using CSS such that the page displays reasonably
> well in all or almost all currently used browsers (without making the
> page overly simple or basic)?
Possible, yes. Wise, yes. Commercially realistic, no. If the question
regards a design-and-forget template for no-clue users to periodically
update "content" through a click-and-hope "CMS" or anything else
involving the word "interface", forget it.
> By conditional processing, I mean sending different html data to the
> user's browser depending on what browser he is using.
HTML is never data. Data is the bit *between* the tags. If I've read
you correctly, you know that.
> ...good-looking result in multiple browsers - that the ideal of using one
> set of html and css instructions for all browsers is not always possible.
>
> If the answer is not no (e.g., yes, or sometimes) then, of course, the
> next question is how to do this - e.g., use Javascript?
JavaScript. Funny. MS conditional comments are almost completely
reliable (excepting their use for DOCTYPE declarations, thanks to a
Mozilla bug).
> ...since I suspect this is not a simple issue,
I think it *is* simple. MS has a taken a different evolutionary path
to other browser vendors, concentrating on the ability to render text
in as many languages as possible. Other vendors have concentrated on
taking "western" scripts and prettifying as much as possible with
CSS2. Ignoring the features of other browsers so the result will look
the same as MSIE is stupid.
> that it's probably controversial,
A simple statement like, "The sky is blue," is controversial to the
CIWAHians.
"Only for twelve hours a day."
"Well no, actually. For some value of 'hours a day' that depends
critically on the user's latitude."
"When you write 'latitude', ITYM 'latitude and time of year,' don't
you?"
"Oh yes. It's off-topic here, anyway. But to try to get back to
something on-topic, the concept of 'colo(u)r of the sky' is inherently
inaccessible and should be avoided when authoring in a WWW context.
Consider, what does 'The sky is blue' mean to a blind user using a
speech browser?"
> and that reasonable people will have different views about it.
Reasonable people don't frequent Usenet.
> Hopefully the resulting discussion will be useful to others, too.
Unlikely. It's off-topic here anyway and belongs in .site-design or
.stylesheets
--
Karl Smith.
Chris Morris
> A simple statement like, "The sky is blue," is controversial to the
> CIWAHians.
> "Only for twelve hours a day."
> "Well no, actually. For some value of 'hours a day' that depends
> critically on the user's latitude."
> "When you write 'latitude', ITYM 'latitude and time of year,' don't
> you?"
> "Oh yes. It's off-topic here, anyway. But to try to get back to
> something on-topic, the concept of 'colo(u)r of the sky' is inherently
> inaccessible and should be avoided when authoring in a WWW context.
> Consider, what does 'The sky is blue' mean to a blind user using a
> speech browser?"
colour-blindness, users who are underground, etc. since accessibility
isn't just about blind users."
"Far better to use structural markup to convey the same meaning, so
<sky time='day' weather='clear'> instead of <sky bgcolor='blue'> or
even worse <sky><font color='blue'> or merely 'blue sky' - how's a
search engine supposed to know if that's about the weather or research
if you don't include proper markup?"
"Though Netscape 4 gets that wrong and displays the sky as clear
rather than as blue."
--
Chris
Stephen Poley
wrote:
You omitted to mention that when using the Northern Britain colour
palette the sky invariably renders as grey anyway.
Andy Dingley
<j...@shell.dimensional.com> wrote:
>By conditional processing, I mean sending different html data to the
>user's browser depending on what browser he is using.
_IMHO_ (and I've been flamed for this many times) it's just about
excusable to do this if NS4 legacy support is vital to you. For
anything newer, you don't need to go to this amount of trouble.
It's also completely impractical to do it at all well. The main
problem is next week's new browser that you didn't know about when you
coded the test.
--
Smert' spamionam
Jim Cochrane
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>
>>From my experience
>>so far it seems to me that sometimes this is the only way to produce a
>>good-looking result in multiple browsers - that the ideal of using one
>>set of html and css instructions for all browsers is not always possible.
Thanks, Spartanicus, for your informative response.
> Serving different html should never be needed, serving different css is
> also not needed, but due to the many shortcomings in IE you need to
> restrict yourself with regards to using certain css features.
>
> CSS bugs and issues should be dealt with by exploiting UAs css bugs to
> serve different css. Usually it suffices to distinguish between
> Netscape4, IE5.x, IE6, and standard compliant renderers.
>
> The common approach to deal with NS4 is to hide all css from it by
> exploiting the fact that it doesn't support @import, there's Tantek's
However, in my admittedly limited experience so far, it seems it's not
always easy to produce an html document that looks acceptable in NS4
without style sheets, while still using css for browsers that support it -
that is, often when I put something together that looks good with css, it
looks lousy in NS4. For example, one page I'm working on displays, from
top to bottom, the navigation section, and then the main content (rather
than the nav. section appearing on the left and the content on the right);
and the main content, which does not use tables, ends up with what should
be two columns, left and right, appearing in the same place, among other
problems. In other words, it looks great in Mozilla, but in NS4 it looks
quite atrocious.
Are problems like this not hard to avoid after "getting the hang of it", or
is this a challeng even for experienced web developers who know css well?
