RSS Feed

The Joel on Software Discussion Group (CLOSED)

You mentioned the market share of Firefox. A data point you might be interested in is Tim Bray's Ongoing weblog (Tim works for Sun if you haven't come across him). Every week Tim calculates some statistics from his logfiles, one graph is browser market share http://www.tbray.org/ongoing/When/200x/2004/12/12/BMS

Looking at the stats for a tech news website doesn't say much to me in concerns to "market share".  Techies go to this site, so he's going to have more Firefox users - just seems like a biased look at this topic to me.

I checked one of our highly trafficked websites that attracts customers of every walk of life, a good mix of everyone, and we still get 92.54% (taken from a sample of 500,000 sessions) of users using an IE product with less than 1% decrease from last year around this time.

If your audience is a non-tech community, is it really important to make your web products support Firefox?

FYI, there are still a HUGE number of people on 800x600 resolutions.  I would be more concerned that web pages meet the needs of these people, as there are way more people with smaller screen sizes than people using obscure browsers.  (33% at 800x600, 7% on Netscape/Mozilla browsers)

"If your audience is a non-tech community, is it really important to make your web products support Firefox?"

Yes.

If you're trying to sell something, you're giving up 7.46% of potential sales because you couldn't be bothered to support more than one browser. If you're not trying to sell something, you're just being needlessly rude to 7.46% of your visitors.

If you're still thinking in terms of designing for one or two specific browsers instead of designing for the web, you're making life harder on some fraction of your visitors for no gain. You also run the risk of having to redesign if and when IE's market share drops to the point where you feel like you have to care about other browsers.

It's obvious that IE and Firefox should be supported by any major consumer web stie. But what about Opera, etc.? Where do you draw the line? What percent of market share do you consider too small to commit development dollars to?

"But what about Opera, etc.?"

After you've already built a site for 2 browsers, it gets much easier to design for 3.  My policy is to design for Firefox and IE and then periodically do whatever fixes are necessary for Opera.  However, for my application, some things just plain don't work in Opera (and never will) -- so opera users get a slightly worse experience.  But the for the most part I at least try and make it possible for them to use the site.

That's exactly what I'm saying.  If we are playing percentages, there are way more people I should cater to first, rather than Firefox which is only 7%.

If it's a general business website, making it compatible is very easy.  If it's a complex web system, I would just assume require IE.

For example, I create an app that 1000 people use and subscribe to at $10 per month.  70 people want to use Firefox.

Now, are these people NOT going to use the system because it doesn't support Firefox?  Or will they just bite the bullet?  What if the people who say No Way is only 10 of the 70 who want Firefox support.

Can I justify spending valuable development time catering to clients generating only $100 per month?  Well, if I can finish the compatibility upgrade in under 20 hours, then ... maybe.

I target IE and Firefox with my application.  I do pretty much all development testing in Firefox and QA mostly uses IE.  The site actually runs a bit better in IE (faster, smoother).

The main reason for keeping Firefox around is to hedge our bets -- even now, IE7 is going to be for Vista only.  That means that we'll have many years of IE6 without new development.  At some point, we might eventually have to recommend Firefox to our clients -- it will less painful to develop for it now than later.

Furthermore, sometimes people just have problems with IE (spyware, etc) that can conflict with our site.  Without Firefox -- we cannot give people any alternative options.

Just curious, but does anyone use 800x600 not because of a very small screen, but because it is the only consistent way we've found to provide larger type and larger everything because we can't see very well?

I see plenty of people who run 17" monitors at 800x600, presumably to get larger type.

I run my monitor at 1600x1200, but my web browser gets at most half of that width. I have other things to do with the remaining screen real estate.

"Now, are these people NOT going to use the system because it doesn't support Firefox?  Or will they just bite the bullet?  What if the people who say No Way is only 10 of the 70 who want Firefox support."

I was one of a few thousand people who emailed my former bank describing "how disappointed we were that they use ActiveX controls and other things based on a unsecure browser that the Departments of Justice and Homeland Security don't even use".

Their site was FF compliant within a week.

Ahh but KC, a few thousand, with serious profit implications, does not equal 10, at a mere 100 a month.

And apparently I'm not the only one to run at 800x600 for that reason...

" If we are playing percentages, there are way more people I should cater to first, rather than Firefox which is only 7%."

That's a dangerous game, though. For instance one of the most profitable small ISV spheres is developing for the Mac/OSX. As a group they're less likely to pirate, and more prolific spenders on software, and of course the domain has less competition.

Knowing the marketshare alone means nothing if 90% of the customers who would likely be interested in your product run Firefox, while you're targetting the irrelevant 93% of the browser market that just wants to visit livejournal.

"That's a dangerous game, though. For instance one of the most profitable small ISV spheres is developing for the Mac/OSX. As a group they're less likely to pirate, and more prolific spenders on software, and of course the domain has less competition."

