All posts by Lachlan Hunt

Volunteer Wanted: CSS Artist

I’m looking for someone that can help me out with the Firefox 5 Minute Challenge that I’ve been working on over at SpreadFirefox. The CSS has been started, but still needs a lot of work, and I don’t have the time, nor the ability to get it working properly anytime soon, so I’m asking one of my loyal readers would like to volunteer to complete the CSS for me. Not only would you be doing a huge favour for me, but you’d be contributing to a great cause – Spreading Firefox!

If you can help out, leave a comment with your contact details or e-mail me <lachlan.hunt (AT) spreadfirefox.com>. Most of the files you’ll need are already available for you to work with, just download them and get started. I can send you all the source files if you need to re-create any of the images for any reason (eg. to fix image sizes, especially for the header and logo), and for the few images that aren’t up yet.

Note: One thing you should take note of is that file paths in the markup and CSS do not include file extensions because this server is set up with multiviews to handle the alternate languages and file formats. If you’re not developing it on a testing server with multiviews enabled, then you’ll need to add the file extensions or the browser won’t find the files.

Since IE users are the target audience for the page, it needs to work fairly well in it. It doesn’t need to be pixel perfect, but needs to be visually appealing for any IE users. It should work perfectly in Firefox because it would be fairly embarrassing to promote Firefox with a page the doesn’t work in it. If possible, I’d like it to work in IE 5.x, but that’s not a priority – aim for IE6 at least.

Firefox 1.0

For those of you that have been asleep for the past few months, and haven’t heard, Firefox 1.0 has just been released! This really is one of the biggest events in the history of the Internet, and the Mozilla site is struggling to handle the load. At the moment, trying to get to mozilla.org is slower than a standard issue, spyware infested copy of Internet Exploder. Firefox is as fast as ever, it’s just the server is overloaded. If you haven’t got Firefox 1.0 yet, get it via the FTP server, you’ll have a better chance.

As reported by MozillaZine, this milestone release of Firefox features a brand new start page. It’s not hosted on the mozilla site, it’s on Google – a smart move based on the amount of traffic (Google have got plenty of bandwidth to share). The new start page is also quite well designed from a marketing point of view: it features the most useful service available on the internet – a Google search!

It’s clean design is certainly a big improvement over the Firefox Product page, which is aimed more at users that don’t have Firefox yet. This one emphasises what a user actually wants to do: get on with browsing the web. It also features a short description about common keyboard shortcuts for new tab, bookmarks, history and search, which shows they’ve really thought about what a new user needs to know. They’ve also added links to a few of the important mozilla pages at the bottom for Firefox Central, About Mozilla, Mozilla store and how to get involved.

However, the page has been written in typical Google tradition, with no less than 51 validation errors and no DOCTYPE specified – it really takes advantage of what Firefox is good at… parsing, interpreting and guessing how to render junk built for IE! </sarcasm>

It really is a shame since part of the message about spreading Firefox is to be able to build with standards, yet the first page thust upon new users doesn’t really help with spreading that message.

iiNet Standards Redesign

Recently, and to my surprise, iiNet have redesigned their entire site. Not only that, but it validates as XHTML 1.0 Transitional, separates presentation from structure (no tables for layout!), makes reasonably good use of alt text (it’s not perfect, but it’s quite good), fairly accessible use of JavaScript (no serious problems caused with it disabled) and even makes good use of sIFR!

iiNet have discussed this re-branding from a marketing perspective. That’s fair enough, a typical customer isn’t going to want to hear about their new found standards compliant and accessible design methods. So, in the process of congratulating them for this fine effort, I’ll take a look at exactly what they have done.

Structure and Presentation

Disabling stylesheets quickly reveals that they have actually put in a lot of effort into this redesign from the previous. The old design was a typical table layout with spacer gifs, invalid markup and a few pages just didn’t work correctly in anything but IE. With the redesign, they’ve used reasonably semantic markup – headings use <hn> elements, paragraphs use <p>, navigation menus and other lists use <li> and there’s no use of presentational class names or ids.

