Best way to blow off Netscape 4 users

[Jonas Smithson]

I'm going to say something now that may seem to completely contradict
a previous post of mine, in which I basically said that taking a "who
cares" attitude about certain browsers (because of their non-standard
CSS rendering) makes no sense.

Well, you have to draw the line *somewhere*... and contradictory or
not, I've decided that I've suffered with Netscape 4 for long enough.
Coding workarounds for its brain-dead CSS rendering has literally
doubled my development time on some projects. It's the only browser I
can truly say that I *hate*, and now that its market share seems to
have slipped below one percent (I hope!), I'm saying "enough."

So I'm planning to simplify my CSS and JavaScript files by stripping
out most or all of the NS4-specific bloat, and never find out what
gibberish it makes out of my pages by the simple method of not
viewing them myself in NS4.

However, I may leave my JavaScript browser-sniffer in place. My
question is, what's the best way to blow off NS4 users? Here are
some options that I've thought of:

1. Do nothing special (even get rid of the sniffer).
Pros: Simplest thing to do. They're going to see garbage
anyway, and they're probably used to it.
Cons: They don't know that their browser is crap, and they'll
blame me (or my clients).

2. Put up a JavaScript alert box on every page load, saying that my
site doesn't support NS4, and did they know they can get a modern
browser for free?
Pros: -Performs a public service - educates them a little.
-Lets them know I know what's happening (i.e. I'm not
just oblivious/incompetent).
Cons: -Might make pages unusable (too annoying) that might
otherwise be readable even in NS4.
-I'm not sure, but I don't think I can put a link (to
a browser upgrade site) in an ordinary JS alert.
I could certainly generate a special little window
(pseudo-"alert") with a link, but that would be work.
-If my sniffer screws up, I may end up annoying
non-NS4 users.

3. Use JavaScript to write some bold red text on every page with a note
(as in #2 above) and a link to some kind of browser upgrade site.
Pros: A little less harsh than the other options.
Cons: -More work for me. Do I have to check each page in NS4
to see that the warning is actually visible? Even if
it is, will I be able to restrain myself from "fixing"
each page after I see what garbage it looks
like in NS4?
-If my sniffer screws up, I may end up annoying
non-NS4 users.

Opinions? How do you handle this?

[Mark Tranchant]

Jonas Smithson wrote:

So I'm planning to simplify my CSS and JavaScript files by stripping
out most or all of the NS4-specific bloat, and never find out what
gibberish it makes out of my pages by the simple method of not
viewing them myself in NS4.
What the others said. Hide your CSS from NS4 and write decent,
well-structured HTML. See my site for an example:

http://tranchant.plus.com/

If you view it in Firefox, click on the little icon with an "A" on it on
the left edge of the status bar, and select Basic Theme. That's what my
NS4 viewers see.

Implementation hint: Firefox displays this icon if you give the
stylesheet a title in the <link> element.
However, I may leave my JavaScript browser-sniffer in place.
Don't. Javascript is a really poor way to browser-sniff. The least poor
way is server-side testing of the user-agent string, but even that is
notoriously unreliable. I do it only to detect IE to fix a couple of
things for its poor users, like <abbr>.
My
question is, what's the best way to blow off NS4 users? Here are
some options that I've thought of:

1. Do nothing special (even get rid of the sniffer).
Other than make sure the CSS call is hidden from NS4 (via @import, or
using media="screen,p rojection"), that's the right thing to do.
2. Put up a JavaScript alert box on every page load, saying that my
site doesn't support NS4, and did they know they can get a modern
browser for free?
No. Extremely annoying, and some people may be using public computers
that they can't upgrade. They'll leave your site pretty quickly.
3. Use JavaScript to write some bold red text on every page with a note
(as in #2 above) and a link to some kind of browser upgrade site.
See 2.
Opinions? How do you handle this?

See above. Sounds like you need to rely less on Javascript, too.

--
Mark.