(I'm anticipating that you, or someone else reading this, will invite me to
post a URL for the page; I might be able to do this, but I'll first need to
get permission from the person for whom I'm developing the web site.)
> voice-family hack to distinguish between IE5.x and 6, and there is the
> child selector trick to distinguish IE from standard compliant
> renderers.
>
> There are quite a few more bugs to exploit, but those listed above
> normally suffice. A list of exploitable bugs:
> http://centricle.com/ref/css/filters/?whitebg
>
> All of this can be incorporated into the same stylesheets, no need to
> serve a different one (which btw would cause far more trouble than it
> solves).
Do you mean that it's possible, and not too difficult or ugly, to code one
css file to take advantage of quirks in the different browers (such as:
"This browser will just ignore this section and use that one.", "This
brower will interpret this section as a comment.", etc.) such that some
browsers will use a particular set of "selector definitions" (or whatever
the correct terminology is) and other browsers will use a different set?
>>If the answer is not no (e.g., yes, or sometimes) then, of course, the
>>next question is how to do this - e.g., use Javascript?
>
> Not for this.
So, I take it you would regard this example as bad practice:
http://www.directorystore.com/index.jsp
? (If I'm reading the source correctly, it runs this javascript script:
var style = "";
if (navigator.platform != "Win32")
style = '<link rel="stylesheet" type="text/css" href="/css/site_mac.css">';
else if (navigator.appName == "Netscape" && parseInt(navigator.appVersion) >= 5)
style = '<link rel="stylesheet" type="text/css" href="/css/site_pcns6.css">';
else if (navigator.appName == "Netscape")
style = '<link rel="stylesheet" type="text/css" href="/css/site_pcns4.css">';
else
style = '<link rel="stylesheet" type="text/css" href="/css/site_pc.css">';
document.write(style);
.)
Jim Cochrane
> On Tue, 16 Mar 2004 22:21:37 -0700, Martin Leese
><ple...@see.Web.for.e-mail.INVALID> wrote:
>
>>Jim Cochrane wrote:
>>...
>>> My question is: Is it possible and wise to use conditional processing to
>>> implement a web page using CSS such that the page displays reasonably
>>> well in all or almost all currently used browsers (without making the
>>> page overly simple or basic)?
>>
>>Visit http://www.richinstyle/
>
> Actually http://www.richinstyle.com/
Thanks, to Martin, and to Stephen for the correction. Looks like a great
resource. (Where would we be without such generosity manifesting all over
the internet? Back in the tech. dark ages, I suspect.)
Jim Cochrane
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>
>> My question is: Is it possible and wise to use conditional processing to
>> implement a web page using CSS such that the page displays reasonably
>> well in all or almost all currently used browsers (without making the
>> page overly simple or basic)?
>
> Possible, yes. Wise, yes. Commercially realistic, no. If the question
> regards a design-and-forget template for no-clue users to periodically
> update "content" through a click-and-hope "CMS" or anything else
> involving the word "interface", forget it.
I'm thinking more in terms of my own web design, which (being a
command-line and vi addict) is going to be at least 95% hand coded.
However, regarding the "Wise, yes", it sounds like some others posting to
this thread would disagree.
BTW, is it commercially unrealistic becuase it's too expensive to maintain?
>> By conditional processing, I mean sending different html data to the
>> user's browser depending on what browser he is using.
>
> HTML is never data. Data is the bit *between* the tags. If I've read
> you correctly, you know that.
Yes - I was being lazy and imprecise.
>> ...good-looking result in multiple browsers - that the ideal of using one
>> set of html and css instructions for all browsers is not always possible.
>>
>> If the answer is not no (e.g., yes, or sometimes) then, of course, the
>> next question is how to do this - e.g., use Javascript?
>
> JavaScript. Funny. MS conditional comments are almost completely
> reliable (excepting their use for DOCTYPE declarations, thanks to a
> Mozilla bug).
Do you mean that javascript is unnecessary because conditional comments can
be used for MS IE? (I guess I need to familiarize myself with what
conditional comments are - Know any URLs for this?)
>> ...since I suspect this is not a simple issue,
>
> I think it *is* simple. MS has a taken a different evolutionary path
> to other browser vendors, concentrating on the ability to render text
> in as many languages as possible. Other vendors have concentrated on
> taking "western" scripts and prettifying as much as possible with
> CSS2. Ignoring the features of other browsers so the result will look
> the same as MSIE is stupid.
Well, my main development platform is mozilla on Linux, so it's not in my
nature to regard MSIE as the main browser for development (although it
needs to be paid attention to, of course, due to its obnoxious ubiquity).
But I think you're implying here (like, I believe, Spartanicus was) that,
with some compromises, it's not too hard to design a site using css that
looks good in all currently used browsers. (Yes?)
>> that it's probably controversial,
>
> A simple statement like, "The sky is blue," is controversial to the
> CIWAHians.
>
> "Only for twelve hours a day."
>
> "Well no, actually. For some value of 'hours a day' that depends
> critically on the user's latitude."
>
> "When you write 'latitude', ITYM 'latitude and time of year,' don't
> you?"
