Please forgive the following newbie question: should one still be testing in ie5,safari,crome,opera etc. when they have so little market share? I only test in ie6,ie7, and firefox. Thanks, cmartin19
You remember, oh, maybe it
You remember, oh, maybe it was 10 years ago already... if you were online back then, do you remember what you used for Internet searching? Prolly Alta Vista, right? Or AskJeeves or something. Suddenly, Google appeared, and within, what, a few months? It became millions of people's default search engine. Kinda scary.
With regards to browsers, market share alone isn't what should guide you. Safari, Opera and Chrome are all pretty low but any one of them could get popular for any reason, even something as silly as a popular movie that prominently shows it or something.
Right now, IE6, 7 and FF will cover 90% of your butt, sure. Older IE's aren't going to get popular again (we hope, O gawd no!) so then testing in them is more for when you have clients or a targetted group who you know is stuck using those browsers. For instance, I'm building a fire dept site. Because the neighbouring town's fire dept is stuck with Windows 2000 machines, I am writing and testing for IE5.5 as well as my regular assortment. For my other sites, I usually just take a quick look in IE5.5 and if some content is covered up I'll write a little width hack for its box-model problems-- I don't usually try to make it look pretty, just still usable. I would recommend this if you are writing ANY commercial site-- the small % of people stuck on older OSes won't grow, but they may still be a large enough % of visitors to mean it matters to your client (the bigger or more rural, the more likely).
Fortunately, if you write decent, semantic, valid, and generally well-written code, you should find Opera, Safari and Konqueror looking pretty much the same as FF (or better as FF has some bugz that can drive you up the wall).
Don't fall into the trap of
Don't fall into the trap of believing you must test for every browser ever written, it's a fallacy, and a folly perpetrated by the web dev community stemming from the dark days of the browser wars and of the emergance of more standards compliant modern browsers.
You should code to Standards, this way the majority of good well behaved modern browsers will render your work as you wish only then do some checks for legacy browsers such as IE6 and to some extent IE7 ( although this browser is much improved it's not as much improved as MS would have us believe)
Thinking that you need to worry about browsers such as 'Chrome' is not logical, if that browser has problems then it is up to Google to fix it with updates and patches , if we hunt around for fixes constantly it will only have the effect of making that browser appear to have no problems while having cost us in terms of dev time which equates to cost. Installing browsers such as Chrome or any other open source style browser is a matter of choice, if it doesn't perform then the user should move to something that does until such time as Google or whomever fixes things.
There is of course another side to this argument which is where certain users are locked into browsers with no choice an example might be safari on OS predating OS X 10.4 ? which can't update to Safari 3 in these cases it behoves us to attempt to produce a reasonable result; but that ought to be the watchword 'reasonable' don't attempt to get all browsers rendering identically, as it is simply too time consuming and in a commercial market simply not viable.
Well put Stomme !
I myself strongly encourage my viewers to get the latest releases of popular browsers. If the viewer is using IE 6.0 or below, I wount even let them see my website. That is a proactive way to force the slow-moving web users to comply with newer standards. "If it aint broke, dont fix it". Well now it is broke so they have to fix it! hahahaha
Certain browsers are dead
Certain browsers are dead and buried IE5 is an example if, in a commercial situation, you are obliged to cater for them then it should be clear that this is an exception and a special case along with explicit time costs involved, unless for a commercial project please don't test for old browsers such as IE5 otherwise the flipping things will never die and users who can upgrade never will.
thank you for help
Thanks for the advice. I will use it!
Quote:If the viewer is using
If the viewer is using IE 6.0 or below, I wount even let them see my website.
After trying to use JAWS with Firefox (half the functionality lost and difficult as hell to use), I'm more aware of how much some people must rely on IE. If you don't have Windows XP, it's going to be pretty hard for you to get IE7. I keep this in mind.
