Drawing a line in the sand: No more IE 5.x support in Basecamp and our future products Jason 26 Mar 2005

42 comments Latest by Arne Gleason

The time has come to move on. We’re phasing out IE 5.x support in Basecamp and not supporting IE 5.x in any future products we develop.

In some ways the decision was a hard one, but it’s also one of those decisions that you’ll look back on and wonder why you didn’t make it sooner. This is symptomatic of many tough decisions.

But in the end we’re betting on the future, not the past. We’re on the Ajax highway and not looking back. Backpack will be the one of the best examples of the benefits of an Ajax customer experience on the web when it launches in 30-45 days. Ta-da List said no to IE 5.x from the start, and so far over 15,000 Ta-da listers have jumped on board.

Sure, we could have hacked in support for IE 5.x, or built alternate versions for older browsers, but legacy support is not the way to build a forward looking product. IE 6, Firefox, and Safari are now the tickets to ride.

So today is the day we announce we’re moving on.

42 comments so far (Jump to latest)

Tristan Dunn 26 Mar 05

Good choice, in my opinion. Can’t wait to see Backpack.

Josh Delsman 26 Mar 05

Thank god. I can’t believe anyone is still supporting IE 5.x. I haven’t from the start. I wish we didn’t have to support IE 6.x :)

Justin 26 Mar 05

Bravo!

Mark 26 Mar 05

7% of your audience, crikey. Here’s hoping you just turned lots of people into Firefox users! :)

Sean Tierney 26 Mar 05

That’s a bold move giving the finger to 93% of the browser market but those are probably the 93% that wouldn’t appreciate your stuff anyways… BTW, I just setup Basecamp for a legal tech consulting project I’m working right now and it has gotten great feedback so far. The attorneys I deal with are generally still running Win 98 but they can all appreciate a great interface when they see one because it just works right. Great job.

Sean

Andrew 26 Mar 05

“That’s a bold move giving the finger to 93% of the browser market but those are probably the 93% that wouldn’t appreciate your stuff anyways…”

The whole point is that it’s NOT 93% of the market. That’s completely wrong. It’s 7% of *their* product’s userbase, and I assume it accounts for more development and support costs than 7% of the userbase adds to profits. Actually, I’d bet that some significant portion of that 7% will simply switch to another browser, so there’s almost certainly not going to be an instant 7% drop in users. Sean, even those attorneys using Win98 could still use Firefox perfectly well.

B. Adam 26 Mar 05

Sean: IE5.x doesn’t have 93% of the browser market, they’re saying it has 7%. And 7% of their market — which is an important distinction. This kind of move should certainly only be made on a case-by-case basis, after closely reviewing your stats as 37s has done.

So, Jason, am I right in assuming when you say “no support” you mean not even unstyled access? Not even a slow, degraded yet still accessible version? Interesting.

A. Casalena 26 Mar 05

Is this a cross platform statement? (IE 5.2 Mac as well?)

I gave the kick to IE 5.x in my apps a bit ago (for Windows users, not having my applications work is the LEAST of your problems if you’ve got IE 5.5 — I’ve never seen an IE 5.5 user who wasn’t totally consumed with spyware).

It gets so tricky with Mac’s becasue of Apple’s failure to provide browser upgrades to their old OS’s. It’s OS 9 I believe that is forced onto Safari 1.0 + IE 5.2 (Someone was telling me Firefox wouldn’t work on OS 9 — but I have no idea if this is true or not).

bonaldi 26 Mar 05

No, there’s nothing that will work on OS 9 apart from IE.
No Safari, no firefox, no camino, no …

(There’s an ancient unofficial hack of Mozilla that has some nasty bugs and a terrible version of iCab, but that’s really it. And at my work — newspaper company — the machines are locked down with only IE.)

So that’ll be us OS 9ers completely out the 37 signals loop. Boo. We only just got gmail working, too.

ChrisR 26 Mar 05

Very little is supporting OS 9 any more. Most new software is OS X only. It’s time to move up. We’re already 4 versions in on OS X. Should software also work on OS 5, OS 6, OS 7, and OS 8 also? The future has been here for a couple of years now. Get on board or miss the train.

Thomas Baekdal 26 Mar 05

“It’s 7% of *their* product’s user base”

I think you are all missing something here. Products like Basecamp, might very well only be sold to people with the latest technology - i.e. web developers. But, Basecamp needs to be accessible and usable for their (web-developers) clients too.

Take Sean Tierney’s case: His clients are not using the latest on the planet - so, this move could force him to stop using Basecamp. It is simply because his clients would no longer be able to use it.

What good are products like Basecamp if the end-users (your clients-clients) cannot use it?

I do agree that technology like XMLHttpRequest (AJAX) is exciting, but I think it is a mistake to base the application on it. Do extend it with AJAX though, but make sure your applications work without it.

I do not think people buy Basecamp, because it uses AJAX. I think they buy it because it is a simple, and easy to use, project management tool.