So, what color is the frequency 440Hz [A above middle C]? :-)
> "Oh yes. It's off-topic here, anyway. But to try to get back to
> something on-topic, the concept of 'colo(u)r of the sky' is inherently
> inaccessible and should be avoided when authoring in a WWW context.
> Consider, what does 'The sky is blue' mean to a blind user using a
> speech browser?"
>
>
>> and that reasonable people will have different views about it.
>
> Reasonable people don't frequent Usenet.
Now that's one I've never heard. :-)
>> Hopefully the resulting discussion will be useful to others, too.
>
> Unlikely. It's off-topic here anyway and belongs in .site-design or
> .stylesheets
Sorry about that. I'll know better next time where to post to.
Thanks!
Jim Cochrane
> On 16 Mar 2004 11:59:11 -0700, Jim Cochrane
><j...@shell.dimensional.com> wrote:
>
>>By conditional processing, I mean sending different html data to the
>>user's browser depending on what browser he is using.
>
> _IMHO_ (and I've been flamed for this many times) it's just about
> excusable to do this if NS4 legacy support is vital to you. For
> anything newer, you don't need to go to this amount of trouble.
Are you implying that it's sometimes easier to come up with something that
looks good in NS4, as well as modern browsers, by using conditional
processing, rather than just designing your site such that NS4 does not use
(does not see) the css?
> It's also completely impractical to do it at all well. The main
> problem is next week's new browser that you didn't know about when you
> coded the test.
>
Thanks!
Darin McGrew
>> The common approach to deal with NS4 is to hide all css from it by
>> exploiting the fact that it doesn't support @import, there's Tantek's
Jim Cochrane <j...@shell.dimensional.com> wrote:
> However, in my admittedly limited experience so far, it seems it's not
> always easy to produce an html document that looks acceptable in NS4
> without style sheets, while still using css for browsers that support it -
> that is, often when I put something together that looks good with css, it
> looks lousy in NS4.
The basic idea is to start with a structured HTML document that is
accessible, and to enhance its presentation with optional CSS. If your
document *needs* the CSS to make sense, then you're misusing CSS. But it's
okay if the document isn't beautiful without CSS.
At this point, you've covered browsers that can handle your CSS properly
and browsers that ignore your CSS. By hiding your CSS from browsers that
butcher your CSS (e.g., NN4), you nudge them into the category of browsers
that ignore your CSS.
> So, I take it you would regard this example as bad practice:
>
> http://www.directorystore.com/index.jsp
What about browsers that support CSS, but have JavaScript disabled?
--
Darin McGrew, da...@TheRallyeClub.org, http://www.TheRallyeClub.org/
A gimmick car rallye is not a race, but a fun puzzle testing your
ability to follow instructions. Upcoming gimmick car rallye in
Silicon Valley: Clue (Saturday, April 3)
Stephen Poley
wrote:
>In article <ncke50lkj727prsu7...@news.spartanicus.utvinternet.ie>, Spartanicus wrote:
>> The common approach to deal with NS4 is to hide all css from it by
>> exploiting the fact that it doesn't support @import, there's Tantek's
>
>However, in my admittedly limited experience so far, it seems it's not
>always easy to produce an html document that looks acceptable in NS4
>without style sheets, while still using css for browsers that support it -
>that is, often when I put something together that looks good with css, it
>looks lousy in NS4.
This is perhaps true. The question is whether that matters.
I've tried keeping a certain degree of support for NS4, and I've noticed
that the percentage of NS4 visitors I have is rather larger than most
sites report. But as of this month even I am getting only about 1% NS4
visitors. I suggest that it is sufficient that they are able to read the
site - it really doesn't seem worth worrying about cosmetics for them
any longer.
Spartanicus
>> The common approach to deal with NS4 is to hide all css from it by
>> exploiting the fact that it doesn't support @import, there's Tantek's
>
>However, in my admittedly limited experience so far, it seems it's not
>always easy to produce an html document that looks acceptable in NS4
>without style sheets
The very few users who still use NS4 will more than likely delight in
encountering a site that uses proper markup, the absence of styling
should never be a hindrance, if you think it is, then take a hard look
to see if your markup is truly solid, or perhaps you are to precious
about your formatting and it's importance to users.
>For example, one page I'm working on displays, from
>top to bottom, the navigation section, and then the main content (rather
>than the nav. section appearing on the left and the content on the right);
>and the main content, which does not use tables, ends up with what should
>be two columns, left and right, appearing in the same place, among other
>problems. In other words, it looks great in Mozilla, but in NS4 it looks
>quite atrocious.
You should not worry about loss of your layout, but proper content
linearity is important imo, not just for NS4 users or others who get to
see the content without styling, but mainly for bots (Google etc.) and
things like screen readers. The common way to restore content linearity
is to use css absolute positioning, however this approach has intrinsic
drawbacks.
Opinions vary on whether absolute positioning causes more problems than
it solves, personally I'm in favor of using it. Take a look at one of my
sites with Opera: http://www.pan-europe.utvinternet.ie/ hit Ctrl-G to
disable the author stylesheets and you'll see that the top navigation
bar, the top bread crumb bar and the panel on the left actually appear
last in the document's source, and that the real content is on top.
Text zooming is adversely affected by the use of absolute positioning
(use IE and Mozilla to test this), but if care is taken the adverse
effect need not be to drastic.