I've been thinking about this for some time, and have decided that the developer has to answer this question for themselves:
with the browsers you are willing to support and those you are not, does your site/company have a sign out in front that says No Jews Or Blacks Allowed, or is it saying No Shirt, No Shoes, No Service? I would NOT build a commercial website today without support for IE6, and I always take a look in IE5.5-- I don't Worry about IE5.5, but I do make sure 100% of my content is available. Because that's accessibility and usability. "Disabled" doesn't just mean blind or in a wheelchair. The fiscally disabled can seriously get to my content and use it. I will not waste time making my sites pretty for them-- that's "design" and is always lower on the totem pole than accessibility.
Seriously it's so easy for us to say, upgrade your OS. But we're not paying for it.
Now upgrade that OS with a $1000 screen reader. Or any other assistive equipment you need to buy or update with that new OS.
Better yet, let's have all those tax-haters pay up so the local libraries can stop using Windows 98 so their patrons aren't surfing with IE5. Imagine the cost of replacing the software of dozens of machines, or worse, replacing the machines themselves because they're just too old to run Vista well.
Chrome, I don't worry about. It's beta. But I test in it. Why? Because somewhere I've done a mistake in my code and often only a single browser shows it, alerting me before I get to my normal "revalidate" schedule. Sometimes this is IE, sometimes it's Opera, sometimes it's Saffy/Chrome. And I won't be hustling to change any design features I don't like when my sites are viewed in Chrome because it left beta and suddenly became 96% of the browser market.
What Hugo said is generally correct though-- if you code to standards you'll generally be covering your butt where it's important anyway. But then, none of us code to browsers, do we? : )
Quote:What Hugo said is
What Hugo said is generally correct though
Phew glad to know I can get a few things generally right, makes the years doing this c**p worthwhile I guess
Lawlz Hugo, it's my way of
Lawlz Hugo, it's my way of generally disagreeing with you without completely disagreeing with you : )
Oh so you want to disagree
Oh so you want to disagree with me? feel free I can take it
Bring it on, crusty! You
Bring it on, crusty!
You may be older and wiser, but I'm snottier and I speak LOLcatz.
LOLcatz If I was younger
LOLcatz If I was younger I'd have a retort to that
I think you meant this pic
I think you meant this pic poes.
ETA: I've meant to compliment you on your new "winter woollies" teehee.
cats, lots of cats
cats, lots of cats :?
Yesh catz wid bad
Yesh catz wid bad Engrish.
Actually, in Perl6 you can write Lolcode and it'll parse! Yesh there was a whole talk about this at the Perl conference : )
You might think that Safari
You might think that Safari and Opera have little market share now, but given that the iPhone only ships with Safari and more and more mobile devices are being shipped with Opera Mobile and Mini, and the fact that browsing the internet on mobile devices is growing at a much faster rate than it is on desktops (and that there are already far more mobiles in the world than desktops), I reckon you'd be better off getting used to those browsers now rather than playing catch-up later on.
(Having said that, if your site looks OK in Firefox, it's more than likely going to be good in Safari and Opera anyway, but I still believe it's a good habit to get into.)
The point is really that we
The point is really that we ought to be moving rapidly away from the dark days of multi browser testing as things improve with the major browsers, when IE6 has finally died and IE7 starting to fade from memory and IE8 comes close to being acceptable then coding to standards rather than to browser foibles should be the approach.
I've come across that l33t style language, but what was far more interesting was the language a visiting network enginneer showed us called [url="http://compsoc.dur.ac.uk/whitespace/tutorial.php"]WhiteSpace[/url]
Think about it, very clever you can hide code within code.
Lawlz, I was just reading up
Lawlz, I was just reading up on whitespace... like the anti-Python : )
It's all missing the point,
It's all missing the point, I'm not saying that we don't test in all browsers , but don't believe that this is clever or correct, and partly it is a rod made for our own backs.
