Live data from Hacker News

Supporting IE Is Too Much Work

bartvanzon.com

141–150 of 176 posts

Re: Supporting IE Is Too Much Work

#141
post #96

Earlier quoted context omitted.

I tried to encourage my team to test better against IE by creating a nice setup on which they could easily use all relevant IE versions. After two days of struggling with all the insanity of Windows licensing, installing and the crap support for running older / multiple versions of IE I've seen more than enough of Microsoft for the next five years. Nothing has changed since 2000. Just setting up the tools for support…

What I have trouble with is why that's valid when on the reverse side of the coin, one must literally buy Apple hardware to test in an Apple environment.

> one must literally buy Apple hardware to test in an Apple environment.

You have to consider the total cost. If the Apple hardware lasts longer, requires less maintenance and makes the developers more productive, it may end up very well paying for itself. My wife uses a Mac and just the time I didn't spend fixing, cleaning and disinfecting a Windows install more than paid for the price difference between her MBP and a cheap Dell. Compare it against a similarly well built machine and things look even worse for the PC. Her current Mac (a 13" i7 MBP) survived a car crash in early March (the machined body is very slightly warped and it'll probably have to make a visit to an Apple dealer for that). Her previous one, a 2006 white MB, sits on my desk as my secondary computer and is our main source of ambient music. It had an aesthetic problem, which was fixed by Apple for free (because other parts had to be replaced due to a recall) last March. It replaced a (still working flawlessly) 1998 iMac in that function. You can say anything about those machines except that they aren't built to last and that their customer service isn't stellar.

If I had to use my Dell with Windows, I know I would be far less productive than I am under Linux. I know because I tried (from 2008 to 2010).

Unless you are developing for Windows (something that pretty much implies you are running Visual Studio), I wouldn't advise you to use Windows as your development platform. And, if you aren't, it's only natural that testing on Windows incurs an extra cost.

Having said that, it shouldn't be that complicated to set up Windows VMs to run automated Selenium-driven tests and plug those into Jenkins.

If you test manually, you are doing it wrong.

Re: Supporting IE Is Too Much Work

#142
IMO part of Microsoft's strategic view is that any developer time spent on non-Microsoft technology is bad, so they saturate the developer with Microsoft-specific requirements and deviations from otherwise well-established standards.

Microsoft's proprietary software development products (Visual Studio, .NET, etc.) present the developer with complex non-standard terminologies, languages, categories and specifications unique to Microsoft tools and platforms.

In the more open sphere of the WWW where Microsoft has at times dominated the market, the same strategy applies: inundate web developers with deviations from the standards and with Microsoft-specific "innovations". It works as long as Microsoft has a large share of the market because you simply cannot ignore their presence. In the end, much, if not most, of your time is used up supporting Microsoft's idiosyncracies, leaving less time for alternatives.

Re: Supporting IE Is Too Much Work

#143

Earlier quoted context omitted.

> If Microsoft facilitated having VMs of each windows installation + browser version available for that OS to developers Which they do, they've been offering complete VM images with various combinations of Windows and IE for some time, the current offering is XP + IE6, Vista + IE7, Win7 + IE8 and Win7 + IE9: http://www.microsoft.com/en-us/download/details.aspx?id=1157... While the images are for VirtualPC, I've yet t…

IE9's debug tools are hands down the best at this point in time. Although Chrome's is cacthing up, pound for pound you will be far more productive in IE9s debug tools then firebug, dragon fly or any other at the moment. I think there is just too much bias at play here.

I have downvoted you because that statement is so absolutely wrong.

I know because I am debugging a stupid, stupid IE only issue and I have to use a combination of Chromes debugger and judicious use of the console/repl in IE.

You can't even inspect an element that was added in code in IE.

Re: Supporting IE Is Too Much Work

#144

Earlier quoted context omitted.