>> All of this can be incorporated into the same stylesheets, no need to
>> serve a different one (which btw would cause far more trouble than it
>> solves).
>
>Do you mean that it's possible, and not too difficult or ugly, to code one
>css file to take advantage of quirks in the different browers (such as:
>"This browser will just ignore this section and use that one.", "This
>brower will interpret this section as a comment.", etc.) such that some
>browsers will use a particular set of "selector definitions" (or whatever
>the correct terminology is) and other browsers will use a different set?
Yes, in my experience if you restrict yourself to using the parts of css
supported by IE then the amount of hacks you need to incorporate into
you stylesheets is limited. If you take a look at the 2 stylesheets of
the aforementioned site: http:/www.pan-europe.utvinternet.ie/css/one.css
This above stylesheet imports the main stylesheet, thereby hiding it
from NS4: http:/www.pan-europe.utvinternet.ie/css/two638.css
Lines 8, 10, 12, 21-24 and line 26 all contain hacks to aid in cross
browser rendering (= helping the old beast IE).
>>>If the answer is not no (e.g., yes, or sometimes) then, of course, the
>>>next question is how to do this - e.g., use Javascript?
>>
>> Not for this.
>
>So, I take it you would regard this example as bad practice:
>
>http://www.directorystore.com/index.jsp
Very much so, skilled developers don't need that sort of nonsense, and
the drawbacks are nasty.
--
Spartanicus
kchayka
>
> (I guess I need to familiarize myself with what
> conditional comments are - Know any URLs for this?)
google should point right to the msdn page
<URL:http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=internet+explorer+conditional+comments&btnG=Google+Search>
> But I think you're implying here (like, I believe, Spartanicus was) that,
> with some compromises, it's not too hard to design a site using css that
> looks good in all currently used browsers. (Yes?)
You can do quite a lot as long as you keep the attitude that a page
doesn't have to look the same in all browsers/environments, which isn't
attainable anyway.
--
Reply address is a bottomless spam bucket.
Please reply to the group so everyone can share.
Andy Dingley
<j...@shell.dimensional.com> wrote:
>Are you implying that it's sometimes easier to come up with something that
>looks good in NS4, as well as modern browsers, by using conditional
>processing, rather than just designing your site such that NS4 does not use
>(does not see) the css?
NS4 is entirely broken when it comes to doing any sort of layout work
with CSS. If you care, you're stuck with tables (and quite possibly
1x1.gif). The @import trick allows you to avoid _breaking_ an NS4 page
with some CSS, but it isn't (of itself) an alternative strategy for
controlling layout.
OTOH, almost all other browsers either understand enough CSS to be
useful, or little enough CSS to ignore everything and degrade
gracefully. This includes mobile phones, Lynx, PDAs etc.
--
Smert' spamionam
Chris Sharman
>>>exploiting the fact that it doesn't support @import, there's Tantek's
>>
>>However, in my admittedly limited experience so far, it seems it's not
>>always easy to produce an html document that looks acceptable in NS4
>>without style sheets, while still using css for browsers that support it -
>>that is, often when I put something together that looks good with css, it
>>looks lousy in NS4.
>
> This is perhaps true. The question is whether that matters.
>
> I've tried keeping a certain degree of support for NS4, and I've noticed
> that the percentage of NS4 visitors I have is rather larger than most
> sites report. But as of this month even I am getting only about 1% NS4
> visitors. I suggest that it is sufficient that they are able to read the
> site - it really doesn't seem worth worrying about cosmetics for them
> any longer.
That's one valid approach, but if you wish to be less brutal to your NN4
users, quite a lot of 'cosmetic' CSS does work in NN4 - if you feel it's
worth the effort you can split your CSS into 'universal' and 'hidden
from NN4' (by trial & error, adding a line at a time to the 'universal').
If you do this, you definitely need to test with NN4 after every change
(html as well as css) - the slightest thing that NN4 doesn't like can
cause catastrophic misrendering. Closing table elements is good for NN4
too, although redundant according to standards & most browsers.
I used to do this, and continue to maintain it on old pages for the time
being, but I do think we've now reached the stage where it's unnecessary
for new pages.
Chris
Brian
["CIWAHian" rant snipped]
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>
>> reasonable people will have different views about it.
>
> Reasonable people don't frequent Usenet.
So it seems. ;-)
--
Brian (remove "invalid" to email me)
http://www.tsmchughs.com/
Jim Cochrane
> Spartanicus wrote:
>>> The common approach to deal with NS4 is to hide all css from it by
>>> exploiting the fact that it doesn't support @import, there's Tantek's
>
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>> However, in my admittedly limited experience so far, it seems it's not
>> always easy to produce an html document that looks acceptable in NS4
>> without style sheets, while still using css for browsers that support it -
>> that is, often when I put something together that looks good with css, it
>> looks lousy in NS4.
>
> The basic idea is to start with a structured HTML document that is
> accessible, and to enhance its presentation with optional CSS. If your
> document *needs* the CSS to make sense, then you're misusing CSS. But it's
> okay if the document isn't beautiful without CSS.