[Chris Morris]

Jonas Smithson <sm************ @REMOVETHISboar dermail.com> writes:

First of all, I just tried the @import idea and immediately saw a
problem: it (naturally) doesn't block inline styles, or styles in the
head. Now I do use external styles whenever possible -- and I'd say
it's practical for about 90% of my styles. But the other 10% -- the
head and inline styles -- make a total mess when the external sheet is
blocked.
Styles in the head you can use a different hiding method for. There's
several that work on NN4, how about:
<style type='text/css'>
/*/*/
body {
break-ns4: yes;
}
/* */
</style>

You might even be able to use the same thing in an inline style <p
style='/*/*/ break-ns4: yes /* */'> but I haven't tested that because
I don't use inline styles myself on live pages.
The sidebar is an absolutely positioned element whose spatial
relationship to its parent will change without CSS. It may even appear
on an entirely different part of the page if the CSS is missing or
mangled.
If you've already checked this, my apologies, but what happens if you
increase your browser font size to somewhat larger than normal - does
the positioning relationship break?
Worse case, the reader may actually think the ancillary material
refers to a different "parent" paragraph than it actually does!
<p>This is a paragraph ... </p>
<div id="sidebar1">T his is the paragraph's sidebar</div>

<p>This is another paragraph ... </p>
<div id="sidebar2">T his is another paragraph's sidebar</div>

Or similar. That way the sidebars are close to the paragraphs without
CSS, and it's no harder than otherwise to position them close with it.
I suppose the simple retort to my example is "that's a true tabular
relationship, use a table" -- but I was trying to get away from table
code!
If it *is* true tabular relationship then use table code. It seems a
slight stretch in this case, but maybe not - URL for an example?
Or "don't put that in a sidebar" -- so I have to limit my design
options with CSS, although I didn't when using tables? Rhetorical
question: So are we to conclude that any page that requires specific
positional adjacencies as a clue to meaning can't be safely laid out
with CSS? I doubt many people here would say that. It seems to me that
many, many pages have such positional relationships!
Absolutely not, as you've guessed. However, the positional
relationships are *initally* in the HTML, in a good design. CSS
positioning may be used to _reinforce_ the relationship, but not to
initially indicate it.
I can understand Brian's objections to my suggesting too obtrusively to
readers anything about their browser: whenever I've visited a site that
said "This site best viewed in... [insert browser name]" my reaction
was always that the designer was both arrogant and clueless. So it's a
little strange that I'm contemplating something like that now. (I
emphasize "contemplating" ; I haven't decided, that's why I'm posting
here.) But if a Netscape 4 user may actually *misconstrue the meaning
of the content on a page* because the positions have been scrambled,
might it not be the lesser of several evils to warn them or even block
the page to them entirely?
There's a concept in web design called 'linearisation' . It means that,
when CSS is disabled, or tables are disabled (since it's important for
both layout methods), or both, the page still makes sense. If your
page doesn't do this then you have your HTML elements in the wrong
order and you need to fix that.
Regardless of whether I...
1. ...block most of the CSS with @import (and have the unstyled HTML
unpredictably blend with the inline styles that I can't block), or
Or move the inline styles into the stylesheet
<p style='...'> becomes <p id='...>
p#... {
...
}
*Either way* I run the risk of users seeing adjacencies that could be
highly misleading. This is a question of *meaning*, not fonts and
CSS should not be used as the sole indicator of meaning in the same
way that colour shouldn't.
backgrounds. That's why it seems like some kind of warning to NS4
visitors -- yes, on every page -- might be the least obnoxious
solution. (?)
No, fixing the HTML code so that it doesn't require CSS to make sense
is the *least* obnoxious solution. Give a URL for a problem case and
we might be able to make suggestions.
(P.S. Did I mention that I *hate* Netscape 4?)

Yes.

--
Chris

[Stan Brown]

"Jonas Smithson" <sm************ @REMOVETHISboar dermail.com> wrote in
comp.infosystem s.www.authoring.stylesheets:

Although, BTW, a *properly written* sniffer is not
as unreliable -- for this purpose, anyway -- as some people seem to
think.

Yes, and a green rhinoceros would be quite an attraction in the zoo.

The problem is, they are both mythical.

