Mobile version (Display Regular Site)

Skip to: Navigation | Content | Sidebar | Footer


Weblog Entry

Notes From All Over Part V

July 23, 2004

Thoughts on the Odeon debacle and developer workarounds for bad sites, Internet Explorer news, and a few random links to start the weekend.

There’s not a lot you can say for a company that turns its back on a third of its customers on point of principle. It’s understood that Odeon felt leery about allowing customers to submit personal data through an unaccountable third party; but look at the numbers. 31 percent of users to the site feel it’s so bad that they refuse to use it.

Paying developer Matthew Sommerville a yearly salary to keep it running would have been a pittance compared to the business they just shut off. Starting with the numbers in this Wired article and assuming all users stop ordering Odeon tickets online (which is a stretch), that’s roughly £600,000 per year. There are enough unknowns that this number is probably far too high, but even half that is high enough.


Wired picked up on it in the article referenced, and extrapolated a common theme: developers are sick of putting up with Other People’s Messes. Witness the flap over Allmusic.com’s redesign. They went from hideous URLs, browser-specific code, and broken functionality to…. something even worse.

Adrian Holovaty took to heart Jeff Veen’s straightforward thinking on routing around the damage and wrote a Firefox extension that fixes some of the problems with Allmusic’s redesign. Is this the start of a new trend?


Pre-publishing update: looks like Zeldman has summarized these in more detail and clarity this morning.


The Internet Explorer team has a new weblog. IE for XP’s second service pack is available as a Release Candidate. Good to see signs of life; too bad the bar was raised so high while they were sleeping.

It’s hard to make too much fuss about IE’s limitations anymore though. After kicking and screaming over the long dormancy, I’ve come to accept that no matter what the IE team does now, it’s not going to actually make much difference in my day to day life for many years yet.

IE gained its market dominance because it was bundled with Windows. The version 3 to version 6 upgrade path was relatively quick because computers were replaced frequently over the past 10 years as technology improved. Consumers and businesses aren’t upgrading like they used to. Combine this with the fact that Longhorn isn’t due out for two more years at best, and you have a very long period ahead where IE6 is the baseline, simply because too many older computers will continue running it.

I expect to continue supporting IE6 well into the next decade. This is what I had in mind when I wrote this prior article on the slowdown. Perhaps the frequent and critical security issues will cause a quicker migration than that; perhaps not.

Whatever the case, I’ve made peace with IE. High-capability browsers will receive extra layout attention and effects, IE will receive somewhat dumbed-down equivalents. Users will never know the difference, and ignoring the limitations is freeing. Armed with these two filters, CSS development with IE is a far less troublesome experience:


 html>body #selector {rule;}
 /* hides the rule from IE only */

 * html #selector {rule;}
 /* hides the rule from everything but IE */

This is the future.


And a grab bag of extra link love just before the weekend: ROI is not a silver bullet. Congratulations Jay. XML on the web has failed. HTML saves more bandwidth than XHTML.


Reader Comments

1
July 23, 01h

I wrote up a review of the new IE update on my site this morning after installing the XPSP2RC2.

http://www.jeremyflint.com/wp/archives/2004/07/23/internet-explorer-and-the-xp-sp2-rc2/

The popup blocker does work well. I guess time will tell how much the security has improved. I am sure there is some 30 year old in his parents basement looking for exploits now.

The thing that came to my mind about this update is that it only affects those using XP. Will there be an update for IE6 on Win2k?

If not, that basically means Microsoft is saying the only way to improve your security is to upgrade, either now to XP or in 2 years to Longhorn.

In the meantime, the updates made to Mozilla Firefox aren’t OS dependent. So when the vulnerability was uncovered a few weeks ago, the patch that went out applied to everyone, regardless of whether you were running XP, 2000, etc.

2
Dave P says:
July 23, 02h

Good link the the IEblog, Dave. What a farce. The site doesn’t even validate. I guess that pretty much sums up MS’s take on standards.

The thought of dealing with IE into the next decade is not very appealing. I’m curious though, when (if at all) do you think we as a development community call Microsoft’s bluff and start serving unstyled content to IE users?