>
> At this point, you've covered browsers that can handle your CSS properly
> and browsers that ignore your CSS. By hiding your CSS from browsers that
> butcher your CSS (e.g., NN4), you nudge them into the category of browsers
> that ignore your CSS.
Yes, this seems to be generally what is recommended for producing a good,
standards-conforming site (as I've discovered from this thread and from
reading ciwah and ciwas and various web sites).
Here is one site I've been working on (for a free-software product I've
written):
http://eiffel-mas.sourceforge.net/
It's not a masterpiece, but I think I've done a fair job of making it
presentable in css-savvy browsers and not too ugly in NN4. However,
criticisms and suggestions are welcome. (Though it's fine in Mozilla,
I haven't been able yet to get the sidebar to look good with a different
background in IE 5, so I've hacked the style sheet to bypass this for IE.)
(The style sheet file is at:
http://eiffel-mas.sourceforge.net/styles.css
. Yes, it needs some cleaning up.)
Thanks very much for the advice.
>
>> So, I take it you would regard this example as bad practice:
>>
>> http://www.directorystore.com/index.jsp
>
> What about browsers that support CSS, but have JavaScript disabled?
I guess that site would not be very usable for those without JS, eh?
Jim Cochrane
>>>>The common approach to deal with NS4 is to hide all css from it by
>>>>exploiting the fact that it doesn't support @import, there's Tantek's
>>>
>>>However, in my admittedly limited experience so far, it seems it's not
>>>always easy to produce an html document that looks acceptable in NS4
>>>without style sheets, while still using css for browsers that support it -
>>>that is, often when I put something together that looks good with css, it
>>>looks lousy in NS4.
>>
>> This is perhaps true. The question is whether that matters.
>>
>> I've tried keeping a certain degree of support for NS4, and I've noticed
>> that the percentage of NS4 visitors I have is rather larger than most
>> sites report. But as of this month even I am getting only about 1% NS4
>> visitors. I suggest that it is sufficient that they are able to read the
>> site - it really doesn't seem worth worrying about cosmetics for them
>> any longer.
>
> That's one valid approach, but if you wish to be less brutal to your NN4
> users, quite a lot of 'cosmetic' CSS does work in NN4 - if you feel it's
> worth the effort you can split your CSS into 'universal' and 'hidden
> from NN4' (by trial & error, adding a line at a time to the 'universal').
>
> If you do this, you definitely need to test with NN4 after every change
> (html as well as css) - the slightest thing that NN4 doesn't like can
> cause catastrophic misrendering. Closing table elements is good for NN4
> too, although redundant according to standards & most browsers.
While the idea of doing this to make a site look reasonably good in NN4 is
appealing, it sounds like it's rather time consuming (i.e., expensive), no?
> I used to do this, and continue to maintain it on old pages for the time
> being, but I do think we've now reached the stage where it's unnecessary
> for new pages.
>
> Chris
Jim Cochrane
I think I'm tending to favor this approach so far. However, it looks like
the folks at http://www.richinstyle.com/ would disagree.
Jim Cochrane
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>>top to bottom, the navigation section, and then the main content (rather
>>than the nav. section appearing on the left and the content on the right);
>>and the main content, which does not use tables, ends up with what should
>>be two columns, left and right, appearing in the same place, among other
>>problems. In other words, it looks great in Mozilla, but in NS4 it looks
>>quite atrocious.
>
> You should not worry about loss of your layout, but proper content
> linearity is important imo, not just for NS4 users or others who get to
> see the content without styling, but mainly for bots (Google etc.) and
> things like screen readers. The common way to restore content linearity
> is to use css absolute positioning, however this approach has intrinsic
> drawbacks.
>
> Opinions vary on whether absolute positioning causes more problems than
> it solves, personally I'm in favor of using it. Take a look at one of my
> sites with Opera: http://www.pan-europe.utvinternet.ie/ hit Ctrl-G to
> disable the author stylesheets and you'll see that the top navigation
> bar, the top bread crumb bar and the panel on the left actually appear
> last in the document's source, and that the real content is on top.
>
> Text zooming is adversely affected by the use of absolute positioning
> (use IE and Mozilla to test this), but if care is taken the adverse
> effect need not be to drastic.
I think I've done a pretty good job of using absolute positioning on this,
rather simple, site:
http://eiffel-mas.sourceforge.net/
The content is up front to please the search engines and it doesn't look
too bad in NN 4.
>
>>> All of this can be incorporated into the same stylesheets, no need to
>>> serve a different one (which btw would cause far more trouble than it
>>> solves).
>>
>>Do you mean that it's possible, and not too difficult or ugly, to code one
>>css file to take advantage of quirks in the different browers (such as:
>>"This browser will just ignore this section and use that one.", "This
>>brower will interpret this section as a comment.", etc.) such that some
>>browsers will use a particular set of "selector definitions" (or whatever
>>the correct terminology is) and other browsers will use a different set?
>
> Yes, in my experience if you restrict yourself to using the parts of css
> supported by IE then the amount of hacks you need to incorporate into
> you stylesheets is limited. If you take a look at the 2 stylesheets of
> the aforementioned site: http:/www.pan-europe.utvinternet.ie/css/one.css
> This above stylesheet imports the main stylesheet, thereby hiding it
> from NS4: http:/www.pan-europe.utvinternet.ie/css/two638.css
>
> Lines 8, 10, 12, 21-24 and line 26 all contain hacks to aid in cross
> browser rendering (= helping the old beast IE).
