The Design of Software (CLOSED)

A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.

The "Design of Software" discussion group has been merged with the main Joel on Software discussion group.

The archives will remain online indefinitely.

Web Design for Older Browsers

If you read a book or take a class on web design, chances are the instructor is going to do his or her best to try to drum into your head the idea that you do your best to make your pages look nice on older web browers.  I think now the security is finally (and long overdue) starting to take a more central role that this practice should be abandoned.  Not that I don't believe in accessiblity.  I just don't think your javascript should have make checks to verify whether or not you are using IE 4, for example.  So what if the user's browser doesn't support css level 3?  In the past, we have always been worried about alienating users with older browsers, but look at it like this:

Someone with an older browser tries to go to your site, and fails.  No big deal.  This time.  But they go somewhere else, and that looks horrid, too.  Then they go somewhere else and that also looks awful.  Now this IE 4 user, who has probably never patched his system, may be seriously thinking about paying a visit to the windows update site to get a newer, theoritically more secure, browser.  He may even switch to firefox or opera, thus creating more competition.  A similar strategy could be used with desktop apps to be sure they update the operating system as well.  Anything that handles sensitive information could refuse to work correctly on older or unpatched systems, thus prompting users to learn to keep them up to date.

What does everyone think of this approach?


BTW, I think today is supposed to be the day the full page NY Times Firefox ad runs.
Joel Coehoorn
Thursday, December 16, 2004
 
 
I worked for a company that decided to reject every browser that was too low a version. Unfortunately, for the fool who did the browser check, it also rejected everyone with a higher revision browser. It was IE 5.0 or nothing. Not 4.0.1 or 5.5. They dealt with doctors. The web front end was to allow doctors to order and check lab results. It took about 1 month to discover the bug and fix it. By that time, all the physicians had cancelled the web service.
Peter
Thursday, December 16, 2004
 
 
Several thoughts.

1. Some corporations may still be standadized on old browsers, as bizarre as it sounds. Citigroup was standardized on Netscape 4 until about 2002 - 2003, AOL was also on Netscape 4 the last I heard, also in the same time period.

2. People were more rigid in their "Anti Netscape" stance before Mozilla, and any site that has a "I see you're not using IE 4+ or higher, please upgrade to the latest browser" detection script is a relic. These teachers are likely carrying over that same attitude.

3. Cross browser/cross platform compatability is still an issue. Things don't render the same everywhere. Just look at slashdot in mozilla. Why does the left column overlap the center?

4. Good design should "fail gracefully" and good usability should downgrade gracefully.
www.MarkTAW.com Send private email
Thursday, December 16, 2004
 
 
"What does everyone think of this approach?"

It's not going to help me from browsing on my cellphone -- which I do quite regularily.  I don't expect perfect rendering but I want to see something!

The same is true of older browsers -- I'm sure if you looked at any of our sites in Netscape 4.x they're going to look strange -- maybe even bad.  But you'll see something.
Almost Anonymous Send private email
Thursday, December 16, 2004
 
 
I want to re-iterate that I'm not against designing for accessibility.  The site should scale down well for the mobile web and be semantically correct, etc.  And it shouldn't intentionally make it difficult on older browsers.

At the same time, you shouldn't have to bend over backward to make sure it will run on older systems.  There needs to be more incentive out there for people to make sure they have the most recent security updates, and this would be one good way to begin providing it.
Joel Coehoorn
Thursday, December 16, 2004
 
 
My sites probably look like very basic HTML on old browsers. They look like pretty much plain text on phones. They ARE plain text in Lynx. Aiming for things to look great no matter what platform is a fool's errand, you'll never do it. There's always platforms you won't expect.

The whole point of standards markup (well, one of the points) using XHTML/CSS is that they degrade on older browsers without you having to do *anything*.

As soon as you get onto client side scripting that's where you start to have to work. Hence, all sites i've worked on have the absolute minimum (usually none) of javascript, etc. Unless you're gmail, and have enough staff to write JS for every browser bug you can think of, it's very very rarely worth trying.

I think that these days most web design courses will be teaching this philosophy rather than the old one, if they're worth anything at all.
Andrew Cherry Send private email
Thursday, December 16, 2004
 
 
The problem with Netscape is that version 4.x thinks it knows CSS when it in fact doesn't fully do it. CSS does not degrade nicely as it should here. I don't know whether other browsers have the same problem, but for Netscape 4.x there is the @import workaround at least.
Peter Monsson Send private email
Thursday, December 16, 2004
 
 
It's always going to be a judgement call. How much do you care? How badly does it break? Who's the target audience? Is the extra money spent on design worth the customers who might be using an older browser?
www.MarkTAW.com Send private email
Thursday, December 16, 2004
 
 
And don't forget the collection of IE CSS flaws as well.

On the other hand, you can always make a CGI and two style sheets and use the browser header to determine which one is sent.

