Forum Moderators: open
a) Web surfers
b) Search engine operators (most notably Yahoo)
I know posting this question will probably mark me out as a complete newbie (and yes, I am such), but well, everyone has to start some place.. :)
Welcome to WmW.
I agree IE is the most widely used browser. But optimising only for it for this single reason is a mistake. If I was a bar tender, I would not tell the doorman to admit only people with brown eyes and brown hairs.
If your site does not display correctly in other browsers, part of the market share will be lost. Also, the browsers landscape can change. For instance, AOL will soon drop IE for NN and they have more than 34 millions users worldwide.
First making sure your site has valid w3C compliant coding. This is done using the validator ( validator.w3.org ). Then you will have to make changes to it to fix various browser bugs. Publishing a web site without a multiple browser quality control will limit his audience.
Yahoo! websurfers are using a wide range of browsers and platforms. You never know, sometimes they use many browsers while reviewing the same site.
I have recently downloaded
4) BrowserG
5) Mozilla
I am also looking for a good free text browser for a Win98 OS.
My favorie testing browser is Opera, it allows you to validate html on line among other things.
This is our stats break down
90.8% IE
8.8% NN
0.2 WebTV
0.1 OP
0.0 KONQ
A lot of the time when Opera renders something diferently than NN or IE, Opera is actually doing it the right way. So by checking in Opera you are publishing pages that will hold up well when the other browsers finally get it right in future versions.
A lot of the time when Opera renders something diferently than NN or IE, Opera is actually doing it the right way.
Tedster is absolutely correct... Opera's STRICT interpretation of HTML is why Opera IS my #1 test browser.
Opera's extensive feature set is why it is also my #1 surfing browser. Opera is a total control browser... and YOU, the end user, are in control!
I happened upon an example today that demostrates WHY validation is more than a good idea: it should be required! Opera displayed the page as was written, errors and all. IE6 and Mozilla 1.0 rc were forgiving and displayed the page as the author intended, even though there were numerous mistakes. This was NOT a personal homepage mind you, but one that offered programming solutions.
The point is this: it takes no more time to write VALID code than it does to write INVALID. Web Standards are written for a purpose, not a for whimsy. Opera, bless her strict little soul, calls 'em as she see's em.
Writing valid code means Opera will display page code as intended... and as validated.
Testing with Opera as you build, leads towards validation. I know I WANT to find my mistakes BEFORE I upload a page. Opera helps me do that.
XHTML + CSS + WEB STANDARDS + Opera = :)
- papabaer
...it's HTML H*ll out there! Got your <!doctype>? No? Oh, well then, you are completely confident that your code is without a doubt, absolutely, 100% grade "A" VALID... aren't you? ;)
as Mr. Bill used to say: "Oh noooooo!" :o
How do I choose which browser to optimize my websites for?
Don't. Make it "Viewable With Any Browser" [anybrowser.org...]
"Anyone who slaps a 'this page is best viewed with Browser X' label on a Web page appears to be yearning for the bad old days, before the Web, when you had very little chance of reading a document written on another computer, another word processor, or another network."
-Tim Berners-Lee in Technology Review, July 1996
The reality is that there is probably not one web site out there that actually displays correctly and looks descent on all browsers / platforms.
I usually code for HTML 4.01 and get it validated like Macguru, suggests. Also i use CSS1, and also get it validated. As well as starting with the validation, i have various browsers of different versions on different OS platforms to view from.
The URL, Luma, suggests is quite handy to see how self-percieved simplistically designed pages, are way over the heads of the old Mosaic (early graphical) / lynx (text) based browsers > can't please everyone !
I am more concerned with WML, than anything else.
However, NN4 doesn't support this, and even if I do it with CSS (td { background-image: URL(PICNAME); }), it still refuses to show the gfx..
I know this is an isolated example, but CSS being the relatively new standard as it is, and the fact that some very important people (IE: YAHOO) uses NN4 (a rather old browser), how can I be sure that CSS will show my page in its full glory?
The thing about CSS is that NN4 doesn't really support it in full.. for example, if I want to make an effect where there is a piece of paper on the web, I have to make the gfx for the "paper" scrollable...
NN4 DOES support background-images for CSS, but like most everything the support has limitations.
I did a page a few months back that used a "sheet of paper" as a background-image. It works fine in NN4, though I will soon move the CSS to an external style sheet and reference the code using the @import rule which NN4 does NOT understand.
NN4 surfers will no longer see the styled content, but this will ALSO present useable content not only to NN4 users, but ALL older user-agents with little to none CSS support.
I will send you the url of the page via stickymail.
- papabaer
"NN4 DOES support background-images for CSS, but like most everything the support has limitations. "
(I have no idea how to do the boxed in quote thing, so this will have to do :p)
NN4 supports background images for the body tag.. not for td tags.. the current code for what I want is:
<table><tr><td background="XX"></td><td>INFO</td><td background="YY"></td></tr></table>
In this simplified version, when the center cell lengthens vertically, the side cells will also enlarge, and the background will tile. This way, if I wanted a "treasure map" effect where the sides of the table look torn and tattered, I can make 2 gif's that show the sides of the map (XX and YY), then when the center cell lengtens, the sides tile, so the "treasure map" can be of any length.
To put all that simply (I can't even understand myself :p), I just need to frame my data with two pictures that form the sides. Since the length of the data varies from page to page, I need the framing pictures to length/shorten as required.
I have no idea how to do the boxed in quote thingStick your text inbetween.
junbin -- also, HELP > HTML Style Codes...