Hmm, I looked at your site before tonight, but right now I'm having
problems connecting with your server - perhaps your web server is down
right now or something like that (e.g.:
telnet www.pan-europe.utvinternet.ie 80
Trying 194.46.8.65...
telnet: Unable to connect to remote host: Connection refused
)
I'll try to remember to look at it later.
>>>>If the answer is not no (e.g., yes, or sometimes) then, of course, the
>>>>next question is how to do this - e.g., use Javascript?
>>>
>>> Not for this.
>>
>>So, I take it you would regard this example as bad practice:
>>
>>http://www.directorystore.com/index.jsp
>
> Very much so, skilled developers don't need that sort of nonsense, and
> the drawbacks are nasty.
Thanks for your helpful feedback.
Jim Cochrane
> Jim Cochrane wrote:
>>
>> (I guess I need to familiarize myself with what
>> conditional comments are - Know any URLs for this?)
>
> google should point right to the msdn page
><URL:http://www.google.com/search?hl=en&ie=UTF-8&oe=UTF-8&q=internet+explorer+conditional+comments&btnG=Google+Search>
>
>> But I think you're implying here (like, I believe, Spartanicus was) that,
>> with some compromises, it's not too hard to design a site using css that
>> looks good in all currently used browsers. (Yes?)
>
> You can do quite a lot as long as you keep the attitude that a page
> doesn't have to look the same in all browsers/environments, which isn't
> attainable anyway.
I think I grok this now.
However, there's another potential problem that I suspect could come up
(and discussion of it is probably worthy of a separate thread), which is
that if one is designing a web site for a client, the client may have a
definite layout in mind, one that focusses more on layout than on content.
I'm sure that in many cases it will be possible to educate the client -
that it's more appropriate to design a site as we've been discussing, but
I think some clients will insist that this is what they want.
Comments?
Jim Cochrane
> I'll start by saying that although I've used html informally for a few
Thanks, everyone, for your helpful feedback and comments. I've learned a
lot from this discussion, though I've still more learning to do.
Mark Parnell
declared in comp.infosystems.www.authoring.html:
> However, there's another potential problem that I suspect could come up
> (and discussion of it is probably worthy of a separate thread), which is
> that if one is designing a web site for a client, the client may have a
> definite layout in mind, one that focusses more on layout than on content.
Design it to look the way they want in whatever browser they use, and it
doesn't matter if it looks identical in other browsers, as long as it
looks OK. Chances are they'll never look at it in more than one browser,
and the customers certainly won't know the difference.
--
Mark Parnell
http://www.clarkecomputers.com.au
Jim Cochrane
I think this make sense, although it could probably, usually, be
translated as: "Design it to look the way they want in the version of
IE that the client is using, and ...".
However, I think I would change that to: "Design it to look the way they
want in the version of IE that the client is using and to look as close
as possible to the way they want in all other versions of IE that are
commonly used, and ..." - or something like that.
Spartanicus
>I think I've done a pretty good job of using absolute positioning on this,
>rather simple, site:
>
>http://eiffel-mas.sourceforge.net/
You have issues:
http://www.spartanicus.utvinternet.ie/test/jim_cochrane.png (server is
down at the moment of writing)
Note the horizontal scrollbar and the spill over of text from the left
panel.
--
Spartanicus
Stephen Poley
wrote:
>In article <c3mamm$8p3$1$8302...@news.demon.co.uk>, Chris Sharman wrote:
>> That's one valid approach, but if you wish to be less brutal to your NN4
>> users, quite a lot of 'cosmetic' CSS does work in NN4 - if you feel it's
>> worth the effort you can split your CSS into 'universal' and 'hidden
>> from NN4' (by trial & error, adding a line at a time to the 'universal').
>>
>> If you do this, you definitely need to test with NN4 after every change
>> (html as well as css) - the slightest thing that NN4 doesn't like can
>> cause catastrophic misrendering. Closing table elements is good for NN4
>> too, although redundant according to standards & most browsers.
>
>While the idea of doing this to make a site look reasonably good in NN4 is
>appealing, it sounds like it's rather time consuming (i.e., expensive), no?
It certainly can be. I did it for quite a while on my site, but then
discovered that even testing NN4 after every change wasn't sufficient.
NN4 has intermittent display faults - most days the content of my page
would display but some days I would get only the background, on pages to
which I had made no changes whatever. So I've now reduced the
NN4-visible stylesheet to half a dozen very simple rules which seem to
work reliably.
Chris Sharman
> I think I'm tending to favor this approach so far. However, it looks like
> the folks at http://www.richinstyle.com/ would disagree.
I've just looked, and wouldn't rate their opinion.
Their "exhaustive all-browser CSS bug list" doesn't include ie6
(arguably the most important), and they advocate setting font sizes in
pixels - which decade is this website from ?
Chris
Neal
wrote:
> http://www.richinstyle.com/
Heh, went there in Lynx and in Opera. Two different websites... I get this
Sterling Relogation in Lynx.