What’s the threshold that mozilla/opera/safari et al. have to reach in market share to make this feasible?

3
hao2lian says:
July 23, 04h

Apart from a handful (30-40ish) of odd “o:p” tag errors, the IEBlog has more forgivable validation than most websites. It’s mostly a collection of little-known validation errors, such as IDs can’t begin with underscores and block elements can’t sprout under inline elements.

What should be less forgivable is the heavy use of inline CSS coupled with “span”, but it’s nice to note they used proper h# tags and lists and that they made more or less readable code, so it’s a cheery note that, at least, they didn’t use Word or Frontpage to make their webpage (an old version of Frontpage, at least).

4
Adam Rice says:
July 23, 04h

Perhaps a renegade cell of Firefox developers will take advantage of IE’s notorious security holes and come up with an IE virus that stuffs the Mozilla rendering engine into IE. This would have the neat effect of upgrading those users least likely to maintain their systems.

Then the virus should close the gate after itself, of course.

5
Dave p says:
July 23, 05h

hao2lian: Are you looking at the same site I am? I caught a missing > off a </ul>, two closing tags without opening tags and more than 10 “font” tags right in their source.

Standards are standards… I don’t have much faith in the future of IE if they can’t code to todays standards already.

6
Lach says:
July 23, 05h

Maybe it’s time to start thinking about loading in external hack stylesheets using some server side logic, instead of specific browser flaws?

This:

1. Would make our current stylesheets cleaner and easier to understand.
2. Would mean we’re ready fdor a possible future case where a browser has no parsing flaws, but still has rendering errors.

Of course you’d want your server-side logic to be up to the task, but that’s not much more than a half hour effort ,including caching logic, only needed to be done once in this day and age.

Eg, your style sheet would start off

@import(hacks.php) (or asp or py or whatever), and the hacks page would send a text/css mimetype and route through the appropriate stylesheet for this UA (including a blank one if nothing extra is needed).

I guess I feel a little uneasy relying upon browser flaws to fix browser flaws.

7
July 23, 10h

re: “Good to see signs of life; too bad the bar was raised so high while they were sleeping.”

Actually, I think it is great that the bar was raised so high while they were sleeping. As much as I want to believe the push from various organizations to upgrade to a better browser is going to have an impact, I really wonder if it will make a difference?

My heart is telling my one thing, but my access logs are telling me another… most of our client’s sites seem to be running at well over 90% IE usage for the month of July. I’d suspect that most blogs within the web development/blogging community are seeing close to 90% not IE.

Is anyone seeing anything different?

8
July 23, 11h

What happens when IE starts supporting the child selector, so the ‘filters’ will stop working, but the (many, complex) issues the filter helped us fix are still unresolved in such a new version of IE?

I use this filter a lot. Will our sites suddenly break when IE supports the child selector?

9
Tom says:
July 23, 11h

Bah! Screw child selector hacks. IE conditionals all the way!

http://msdn.microsoft.com/workshop/author/dhtml/overview/ccomment_ovw.asp

10
Dave S. says:
July 23, 11h

First, that’s an If, not a When. It’s not a for-sure thing, it’s only one possible scenario out of many. There’s no sense in second-guessing it now, when nothing is known about future support.

If a potential Release Candidate exhibits this behaviour there will be a very vocal outrage from developers. Not to say that it will change the situation, but there will be pressure on the IE team to make sure other issues are resolved first.

There’s no guarantees when using CSS workarounds and hacks; that much is granted. But there’s no point in thinking too far into the future when we’re stuck with the limitations of today’s browsers. They solve real-world problems that can’t otherwise be solved. End of story.

11
ant says:
July 23, 12h

“Consumers and businesses aren’t upgrading like they used to.”

I wonder who is then?

http://www.newsday.com/business/ny-bzearn233903380jul23,0,1404853.story?coll=ny-business-headlines

12
July 24, 06h

Just so no one here has to spend hours debugging, remember that:

html>body {hides from IE}

whereas

html > body {hides from IE *except* IE5.0 on a PC;}

13
July 24, 07h

Dave P Wrote: ” The thought of dealing with IE into the next decade is not very appealing. I’m curious though, when (if at all) do you think we as a development community call Microsoft’s bluff and start serving unstyled content to IE users?”