They have, unfortunately, used a few style attributes, but not many. Most of the presentation is specified in an external stylesheet. Ideally, they would use a semantic class name on those few elements that currently use the style attribute, but the damage caused is minimal

As I mentioned earlier, the markup does validate as XHTML 1.0 Transitional, with the exception of one image missing an alt attribute on this page discussing their re-branding. However, it is a presentational image, and only requires an empty alt="" attribute anyway. The home page nearly validate as Strict, however the only errors seem to be structural, due to <input> elements being directly inside a <form> element, the use of a name attribute in a <form> element and the use of a target attribute, which I strongly discourage. They’ve also used an invalid value: target="_new". The HTML 4.01 Specification states:

Except for the reserved names listed below, frame target names (%FrameTarget; in the DTD) must begin with an alphabetic character (a-zA-Z). User agents should ignore all other target names.

This means, that except for the special defined values, _blank, _self, _parent and _top, the value must begin with an alphabetic character. Thus, _new is invalid, even though the validator does not detect it. But, you must keep in mind that the validator is just a tool, and cannot check every conformance requirement, only those specifiable with the DTD. So, technically, they should be using _blank, but ideally, they should remove the target attribute completely, since the user should decide when they want a new window, not the author.

XHTML vs HTML Markup

Update: One thing I forgot to mention earlier, and hence why I’m adding this update, is that technically, they should not be using XHTML since they are serving it as text/html, and doing so is considered harmful. If they’re going to use XHTML, they should be using content negotiation to deliver it correctly as application/xhtml+xml to descent UAs that support it, and text/html to IE, and other legacy UAs that don’t. However, as many of you will know, this issue has been discussed recently. Some say it’s OK, other’s (like myself) think it should be avoided, and other’s insist that it should not be done. These people categorised into either the Stict or Transitional Party.

If they’re not going to serve XHTML properly, then they may as well use HTML 4.01 Strict. I recommend they change to Strict, because Transitional actually triggers Almost Standards Mode in Mozilla. It is near enough to standards compliance, but it adds a small quirk that should not be there, and only exists to support the thousands of pages that depend on IE’s bugginess, yet still use a valid DOCTYPE.

Images and Alt Text

As mentioned, there is one image that I found without alt text, but other than that, they seem to have actually done a reasonable job. Although, ideally, (in this case) the images should have exactly the same text as the images do, however, they have used text with a similar meaning, and viewing without images doesn’t loose too much.

For example, one image they have at the moment, states christmas broadband specials. Free Setup + Modem. Save up to $199.95. However, they have set alt="christmas special - free setup/modem". It misses the price, but it still passes a message that is close enough, especially compared with the vast majority of sites that use very poor, or no alt text whatsoever.

They have made use of Image replacement techniques for the navigation items, though not in the most accessible way. However, that’s limitation with CSS and image replacement techniques in general. iiNet have done image replacement by setting the background image on the <a> element for each link, so that hover effects still work in IE, and set the font to 1px, white to effectively hide it from view. However, like many image replacement techniqes, this is inneffective in the rare case that images are disabled, but CSS is enabled.

For many the headings, they have made use of sIFR, which was designed and developed to be accessible in the majority of cases. It has known limitations, but so far, is the most accessible image replacement technique available.

JavaScript

The site does make some use of JavaScript, however the site does not require it. With JavaScript disabled, the only issue I found was that the what’s webmail and what’s toolbox links don’t work. They are JavaScript links with the purpose of showing additional information about the webmail and toolbox services. Ideally, with JavaScript disabled, that information should be visible by default, but the additional information is not that much, and can be obtained by disabling stylesheets also. The links should also be added using JavaScript, so that useless links do not appear for users with JavaScript and/or CSS disabled, but again, it’s a minor issue.

So, in conclusion, I would like to congratulate iiNet for taking the initiative to move towards standards compliance, and for actually hiring a web developer that knows what they’re doing. Well Done!