So much for a style website that can't degrade.
kchayka
> In article <sc9l501slof30jrii...@4ax.com>, Stephen Poley wrote:
>> On 18 Mar 2004 22:28:16 -0700, Jim Cochrane <j...@shell.dimensional.com>
>> wrote:
>>
>> I've tried keeping a certain degree of support for NS4,
>> I suggest that it is sufficient that they are able to read the
>> site - it really doesn't seem worth worrying about cosmetics for them
>> any longer.
>
> I think I'm tending to favor this approach so far. However, it looks like
> the folks at http://www.richinstyle.com/ would disagree.
richinstyle hasn't been updated in years, probably not since NS6 was
first released. It is good as an historical reference, but I wouldn't
use it as a guide for modern browsers.
Jim Cochrane
I don't know - A couple people from this NG mentioned it recently, I
believe.
Jim Cochrane
> Jim Cochrane wrote:
>
>> In article <sc9l501slof30jrii...@4ax.com>, Stephen Poley wrote:
>>> On 18 Mar 2004 22:28:16 -0700, Jim Cochrane <j...@shell.dimensional.com>
>>> wrote:
>>>
>>> I've tried keeping a certain degree of support for NS4,
>>> I suggest that it is sufficient that they are able to read the
>>> site - it really doesn't seem worth worrying about cosmetics for them
>>> any longer.
>>
>> I think I'm tending to favor this approach so far. However, it looks like
>> the folks at http://www.richinstyle.com/ would disagree.
>
> richinstyle hasn't been updated in years, probably not since NS6 was
> first released. It is good as an historical reference, but I wouldn't
> use it as a guide for modern browsers.
Is there a similar site around that is current?
Jim Cochrane
Damn - I guess you're using mozilla on Windows. I tested it on Linux.
I've changed it to allow a little more room for the left sidebar, and
changed the positioning to use cm instead of em. Don't know if that will
help, but if you get a chance to try it again, please let me know.
As far as the horizontal scroll bar, that really baffles me. I have
Mozilla 1.5 on Linux and when I grow or shrink the window, the text wraps
appropriately, no horizontal scroll bars. I wonder why it behaves so
differently in Windows. What version of Mozilla are you using?
If I'm going to run into these differences even with the same software
package on different OSs, it seems that what I"m trying to do is even more
challenging than I thought.
Thanks!
Mark Parnell
<sbpoleySpi...@xs4all.nl> declared in
comp.infosystems.www.authoring.html:
> So I've now reduced the
> NN4-visible stylesheet to half a dozen very simple rules which seem to
> work reliably.
I've done much the same thing. NS4 simply isn't worth the effort. As
long as the site is still usable.
Mark Parnell
declared in comp.infosystems.www.authoring.html:
> As far as the horizontal scroll bar, that really baffles me. I have
> Mozilla 1.5 on Linux and when I grow or shrink the window, the text wraps
> appropriately, no horizontal scroll bars. I wonder why it behaves so
> differently in Windows. What version of Mozilla are you using?
I don't get a horizontal scrollbar until my window gets down to about
420px wide, using Moz1.6 on Win2k.
Spartanicus
>>>http://eiffel-mas.sourceforge.net/
>>
>> You have issues:
>> http://www.spartanicus.utvinternet.ie/test/jim_cochrane.png (server is
>> down at the moment of writing)
>>
>> Note the horizontal scrollbar and the spill over of text from the left
>> panel.
>
>Damn - I guess you're using mozilla on Windows.
Opera (7.5p3, but 7.2x should behave similar).
>I tested it on Linux.
>I've changed it to allow a little more room for the left sidebar, and
>changed the positioning to use cm instead of em.
Bad move.
>Don't know if that will
>help, but if you get a chance to try it again, please let me know.
The spill over is gone, but the panel looks to wide now.
>As far as the horizontal scroll bar, that really baffles me. I have
>Mozilla 1.5 on Linux and when I grow or shrink the window, the text wraps
>appropriately, no horizontal scroll bars. I wonder why it behaves so
>differently in Windows.
To see the scrollbar open and close Opera's hotlist.
--
Spartanicus
Karl Smith
> Here is one site I've been working on (for a free-software product I've
> written):
>
> http://eiffel-mas.sourceforge.net/
I usually scoff at sites that have long-winded "copyright statements",
"privacy policies", "disclaimers" etc. But since that is market
analysis software, you *NEED* to put some serious disclaimer on that
page.
Courts in Astralia have held that you are publishing here if your site
is accessible from here, regardless of where you or the web server are
physically located. You wouldn't place an advert in an Australian
print magazine for such software without a disclaimer of liability for
investment decisions, would you?
> The style sheet file is at:
> http://eiffel-mas.sourceforge.net/styles.css
#navigation {
position: absolute;
position: fixed;
left: 0; top: 0;
font-size: 90%;
padding: 0 0 0 0.8em;
width: 3.75cm; height: 100%; <-- Woof!
}
Is that width supposed to be in centimetres?
--
Karl Smith.
kchayka
> In article <c3psuc$2aemr9$1...@ID-224809.news.uni-berlin.de>, kchayka wrote:
>>
>> richinstyle hasn't been updated in years, probably not since NS6 was
>> first released. It is good as an historical reference, but I wouldn't
>> use it as a guide for modern browsers.