Would testing and making sure everything looked well with Opera 6.01, NN 6.2 and IE 6.0 -- then validating with the W3C for XHTML/CSS compliance insure atleast acceptable results on IE 5.x and up and other browsers with such high numerical value? I'm assuming you cannot have more than one IE browser installed at the same time...
I don't have the luxury of more than one CPU right now, I can only test with one Microsoft browser, IE 5.5 or IE 6.0 _-> ?
Lastly, how important do you believe testing with Mozilla is? That is, if Opera is the Man when it comes to standards?
Best-
Those are very good questions. I've witnessed enough "exceptions" to be wary.
Opera is my primary testing browser, I've had good results with consistancy when Opera displays as intended, IE6 generally follows suit (exceptions include a number of IE's CSS quirks, including truncated content ala ALA's documentation ;)).
Mozilla also will treat certain element spacing differntly. For example, I had a 200px by 150px <div> postioned with the right-side of the container 5px to the right of an iframe boundry. All that was contained in the <div> were a few links.
There was no padding or margins assigned to the <div> so the text links were only contained by the <div>'s set dimensions. You would assume that there would be no element overlap. Yet though the <iframe> displayed without scrollbars in Opera and IE6 & 5.5 (quirk mode emulation), Mozilla inserted scrollbars which were only eliminated AFTER I repositioned the <div> an additional 5px to the left.
Opera will also add spacing after a paragraph that will NOT be contained by a div (set dimensions) and may cause scrollbars in an iframe when positioned near the bottom of the element.
Spacing of standard HTML elements can vary enough to warrant a thorough "look" across browsers. Headings and Lists are to other elements to be wary of... if you want to avoid unexpected results!
The good news is, all of the exceptions and "variations" I have come across (so far!) are controllable via CSS. Once you become familiar with the variances, it is easy to approach near uniformity across the major (modern) browsers.
Regarding IE6 and testing for IE5.5 Quirk Mode: I had assumed that removing the <!doctype> from the beginning of my HTML document and then displaying the page in IE6 would trigger 5.5 quirk mode emulation... it does, BUT! there seems to be at least some difference between simulated IE5.5 (via quirk mode) and IE5.5 proper. Not long ago DrDoc had a display problem with 5.5 that I (and several other WMW members) could not replicate via IE6/5.5 quirk mode. I don't believe there are too many differences, but I really cannot say.
While there are still a number of significant exceptions, the amount and variance is nowhere near the level before modern Web standards browsers. Proceed with caution! ;)
Lastly, how important do you believe testing with Mozilla is? That is, if Opera is the Man when it comes to standards?
IMHO, Mozilla is the Man when it comes to standards. In my experience it is far more reliable than Opera when it comes to rendering and manipulating the DOM.
It's biggest drawback (besides still being "pre-release" software) is its system resource footprint.
All IMHO, of course. YMMV.
He was embarrased when he was shown his own site on a Netscape browser; a blank screen. No wonder it had such a small hit rate.
Every site should be thoroughly tested in the following common browsers:
1) IE 5 and/or 6 (because of popularity)
2) Netscape 4 and 6 (4 because of Yahoo and to help validate HTML)
3) Opera 5 or 6 (because of it's strict HTML standards and also it's failure to be polite to JavaScript code - it will be slow and it will find every minor error that IE and NS will treat as everyday)
You should also try the following:
1) Rename your .css files to check your site will work nicely without them. Not all browsers support css.
2) Use Opera to disable combinations of Frames and JavaScript (if your site uses either of them) to show all potential problems that visitors may face.
3) Change the default background to grey (in a browser you only use for testing purposes). Some Apple Mac browsers and RiscOS browsers use grey as a default background. If your page background is intended to be white and you do not use <body background=white>, you will look very poor on the Mac platform.
4) If possible (and a business site) try showing pages in larger fonts for a simulation of a visually impaired persons view of your site. And if you have access to a speech tool, try this also. If you are selling products, this can be important to get extra sales and repeat orders because they will remember your site is accessible to them. That will be unusual rather than the norm.
Tips:
If you are struggling with JavaScript errors in Netscape, type "Javascript:" into the navigation bar (where you normally put a web address) and see an errors list. This works in at least NS version 4 and later.
And don't forget - if you have a large monitor, shrink the size of the screen down to 800x600 and 640,480 to see how it looks. I find using % values for widths much better than fixed values but some sites seem to differ in their approach to this.
All of this can take a long time, but performing the above tasks little by little will help you amend your site to be accessible to everyone.
Don't get carried away with clever JavaScript, just use basic functions otherwise people will be greeted with a nice error message. Don't get carried away with complicated css files unless your site will look fine without it.
As stated above, Yahoo use version 4 of Netscape and this can be important. Web surfers use a wide variety of browsers and making it compatible with all is impossible, but making it compatible with most is not as difficult as it sounds.
However the most important step before you test in any of the browsers above is to validate your code. If it's not valid you are asking for problems. One well known problem is with NS v4, miss out the </table> and the whole table is not displayed. NS v4 is also well known for showing blank pages.
it may appear not as great when using old browser but if you like to see the great things the internet could give, why not update? NN4 and old browser or lynx user have their own reason in using them, they all need information only.
less and clean codes,few hacks (well, we could not live without it). faster loading. and everything nice.
go forward.
exactly...I refuse to spend my entire career checking every site I build in every new version of every browser...so I build to the w3c standards first and only consider browsers in terms of known major bugs...then when the site is just about finished I'll tamper with the styles to sharpen it up a bit in some of the more popular browsers
but the web was intended to be device independent...I don't want to lose that