--
Stan Brown, Oak Road Systems, Tompkins County, New York, USA
http://OakRoadSystems.com/
HTML 4.01 spec: http://www.w3.org/TR/html401/
validator: http://validator.w3.org/
CSS 2 spec: http://www.w3.org/TR/REC-CSS2/
2.1 changes: http://www.w3.org/TR/CSS21/changes.html
validator: http://jigsaw.w3.org/css-validator/

[Stan Brown]

"Mark Tranchant" <ma**@tranchant .plus.com> wrote in
comp.infosystem s.www.authoring.stylesheets:

Jonas Smithson wrote:

2. Put up a JavaScript alert box on every page load, saying that my
site doesn't support NS4, and did they know they can get a modern
browser for free?

No. Extremely annoying, and some people may be using public computers
that they can't upgrade. They'll leave your site pretty quickly.

Agreed -- about as annoying as that voice at the checkout reminding
me after I scan EVERY DAMN ITEM, "Please place the item in the bag."

People who have Netscape 4 on their own computers already know that
later browsers are available, and they know how to find them if they
want them. And I echo Mark's comments about public computers.

--
Stan Brown, Oak Road Systems, Tompkins County, New York, USA
http://OakRoadSystems.com/
HTML 4.01 spec: http://www.w3.org/TR/html401/
validator: http://validator.w3.org/
CSS 2 spec: http://www.w3.org/TR/REC-CSS2/
2.1 changes: http://www.w3.org/TR/CSS21/changes.html
validator: http://jigsaw.w3.org/css-validator/

[Brian]

Neal wrote:

Brian wrote:

Neal wrote:

How about something styled display: none?
Why? What is behind this desire to remind users that their
browser is, according to the author, sub-standard?

Hey, not MY desire.

Sorry, Neal. I was on a bit of a tirade last night, wasn't I? My
apologies, it was not even your idea. :-/
if you're looking to send content to no-CSS viewers, that's it.

As is so often the case, the question should be "how do I...?" but
"should I...?"

--
Brian (remove ".invalid" to email me)
http://www.tsmchughs.com/

[Brian]

Jonas Smithson wrote:

Neal wrote:

How about something styled display: none? NN4 will see it, as the
CSS is imported. Only other folks to see it will really be Lynx.
That sounds clever;

You'll look clever, too, the next time the Googlebot indexes your
site. ;-) (See my response to Neal about this.)
(Although, BTW, a *properly written* sniffer is not as unreliable
Oh? Since js is optional, and ua strings can be forged, how can such a
thing be reliable?
It's true that most sniffers are unnecessary even for JavaScript's
own purposes, since object detection usually works better,
Object detection always works better in js.
many sniffers are badly coded -- but mine isn't.
Ok, genius, let's see your brilliant code! I wait with baited breath. :-)
First of all, I just tried the @import idea and immediately saw a
problem: it (naturally) doesn't block inline styles, or styles in
the head.
That's true. I didn't mention it, because I use neither of those
techniques. You may need to ditch any advanced inline css to protect
NS 4 from itself.
I'm starting to rely on CSS-positioning for entire page layouts
(not just for sections within an outer table), and that opens up a
whole other can of worms if most or all of the CSS is blocked.
I do all layout in CSS for my sites. All of it. I write in HTML
4/strict, which offers little presentation markup. NS 4 users get a
plain site, with no floats, no positioning, no styling at all. It was
the right choice for me. Only you can decide whether it's right for
your projects.
Consider the case of various main units of text which each "own" a
small ancillary sidebar, connected with CSS positioning.
Place it in a div after the paragraph, perhaps. We'd need a url to
provide more specific advice.
if a Netscape 4 user may actually *misconstrue the meaning of the
content on a page* because the positions have been scrambled,

Then you need to rethink the page. CSS is optional no matter what the
browser. Your content should be comprehensible without it. If it is
not, then rethink your markup.

--
Brian (remove ".invalid" to email me)
http://www.tsmchughs.com/

[Michael Fesser]

.oO(Jonas Smithson)

Brian wrote:

it is trivial to simply hide all CSS from it.
@import "hidefromns4.cs s";

I hadn't thought of that. Do you mean fetch *all* my external styles --
i.e. my main external stylesheets -- via @import rather than via <link
rel... etc. [...]

You can also stay with the link-element and an media-attribute. NN4 only
recognizes media='screen'.

From time to time I add a simple stylesheet only for NN4 (could be
written in JSSS for example), but the default stylesheet for all other
browsers is written according to the CSS 2.1 standard without any hacks.

Micha