>
> Is there a similar site around that is current?
Try http://macedition.com/cb/resources/abridgedcsssupport.html
Retlak
> >not always easy to produce an html document that looks acceptable in NS4
>
> This is perhaps true. The question is whether that matters.
>
> But as of this month even I am getting only about 1% NS4
> visitors.
Personally, I take a "hard line" with NS4 users. I ignore them. That
saves me a lot of work and loses very little.
You might say, "Ignoring minority users is bad in principle. Do you
also ignore blind users?". But the two cases are not comparable. The
0.4% of users who are blind will always be there. (Also, there's
nothing they can do about it.) The 1% of users who use NS4 is
decreasing every month, and if they can't be bothered to upgrade from
last century's browser, they're probably not promising customers.
YMMV of course.
Jim Cochrane
.In article <44m16091din7tc4cv...@news.spartanicus.utvinternet.ie>, Spartanicus wrote:
>> Jim Cochrane <j...@shell.dimensional.com> wrote:
>>
>>>>>http://eiffel-mas.sourceforge.net/
>>>>
>>>> You have issues:
>>>> http://www.spartanicus.utvinternet.ie/test/jim_cochrane.png (server is
>>>> down at the moment of writing)
>>>>
>>>> Note the horizontal scrollbar and the spill over of text from the left
>>>> panel.
>>>
>>>Damn - I guess you're using mozilla on Windows.
>>
>> Opera (7.5p3, but 7.2x should behave similar).
OK - Perhaps the table at the bottom was causing the scroll bar. The
table is now gone.
>>>I tested it on Linux.
>>>I've changed it to allow a little more room for the left sidebar, and
>>>changed the positioning to use cm instead of em.
>>
>> Bad move.
>>
>>>Don't know if that will
>>>help, but if you get a chance to try it again, please let me know.
>>
>> The spill over is gone, but the panel looks to wide now.
>>
>>>As far as the horizontal scroll bar, that really baffles me. I have
>>>Mozilla 1.5 on Linux and when I grow or shrink the window, the text wraps
>>>appropriately, no horizontal scroll bars. I wonder why it behaves so
>>>differently in Windows.
>>
>> To see the scrollbar open and close Opera's hotlist.
If you get a chance, I'd appreciate hearing how it looks now. I suspect
the scroll bar is probably gone, but the side bar may still be a problem in
Opera.
Thanks.
Jim Cochrane
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>
>> Here is one site I've been working on (for a free-software product I've
>> written):
>>
>> http://eiffel-mas.sourceforge.net/
>
> I usually scoff at sites that have long-winded "copyright statements",
> "privacy policies", "disclaimers" etc. But since that is market
> analysis software, you *NEED* to put some serious disclaimer on that
> page.
You're right. I've been neglecting this for too long.
I've added a disclaimer at the bottom. Being my first attempt, I think it
needs a bit more work. Comments are welcome.
http://eiffel-mas.sourceforge.net/
Thanks very much for pointing this out.
> Courts in Astralia have held that you are publishing here if your site
> is accessible from here, regardless of where you or the web server are
> physically located. You wouldn't place an advert in an Australian
> print magazine for such software without a disclaimer of liability for
> investment decisions, would you?
>
>
>> The style sheet file is at:
>> http://eiffel-mas.sourceforge.net/styles.css
>
> #navigation {
> position: absolute;
> position: fixed;
> left: 0; top: 0;
> font-size: 90%;
> padding: 0 0 0 0.8em;
> width: 3.75cm; height: 100%; <-- Woof!
> }
>
> Is that width supposed to be in centimetres?
I suppose. Is it a bad idea?
Jim Cochrane
> Jim Cochrane wrote:
>
>> In article <c3psuc$2aemr9$1...@ID-224809.news.uni-berlin.de>, kchayka wrote:
>>>
>>> richinstyle hasn't been updated in years, probably not since NS6 was
>>> first released. It is good as an historical reference, but I wouldn't
>>> use it as a guide for modern browsers.
>>
>> Is there a similar site around that is current?
>
> Try http://macedition.com/cb/resources/abridgedcsssupport.html
Nice site.
Thanks.
Spartanicus
>If you get a chance, I'd appreciate hearing how it looks now.
Opera is only a 3.5mB download and available for Win, several Nix
flavours and Mac.
>I suspect the scroll bar is probably gone
Nope, don't worry to much about it, a reload causes the horizontal
scrollbar to disappear, so it could be an Opera quirk.
--
Spartanicus
Jim Cochrane
> Jim Cochrane <j...@shell.dimensional.com> wrote:
>
>>If you get a chance, I'd appreciate hearing how it looks now.
>
> Opera is only a 3.5mB download and available for Win, several Nix
> flavours and Mac.
I'll have to break down and DL is for my Linux system soon, I suppose, if I
want to get serious about web site design.
>>I suspect the scroll bar is probably gone
>
> Nope, don't worry to much about it, a reload causes the horizontal
> scrollbar to disappear, so it could be an Opera quirk.
>
How odd.
Thanks for your help.