Your Safari example is a great case in point Safari had a great bug where it couldn't handle a certain amount of scripting, it had memory stack issues, or some other rather technical fault, but it was a fault and nothing could get around this other than reduce the amount of scripting running, so choice was reduce scripting until Safari worked and make all other browsers suffer or prevent Safari from running the script, we chose the later as in that instance the scripts were considered important, so Safari users could moan all they liked but it wasn't our fault, course we could offer the advice that they install Camino :shrug:
Actually in this case it was
Though the stack issue in Saffy could have also played a part, if it could use the JS in the first place.
but don't believe that this is clever or correct, and partly it is a rod made for our own backs.
I know. Nobody said it was ideal : ( — we don't have to test roads for every different type of car that rides on it. I just have to say, if you're building for Teh Werld, you have to decide if you will test for Teh Werld or just the big names (IE and FF generally). I decided I have to test everyone for everything and I tell n00bz to do that as well if their aim is a works-for-everyone website.
You're possibly fortunate to
You're possibly fortunate to have the luxury of time to test in all browsers, in the real real world, understaffed, with too much work to get through it simply is not viable to afford all but a cursory glance at minority browsers; I have minor layout issues in Safi 2 and they will remain so as the browsers simply does warrant the time cost involved in finding fixes at the cost of project deadlines.
I would caution against that desire to tell n00bz as you put it to test in all browsers, get them to understand they are coding to a set of Standards and only then testing in certain browsers with large market share that are known to have explicit and horrendous rendering issues.
Test test test
It's sad that bad management leads to understaffing which leads to possible usability and accessibility issues for users of non-major browsers. Cursory glance is something I cannot accept for anything I build, and if I was forced to do so I'd probably have to look for another job... then again, the number of times I've said I'd quit my current job every time I see my nice code turned to crap once it gets wrung through a template machine and a bunch of javacrap added shows I prolly would just stay...
The blind are for example definitely a minority... I could even bet not a single blind visitor has run across any of my companies many sites and probably win that bet. There are other minorities. If they are restricted by user agent then I will keep them in mind and I cannot encourage people to say, "well it looks ok in FF and IE so I'm good!"
The amount of extra work I did for the colourblind on our calendars for our home rental site probably wouldn't have been considered worth it by a larger firm or a company with tighter schedules, but I'm glad I did it because I know I did the right thing. I hate being frustrated on the web and I know others do too. I want to avoid that. What, tell them to go buy new eyes??
My company doesn't value testing, though the boss has said he wants more of it due to software glitches pissing off some customers enough to call him. We don't test anything, except I test my static HTML and CSS. And I expect this company to either shape up or get their butts kicked in the marketplace because of it. Swim or sink. Testing is almost more important than building new. Why do people buy Japanese cars instead of the American ones? They test test test. If GM tested their cars as much, they'd see why the engine dies at 24k miles (this happened to a co-worker of mine, almost-new car, and when they said, tough, your warrenty was until 23k miles, she said fine, my next car is a Toyota and hung up).
Sorry, I just can't tell people to ignore something because it's not the most popular now. If they don't have the time to test, then they don't have the time *shrug* I guess I'm assuming if they're asking, they have the time.
My husband got so fed up with his job that was so busy selling products that didn't exist yet, forcing the programmers to build quick and dirty products that didn't get a lick of testing, which broke constantly due to this, and long hours fixing even more broken crap, that when some guy on teh Internets asked him if he wanted to work for a foreign company for only a bit more money, but a company that cared about its code and tested the hell out of it, cared about their products and their customers as much as selling them, he was teh g0ne. The old job stressed him terribly.
I know I'm fortunate in my job-- I got paid to learn while I went, and generally get plenty of time to feel my way through nasty things (like Flash, ug) or whatever, and do it right. That is lucky. And if this group in the Netherlands can actually do what they say (they want to make companies understand how important valid, semantic, accessible code is) then I'll be happy to join them (gonna see this month). Then I can become even more of an evangelista! And snottier!