Which is much simpler than changing your HTML generation code depending on the browser.
Flamebait Sr.
Friday, December 17, 2004
 
 
I think the question I'm really trying to ask is:

Is it a good idea to use new software and more noticable upgrades or eye candy as a way to ~encourage~ users to be more responsible about how they keep their computers and browsers patched and secured?
Joel Coehoorn
Friday, December 17, 2004
 
 
Unfortunately, clients wouldn't want their website to look noticably crappy in commonly used browsers.  That's just a fact.  They don't care about making people upgrade.
Almost Anonymous Send private email
Friday, December 17, 2004
 
 
Yeah... Because when I see crappy design that doesn't work, I just immediately think of security. Doesn't everyone?
www.MarkTAW.com Send private email
Friday, December 17, 2004
 
 
The people who preach about web design and degrading on older web browsers are stuck in a 1999 time warp. Fortunately the people who visit your site are not. Anyone who looks at browser stats from their weblogs knows that any version of IE before 5.01 is well under 0.1% these days, Netscape 4 is gone, gone, gone, and all the old Mozilla/Opera/Firefoxes are gone too since the people who installed those are early adopters who get the latest thing. Remember the old stories about impoverished africans with vt100s dialed up at 300 baud using Lynx? Guess what: they don't exist! Out of millions of hits on the sites I control lynx is firmly ensconced at exactly ZERO market share.
Joel Spolsky Send private email
Sunday, December 19, 2004
 
 
Personally, when I talk about expected degradation of experience, i'm no longer talking about age. When I plan for degradation i'm thinking about things that will still work on PDAs, Phones, etc. For most people, that's not relevant, but for some, it is. The problem with a lot of "old school" design is that there's no way it would work nicely on small screens - people designed for, say 640x480+ - but that doesn't work on my Smartphone. Modern XHTML+CSS design will, if you do it properly (you might have to provide a seperate stylesheet, you might not).
Andrew Cherry Send private email
Sunday, December 19, 2004
 
 
Lynx may not be important per se, but it's one of the browsers that I test against because it's a good indicator of whether a site will work with nontraditional user agents that I *do* care about but can't test. If a site doesn't work in Lynx because it relies on JavaScript, search engines probably won't be able to index it. If a site doesn't work in Lynx because necessary information is only presented in a visual way (graphics used for text, form controls related to their labels only by position, etc) it's probably going to be a problem for blind people. Common sense requires that I not exclude search engine indexers, and federal law requires that I not exclude blind people. Lynx is the most useful canary in that particular mine shaft.

I agree about the decreasing importance of older graphical browsers like Netscape 4. It's not a problem to make a site useable in those browsers, but making it look the same is really cramping my style. I'll be happy when institutional policy catches up with reality in that area.
comp.lang.c refugee
Sunday, December 19, 2004
 
 
I agree with Andrew and the comp.lang.c refugee. If you ensure that your site stays readable without CSS just relying on the (X)HTML markup itself, it shouldn't be a problem on older browsers. I also use Lynx for testing this, but using the Developer Extension for Firefox I find myself more often just switchiung off CSS by using Ctrl-Shift-D to verify compatability.

I'm definitly not designing for NS 4.x or IE 4.0 or any older browser, since (as Joel correctly stated) they are gone in real life. Even on our old HPUX 10.20 machines, people use Firefox or Mozilla in the meantime, rather then NS 4.x (which was the only brower running on that plattform up to two years ago).
Gerd Riesselmann
Monday, December 20, 2004
 
 
Read the last paragraph here http://daringfireball.net/colophon/ I'ts better than the usual 'Best viewed with...' comment you still see on some sites.
MT Heart
Monday, December 20, 2004
 
 
"It's not going to help me from browsing on my cellphone -- which I do quite regularily.  I don't expect perfect rendering but I want to see something!"

What kind of browsing people do on cellphone besides downloading a new ringer?  Airtime is so expensive for web access that it ain't worth it.
Le Poete Send private email
Monday, December 20, 2004
 
 
"What kind of browsing people do on cellphone besides downloading a new ringer?  Airtime is so expensive for web access that it ain't worth it."

I have $50 CDN unlimited GPRS access that I can use from my cellphone or my laptop via bluetooth.  My phone is a SonyEricsson P800 -- Symbian OS, Opera for the browser, bright 320x480 screen.  I can read pretty much anything I want online and it works quite well.  I regularily read these forums from it.

I once used the browser in the phone to browse to the homepage of my car dealership to get the number of the roadside assistence because I locked my keys in my car.
Almost Anonymous Send private email
Monday, December 20, 2004
 
 
...and I've never downloaded a ringer...
Almost Anonymous Send private email
Monday, December 20, 2004
 
 
That may be, but have you downloaded a ringer for a ringer?
The Dude
Wednesday, December 22, 2004
 
 
About 1% of the traffic to my site is from Netscape 4.x. The question is which 1% is it, and is it traffic I want to lose?
www.MarkTAW.com Send private email
Friday, December 31, 2004
 
 
Coincidently, I spotted some real life Netscape 4.x users during the holidays.

The first was a bank employee. They use OS/2 at the bank (I heard it is still common in this business) and since this OS is - well - rather old, this seems to be the newst browser available. At least there is no Mozilla/Firefox available for OS/2 at the according download pages.

The second was my wife's sister. She used Netscape 4.02 (the internet connection software choosed it as default and she never changed it, though she had IE 6 installed). Now she's using Firefox, however.
Gerd Riesselmann
Monday, January 03, 2005
 
 

This topic is archived. No further replies will be accepted.

Other recent topics Other recent topics
 
Powered by FogBugz