Now! I will be on my site, but for now I’m only serving application/xhtml+xml for the homepage, so tough luck to any IE users wanting to see it, but that’s causing another problem now, since it seems that google won’t recognize it, and just says “File Format: Unrecognized” when it comes up in a search.

Dave P Wrote: “I don’t have much faith in the future of IE if they can’t code to todays standards already.”

Yes, I agree, but what motivation have the IE team got to write standards compliant pages, when they know that their browser chokes on it, especially given the fact that IE’s standards mode is essentially identical to quirks mode, with just a few extra bugs thrown in for fun? Do you really think that given Microsoft’s history, that they actaully care about serving valid content to others?

Lach Wrote: “@import(hacks.php) (or asp or py or whatever)”

That’s a resonable idea, except you have to be aware the browser sniffing techniques are not always reliable. It’s quite easy to alter the useragent string. For example, see Dav’e hilarious “Nutscrape” post:
http://www.mezzoblue.com/archives/2003/05/13/nutscrape/
(BTW, Dave, the image of the referral log is missing for that)

Although, it would be better to just import a file *without* the file extension.

eg. @import(/style/file); or
<?xml-stylesheet href=”/style/file” type=”text/css”?>, etc…

and then set up the server with Multiviews, or similar, so that the implementation can change from plain CSS, to dynamically generated with Perl, PHP or whatever. Theoretically, if there were ever a new stylesheet format as widely implemented as CSS is, it could content negotiated to deliver text/css or application/new-style-format, and the URIs wouldn’t need to change. That’s the way I’m doing it on my website at the moment, and may be converting to dynamically generated to send alternate style to IE, with the same URI.

14
Lach says:
July 24, 09h

Lachlan, I agree absolutely. However, I really don’t care about people who alter their UA string. If they’ve changed it so it doesn’t include at all the string ‘Opera’, ‘Gecko’ etc, then they know what they’re doing. Additionally, I doubt this would count for greater than a hundredth of a percentage point of net surfers, and I’d also think most people competent enough to alter UA strings would be using browsers other than IE (which is where 99% of UA specific hacks will need to go)…

15
Sophie says:
July 25, 11h

Derek Featherstone says:
“… most of our client’s sites seem to be running at well over 90% IE usage for the month of July. I’d suspect that most blogs within the web development/blogging community are seeing close to 90% not IE.

Is anyone seeing anything different?”

I’ve seen 16 to 18% Mozilla-Safari-Opera figures for more than a year. This morning among the last 20 visits, 7 are from one of these browsers, and it has been a usual figure to get 5 out of 20 recently.

16
July 25, 12h

question: You mentioned

* html #selector {rule;}
/* hides the rule from everything but IE */

Does this work as a separate filter? Or only when specifying a child selector (html>body) before? If it works separatly, a IE6 filter is easy to implement, before the IE5.5 and IE5 middpass filters from Tantek.

Do you have a link describing the mentioned * html rule? It is the first time I see it.

17
Dave S. says:
July 26, 01h

> Does this work as a separate filter? Or only when specifying a
> child selector (html>body) before? If it works separatly, a IE6
> filter is easy to implement, before the IE5.5 and IE5 middpass
> filters from Tantek.

It works separately. Each filter is stand-alone.

> Do you have a link describing the mentioned * html rule? It
> is the first time I see it.

I can’t find a specific page at the moment, but if you browse around http://positioniseverything.net/ you’re *sure* to find something about it. (and probably a whole lot more you never knew either)

18
July 26, 06h

The advent of content altering “fixing” extentions is in my view a serious obstruction of ones freedom of expression. Just because someone delivers bad work doesn’t give someone else the right to alter that content because they feel that they know what better for mankind. Intentions so far are pretty good, so far…

19
Martijn says:
July 27, 12h

Thanks Dave,

It’s the ‘Tan hack’ : http://positioniseverything.net/articles/box-model.html

In hindsight I knew about it, but never used it before.