That makes sense for the Mac world but do you really think that Firefox users would fall into the "less likely to pirate" category?  The only non-technical people I know who use Firefox do so because they got spyware while browsing warez/mp3/movie download sites using IE.  Not exactly the kind of people who would make great customers for a software company.

"That makes sense for the Mac world but do you really think that Firefox users would fall into the "less likely to pirate" category?"

Fair enough, but on the flip side if you're promoting a software development tools, for instance, the proportion of Firefox users are overwhelming (including amongst Microsoft-technology developers). There are a lot of niche type markets where Firefox use is overwhelming.

> IE7 is going to be for Vista only. 
> That means that we'll have many years
> of IE6 without new development

Not quite right, right now I'm using IE 7 Beta 1 (obtained via Microsoft Partners website) on a WinXP and I'm pretty sure many people will install it on their desktops as soon as it becomes available to the public.

Btw, IE 7 looks pretty nice and the coolest thing is tabbed browsing, which is implemented better than Firefox.

It also features an embedded RSS reader, auto discovery of RSS feeds on web pages, a Search box that can do Google, Yahoo or MSN searches, and a much cleaner toolbar only with Back and Forward buttons. The "Go" button is replaced with a "Refresh" button which also acts as "Stop" when the page is being loaded.

"The "Go" button is replaced with a "Refresh" button which also acts as "Stop" when the page is being loaded."

I've used browsers like that in the past, and hated it. Many times I've had a site finish loading between when I decide to click "stop" and when I actually click it. The result is that I end up reloading.

Bad idea.

At my site we get between 350,000-600,000 page views per month.  It's a fan site for an old computer game.  We are at over 50% of our users on firefox, IE 6 at 36%, and some other versions of IE and opera with smaller percentages.

We are hardly a cutting-edge community (the game was released in 1997!).

We are missing an important point here.

What techies/geeks use today, many more people, if not everyone, might use tomorrow.

So, don't say, "50% of my visitors use Firefox, but mine
is only a techie site".

Much rather, put it as "50% of my visitors use Firefox, AND mine is a techie site!"

Paul Graham wrote a couple of good articles on the
"Mechanics of Popularity" which I recommend to everyone:

http://www.paulgraham.com/mac.html
http://www.paulgraham.com/popular.html

The first one is about hardware, and the latter one is
about programming languages, so some aspects surely don't
translate well to Web browsers, but the main point most
probably does.

Once upon a time, only geeks used computers.
Once upon a time, only geeks used C.
Once upon a time, only geeks used the Internet.
Once upon a time, only geeks used Perl.
Once upon a time, only geeks used Firefox.

Get the pattern?

(And yes, I know there are 3 gazillion counter-examples,
but that's not my point, so please don't mention them. ^_^)

From a practical standpoint, If find it's easiest to write markup and CSS that works in the most standards-compliant browser (e.g., Firefox or Safari), then add the necessary workaround for IE.  This helps make the site future-compatible.

If you write it for IE, the layout could rely on IE's many rendering bugs.  When IE7 fixes those bugs, not only will your site break in Firefox, but now in IE.

It's counter-intuitive to develop for your largest browser second, but I find it's the most efficient way.

"Once upon a time, only geeks used C.
Once upon a time, only geeks used Perl."

I believe this is still the case.

"From a practical standpoint, If find it's easiest to write markup and CSS that works in the most standards-compliant browser (e.g., Firefox or Safari), then add the necessary workaround for IE."

I also find that to be the case. Usually the only adjustments that I then have to make for other minority browsers have to do with default margins and padding that I didn't overwrite.

Firefox also has a better debugger (although I do use the Microsoft Script Debugger on occasion) and the DOM inspector.  Very useful for development.

"Firefox also has a better debugger (although I do use the Microsoft Script Debugger on occasion) and the DOM inspector.  Very useful for development."

You can add a very useful DOM inspector to IE with the Developer Toolbar [1].  I personally find this much more useable than the one that comes with Firefox.  It's faster, has more features, and has a "Select Element By Click" feature that actually works.  Throw in Fiddler [2] to make IE an even better development tool.

[1] http://www.microsoft.com/downloads/details.aspx?FamilyID=E59C3964-672D-4511-BB3E-2D5E1DB91038&displaylang=en
[2] http://www.fiddlertool.com

Safari also has a decent if minimal DOM inspector. Like other Safari development goodies, it's hidden:

http://developer.apple.com/internet/safari/faq.html#anchor14

I've used an IE DOM inspector called IEDocMon. It works, but I'm not thrilled with it. I'll have to give the IE developer toolbar a try.

People are fully justified ignoring Firefox (or Opera) if they choose to. Sure, they may be ignoring 7% of the market, but so what?

If you make a website only in English, you are ignoring the enormous number of web users who don't speak fluent English.

If you make a Mac only version of software, you are ignoring all the PC users.

These are all design decisions with their own tradeoffs that are have to be made.