Coincidently, on my desk is book. It is yellow with a huge exclamation mark. In the book there is a sentence. “Don’t shut out visitors with old technology: Offer alternative versions and technology upgrade information” (Page 106 - Guideline 19).

The book is called “Defensive Design for the Web” - by 37signals…

bonaldi 26 Mar 05

Right, ChrisR, but do you think Basecamp is going to be the driving force to get my company on the train? Hardly. Riding that train would be very expensive for us, because our custom publishing system - which is still supported - only runs on OS 9 and buying a new version is unjustifiably expensive when everything else we have is running fine: papers come out every morning.

So we lose Basecamp, basically.

JF 26 Mar 05

Thomas, at some point you have to decide who to support, what to support, and how to support it. At this time we’ve decided that we’ll no longer support IE 5.x (a browser that is over 4 years old and dead in the water). Should supporting IE 5.x cripple the experience for everyone else running the latest browsers? We say no.

Sure, there are ways to degrade gracefully so anything can view/do anything, but it’s time consuming, requires more software to do less, involves messy backwards compatibility code, and it has a serious negative effect on progress. In this case we’re choosing progress.

We are betting that the enhanced customer experience for the 93%+ majority will be worth the negative impact on the shrinking minority. Plus, many of the people in the minority will have the opportunity to join the majority by upgrading their browser. An upgrade to the latest technology won’t only benefit them while using Basecamp, but it will benefit them elsewhere on the web too.

If a few percent are left who can’t or won’t, then yes, I suppose we’ll be losing them as a customer. I don’t like losing customers, but we’re dedicated to building the best possible product we can. In our world that means crafting the best possible customer experience for the largest majority of our customers.

Sure, if it was 51/49 or 60/40 or 70/30 or 80/20 we’d probably hold off. But there has to be a cut-off somewhere. We think a 93% majority, and the modern browser switch momentum, gives us a mandate to innovate on their behalf.

Hoshi Sato 26 Mar 05

The book is called “Defensive Design for the Web” - by 37signals…

Wow, Thomas really laid the smackdown on 37Signals. :p

JF 26 Mar 05

One more thing… Of the 7%, IE 5.x for Mac OS 9 is a very small slice. Mostly it’s IE 5.x for PC folks.

JF 26 Mar 05

The book is called “Defensive Design for the Web” - by 37signals…

Yup, we’ll roll with that punch. But really, the book is full of guidelines, not rules. If that rule was hard and fast then we’d have to support IE 3 and 4 (and who knows what else). As with anything, it’s a judgement call. You weigh the pros and the cons, you make a decision, and you move on.

Fred, the real Fred 26 Mar 05

Arrhhhaaaa, the Pirate Code is more like guidelines then them bein’ rules…..

Fred, the real Fred 26 Mar 05

Joking aside, we’ve said no to IE5.x over a year ago with only a mild uproar.

Sunny 26 Mar 05

If you are still using IE5.x, you don’t much care for user experience. Good move.

Jesterace 26 Mar 05

Goodbye IE 5.5 heh, I haven’t used that browser in I don’t know how many years. Yes lets look forward!

Rob L. 26 Mar 05

Hey, if anyone knows Bonaldi (he/she didn’t leave what looks like a real email address), they should really let him/her know that there IS a Gecko-based version of Netscape in the 7.x range that runs on Mac OS 9. Not sure which Mozilla build it’s comparable to but it’ll probably be OK for basecamp.

http://browsers.evolt.org/?navigator/mac/7.02

Testing... 26 Mar 05

Heh, I’ve spent approximately 2 hours testing the presentation of my latest client’s site templates in IE 5.0 and 5.5 (Windows). As usual, debating whether or not to fix certain things for IE 5 that would unnecessarily add to the code (<p> inside of list items to force proper blocking). This announcement has ended the debate, I hereby also cancel any and all support for IE5.X.

Incidentally, if you’re on OS 9 and using IE 5 it’s time to move on. If your company uses a specific application written for OS 9 why not run it in the CLASSIC environment of OS X?

That way you’re not held back in ALL other aspects of your computing experience, just because of one application or workflow.

Can you imaging booting your brand new dual G5 with 23” Cinema Display into OS 9? Sad…

brian breslin 26 Mar 05

Good call, I’ve stopped caring for IE 5, it doesn’t handle CSS[well], and really impairs everything more than its worth.

Can’t wait to see the basecamp updates. I’m just getting started using basecamp, but the potential is really there. I’m loving it so far, even though I feel i’ve barely touched the surface….

Alex 27 Mar 05

How is Ajax different from ASP.NET 2.0’s client callback mechanism?

Thomas Baekdal 27 Mar 05

Alex: It is basically the same (ASP.NET 2 vs. AJAX). Both ASP.NET 2 and AJAX use the XMLHttpRequest DOM to handle the callback.

Steve, from Co, FLA and now RDU 27 Mar 05

Is there any security risk in using Ajax or ASP.NET’s callback?

