IE only isn't a web service ... it's crippled desktop software. My last company used an EXTREMELY expensive service called Windchill. It was used because CVS was free, worked cross platform, and relatively bug free. Fractal wrong, so to speak.
Windchill only comes available through IE. That meant that the 1/2 of the company on linux boxes needed dual boot machines to use it.
Running IE precludes an operating system and hardware on which it runs. If you asked about developing for Firefox only then I'd strongly recommend it. This browser can be run on many operating systems. Most importantly, early adopters who are likely to use this browser are more likely to be an early adopter of your service. For the same reason, I'd recommend developing a Firefox plugin rather than an IE plugin.
>"If you asked about developing for Firefox only then I'd strongly recommend it."
I know that this is a tech site, and that I am probably get modded down for saying this: But that is absolutely ridicolous. Basically what you are doing if you are only developing for firefox is opening a shop where you only let 10% of your customeres through the door. The rest will have to go home, or go to you competition.
First of all, FF has a global market share of about 21%.
If you were looking to test an idea to the "early adopter" crowd (the "echo chamber"), starting with FF/Safari is the smart way to do it. What percentage of tech bloggers of note use IE?
Assuming your v1 test is successful and it's time to "cross the chasm", of course, IE is a must.
If you'd like a little data thrown in, our userbase (after being on TechCrunch twice and LifeHacker 3 times in the last few months) is only 21% IE.
Even at 21% market share if your core audience is intelligent/computer savvy theres a good chance your user base could have a higher FF ratio. I work for a website that has a target market of college students and our FF user base is much higher then 21%, most likely because they are young and computer educated.
No, the other option is for your customers to install Firefox. Now this isn't always something you can demand/expect of your users. But for example for a free service or during a beta period it's perfectly reasonable. Most IE market share is not early adoption friendly. If they wouldn't fear new technology they would be on Firefox already. Either they never heard of Firefox or don't want to try it out. Neither one is a quality that indicates early adopter.
>"No, the other option is for your customers to install Firefox. Now this isn't always something you can demand/expect of your users. But for example for a free service or during a beta period it's perfectly reasonable"
So you think that users will go through the trouble of installing a new browser to see a free website or a product in beta? Good luck with that...
It seems to me that a lot of really good programmers (of which you seem to be one, judging from your comments) don't mingle a lot with "normal" people. These are the ones that have normal lives, use their computer only as a tool that is expected to just work, and use IE because it came with the computer.
And they are the ones that will pay for products. Early adopters are a really hard crowd to make money from, they expect everything to be free, and they will move on as soon as something more exciting comes along.
I would never launch a product, not even in a beta, that only works in firefox, Not only do you lose more than half of your potential customers, but it also sends a signal that either you are not very good at what you do since you only seem to be able to make it work in one browser, or else that you don't much care. Both are pretty bad.
Well it's certainly true that I probably mix with many regular users. But of the people I know my parents are the least likely to buy products from the web or sign up for a subscription website. But yes, they do run IE because it came with the computer.
On the other hand a lot of my friends run either Ubuntu or OSX and shop online often. They tend to try out new services while they're still in Beta and sign up for paid services.
So that's what I personally base my decision on when I take a first cut at building a webapp. YMMV
Well unless your friends are your core audience you should be wary of that approach. You tend to hang out with people that hold the same values that you do yourself. If you think that they are representative of the population as a whole you are headed for trouble.
I interact with "normal" people, and at one family I know switched to Firefox because of the piclens plugin.
It's actually not that uncommon for Windows users to install software on their computer off the Internet. They do so more than I do, actually. If I can't apt-get it I usually pass.
When you're running a business, you have to succeed in all areas of activity from coding to marketing to admin and a whole lot of miscellaneous tasks. Do the 10% in each area which gets 90% of the result. Then backfill as appropriate. If you're developing an AJAX website then it is significantly easier to develop it using Firefox. Early adopters probably won't notice that it doesn't work on other browsers. Then fix the compatibility when you've got time.
I'm astounded at the level of complexity facing contemporary web designers. I'm also appalled by the level of compatibility and stability between websites, browsers and plugins. 10 years ago, I could design a webpage and test on Netscape4. With no further testing, I knew how it would work or degrade on all browsers down to Lynx and Mosiac1.2. Many "designers" couldn't achieve that compatibility, even with a few iterations of feedback. That was just static HTML. Nowadays, you've got the added complexity of JavaScript1.5, CSS and smart web frameworks.
Yes, absolutely focus is essential. But once you decide what to make you want as many customers as possible. And that obviously includes not turning away half of your customers at the door.
Which is what you do if you don't support IE. Unless of course you have chosen a niche where 95% of your users use firefox. I have yet to see the stats for such a site.