Adam Michela 28 Mar 05

Welcome to the club :)

Anonymous 28 Mar 05

Can you imaging booting your brand new dual G5 with 23” Cinema Display into OS 9? Sad…

It’s not even possible. There aren’t any current offerings from Apple that allow you to boot directly into OS 9. Someone will/should correct me, but I think that ended with the 1.25Ghz G4 about a year and a half ago.

monkeyinabox 28 Mar 05

If you have a “brand new dual G5 with 23 in Cinema Display”, true you are not going to run OS 9. The point is not everyone is always urgrading as fast as you’d like to think, especially in the business world where running Windows 98 and OS 9 still happens. If you are not running newer applications, and have older hardware that runs an older OS just fine, then a forced upgrade doesn’t always happen. Windows Update pretty much forced everyone to move past IE 5.x, but we all know how many people religiously pay attention to security patches and updates.

Dan Boland 28 Mar 05

I, for one, support the decision. My feeling is that if an older browser is becoming dead weight, drop it. I don’t think it’s unreasonable.

Jack Shedd 28 Mar 05

I’d be all for it if it wasn’t for my clients. If it was just an internal tool, no biggie, but there’s no way I can shove Firefox or IE6 down my corporate client’s throats. They’ll upgrade when the company decides to upgrade, and a three person design team is going to force their hand.

Bad move.

mrod 28 Mar 05

IE 5X (win) is shot with so many security holes, the users are treading on thin ice anyway. If their not experiencing problems all ready…

pkmk 28 Mar 05

please support opera 8
it’s a decent browser.

Britt 28 Mar 05

Excellent move. With all the security issues, I can’t imagine anyone using IE5.x has a good experience on the web anymore.

Anonymous 28 Mar 05

I always thought the phrase “drawing a line in the sand” was mis-leading since re-drawing that line would be fairly easy. Is the reference to some sort of gladiator game?

Matt 28 Mar 05

Actually I surprised that you have 7% using IE5.x. My site SoftScout.com - Business Software Encyclopedia only has 3.8% using IE5.x

Matt

Bryan C 29 Mar 05

Hurrah! I spent most of a week rewriting the CSS for the site I maintain. Most of it was necessary maintenance and cruft removal, but I spent way too much time chasing down obscure bugs in IE 5.5 and IE Mac 5.2. This is the last time I plan to do that.

OS9 and IE Mac 5.2 I can see. But I’m baffled why any company would still run IE 5.5. IE6 is a free upgrade, it’s been out for years, it’s backward compatible, easily deployed… IE 5.5 is, as many have already said, a security disaster waiting to happen. What’s the holdup? Y’all waiting for IE7? :)

matthew 29 Mar 05

sort of on topic regarding your brand new love affair with AJAX/xmlhttp/javascript.

from what i understand regarding screen readers (you know, what the blind use to browse the internet); your ajax tricks won’t work because the products won’t reread a screen like that, they look for the browser refresh.

and then there’s the fact that javascript alone isn’t really accessible.

okay, so that’s fine — maybe you have two versions of a site depending on what browser is there or what features are turned on? it’s not like it’d be hard to detect javascript and redesign the link programatically. but no. if i turn javascript off, tadalists breaks. it just doesn’t work because you designed it to work one way and only one way.

so, i guess my question then — is this book, “Defensive Design for the Web” being used more to prop up a wobbly table than as a set of guidelines?

JF 29 Mar 05

so, i guess my question then — is this book, “Defensive Design for the Web” being used more to prop up a wobbly table than as a set of guidelines?

As I said before, at some point you have to decide who to support, what to support, and how to support it. At this time we’ve decided that we’ll require Javascript to use our products.

Javascript has been required for Ta-da from the start. Nothing has changed there. It’s required for the experience we want to provide. If people don’t have Javascript turned on (or available at all) then they can’t use Ta-da. We’re OK with that.

The key is to figure out what you are comfortable with as a company policy and then go with it. One size may fit all at some level, but it seldom fits really well. We prefer to craft a great experience for those who have the tools to experience it rather than a sub-par experience for everyone.

markku 29 Mar 05

Excellent decision. It takes balls for someone to do this, and I’m glad you took the plunge. I sure do hope that major web services follow suit if they really need it.

Supporting old browsers really is depriving everyone of a better user experience.

JF 29 Mar 05

Excellent decision. It takes balls for someone to do this, and I’m glad you took the plunge. I sure do hope that major web services follow suit if they really need it.

Thanks for the encouragement, but we’re a small player so I doubt people will follow us.

Arne Gleason 30 Mar 05

“If that rule was hard and fast then we’d have to support IE 3 and 4”

We’re not talking about a bit of latitude in a guideline — You’re going from a published stance “Don’t shut out visitors with old technology” to an announcement of only supporting the latest version.

Personally, I always thought the rational for supporting any browser was customer base vs. cost to produce. It always seemed to me the “you must support every browser” sentiment came loudest from those with a distain for IE. The morally elevated position of leave-no-browser-behind rings hollow.

Post a comment

(HTML is allowed)