I actually considered doing the web application I'm doing at the moment as firefox only, purely because it heavily relied on SVG. It's very much a niche market where my competitors are desktop applications, and it will be a fairly expensive subscription service so it didn't seem unreasonable to demand users download firefox.
Eventually though I just came to realise that the biggest advantage of the web was the ability to have anyone come to my site and be ready to use the product. So in the end I've gone with flash.
That's likely a good decision. I wonder what Firefox's market share would be though if sites like YouTube or MySpace required the users to download Firefox first. Most of these people are the same ones that install random "free" screen savers and games that their friends send through email. I doubt they would mind having to download just another piece of software.
See, I wholeheartedly agree with you about IE, but flash suffers from many of the same problems.
It's a plugin, so right out of the gate a whole class of people won't be able to use it. Plus, even though many have flash installed, versions vary widely, so unless you code for flash 6, again, good luck working with everyone's computer.
There's a right way to do things, with built in browser support, across all the major browsers. We spent time up front to make sure things did work, and now we continue to reap the benefits.
I did once see on a comment here that more people had flash in their browser than had javascript turned on, but I can't vouch for that.
Of course I am using SVG which means firefox or a badly supported plugin for IE. If you didn't need flash's interactivity, graphics or sound it would be pretty silly to use it.
Flash may "reach" my desktop, and even work sometimes, but I find myself having to kill and restart Firefox a few times a day when it quits working. Even when it's working, there are a lot of sites that don't work correctly. Of course, if you're going for market share and don't care about people like me, Flash is just fine.
I only care if you're an Orthodontist :) Really though, if you can realistically use html and not flash I'm sure its better. For my application, I've tried it both ways and flash outperforms native firefox SVG support, let alone trying to get it to work in IE with the adobe plugin.
I think the big flash question remains whether to use long polling AJAX or flash for server push type applications.
Given the standard of software for vertical markets ( http://thedailywtf.com/Articles/Rutherford,-Price,-Atkinson,... ), anything via a browser would be an improvement. In the worse case, you could wrap a browser component around your own client. This would allow you to install the browser of your choice on the platform of your choice.
Yeah looked into using XUL to do something like that. Now I'm down the track a bit if I needed a desktop or offline version I'd use Air. I'm not exaggerating when I say the interactivity of the flash version killed the native firefox SVG one. I appreciate that there is a not insubstantial anti-flash sentiment around, but its not likely to involve my customers and frankly I they're wrong in my applications particular case.
You are of course in an enviable position... I would hope that most people actually working on a web site use something like Wine to be able to test it in IE.
No way do I use it one minute more than I have to, though.
I almost never test things in IE. I'm counting most of the time on the JS libraries I use to be cross platform and defer to a JS guru when some IE only bug shows up. Most of the time one of the many other people that have to look at a prototype will be running IE and report an issue. It's not worth it for me to bother (is it possible to run IE7 in Linux now? IE6 is obsolete).
I run Windows in a virtual machine in Linux to test things. Sure, if you use some library like JQuery or Mootools most of the time your javascript will be cross-platform. But what about the CSS?
Of course IE6 is obsolete (since v0.1 of Firefox-Firebird), but that depends of your target market.
I was actually referring to Microsoft dropping support for IE6. Even though IE6 market share is still high with no Windows Genuine Advantage requirements for IE7 some sites are dropping support for IE6 and requiring IE6 users to upgrade to IE7.
I suppose this will be the general reaction here. If it's possible we could we have of an insight to YC.news stats of what the majority uses here, I guess Firefox/Safari wins over IE?
it's not much of being IE centric as to designing an api first and insteaad of losing time on stupid JS/DOM dependancies (things you can fix later for less cost), just get it maybe to most users for whom IE is the internet.
IIRC there are more IE related JS/DOM bugs than Firefox ones. In FF you also have Firebug which makes development less painful. So why would it help to make an IE version only first? The only reason I can think is if you're using ASP.NET and the funky HTML generation that it and Visual Studio do.
of what i know of IE suffers from certain memory leakage issues and that alone makes you write more careful JS code, that is if you are aware of them. So, if you can avoid these and achieve satisfactory functionality between all recent versions of IE (especially when putting these on their limits) , you can later fix your code working for Mozzila. Of course you could do it simultaniously but would it worth to lose even more time in cotrast of going online. (not a .NET. fan)
I would not! Though I could use my IE for Wine, why would I go through the trouble just to visit a site? If they're not going to cater to me as an Ubuntu user, then they can count me out of their target population.
Its like what paul said, except for me, even though I technically "can," I won't go through the extra trouble.
Windchill only comes available through IE. That meant that the 1/2 of the company on linux boxes needed dual boot machines to use it.