I feel like people are lying to the,selves about IE9. It's still not good enough! The key words you used were that hard implying you know it's still just a little tough. Ever try to get a gradient to work in IE9? First off, when it works it doesn't render anything like it renders in any other browser, and second, you still have to jump through hoops or all your gradients will be blue even if they're not meant to be b…

Nobody is saying IE9 is great, but you can't deny it's good enough for most apps. It's fast, has a decent javascript engine, a hardware-accelerated canvas, some HTML5 support. Their greatest sin was leaving out text-shadow and gradients, but IE10 is just around the corner. Even if it was complete crap, it has a 30-50% share, you can't justify killing off that much business unless you develop a Mac/Linux-centric or br…

That is true. It is good enough but good enough is no longer good enough. We've been spoiled with Firefox and now Chrome. There's nothing wrong with being spoiled either! Practically speaking, it's not a good idea to leave IE out no matter what the version. That said, I'm so glad sites are not only excluding IE but publicly bragging about for better or worse. My hope is that from this point on beginning with IE10 Microsoft adds support or better support for new html5/CSS3 technologies to the point where we think of it as an equal to other browsers. Developing for Chrome, Firefox, and Opera is trivial but still not IE even version 9. It's not about IE bashing. I don't care who makes the browser or what it's called so long as I'm not developing multiple versions of the same site just to support a single browser.

As others have mentioned, market share can be misleading and audiences count. But even so, you're still right that it's not practical to leave IE out. But since people are doing it I hope it encourages IE to take notice and improve instead of digging in their heels and insisting they're just as capable as the next browser.

The real problem however is fragmentation. I'm sure IE will improve or die in the future but Microsoft just shot themselves in the foot by restricting the amount of people who can upgrade. People often say that FF and Chrome will one day be fragmented and a drag on front end developers too but I disagree. Even older versions of those browsers support far more than than IE8 and below support and in some cases they even include features IE9 doesn't support. In the future there will be far less fragmentation of non-IE browsers because the vendors are way better at backwards compatibility. We'll easily still have to be concerned with IE7-10+ in the next five years because the number of people using them will fall at a snail's pace while market share for old versions of other browsers falls much faster. This is because many IE users simply cannot upgrade even if they wanted to while everyone else can. Nowadays Microsoft seems to be pushing updates more. Good for them. It's too little too late.

For now we can't deny IE still isn't good enough and I'm not about to congratulate them for being one step behind everyone else with each new version of IE. While its not practical to exclude Internet Explorer I hope the trend continues just so that it pushes Microsoft to continue improving the browser like it has recently. Maybe in a few years, when old versions of IE (the terrible 6-8 versions) have finally lost almost all their market share and the still shitty but far more manageable versions 9+ are all we have to contend with (and I'm sure by that time IE will finally be on a level with its competition) we can quit being "whiners" and "lazy" and "hipsters" and "elitists" and "ignorant" about supporting it.

Re: Supporting IE Is Too Much Work

#145
post #22

I think a general problem with "webdevs" is they have "year zero" mentality - anything prior to the release of their favourite framework, didn't really exist, and if they do acknowledge its existence, it's only to observe how worthless it is. It's almost religious in nature. You see this in particular in the Ruby world where they are forever reinventing the wheel, but it's certainly not limited to them. IE6, because…

There are many great communities, I will single out Ruby because you chose to do so in your post.

For many in the Ruby community, nothing does exist prior to their favorite framework. This is because a core ethos of the ruby communities has been shaped by people like _why and Ryan Bates at railscasts, people open up some of their best code as community gems for all to use. And "non-programmers" have realized that coding can be a fulfilling form of self expression.

Coding as an expression of creativity, where dumb questions get answered instead of ridiculed or RTFM'ed, and where people are proactive in sharing solutions (each "reinvention" of code, each gem dealing with the same damn thing, addressing the same scope, but each one still solves the problem and has value as personal expression).

The communities push for "best practices" without shunning personal expression or eating it's young is why it has so many new adopters that are excited about it, as opposed to the reaction they get from other high profile open source projects, they are embraced and encouraged instead of treated like a nuisance.

Yeah, you may see the an almost religious excitement among new devs, but it is just the excitement of a child who has found someone to learn from, and they have had their vista for creative expression expanded, this is hardly a bad thing.

Have you ever had to deal with the Linux kernel team? If you knew some of the asshattery open-mpi has had to deal with over the years you might better understand. The ruby community has a lot of crap in it, much of it being produced by new devs who have a zeal that makes them annoying, but in my mind the communities distinguishing mark is its kindness and helpfulness, they go out of their way to help new devs, no wonder there is so much zeal coming from new adopters. I wouldn't have it any other way.

Basic was my first language over two decades ago, I am glad that my kids will have ruby, _why, shoes, and a community that will answer all their stupid childish questions without ridicule or disdain. A flawed but intentionally friendly community is refreshing in opensource, the land where people just defend their kingdoms and castles.

Of course I understand your nostalgia for day 1 of IE6, I still have nostalgia for Gopher, but if someone told me I had to develop for it because a company was exploiting its monopoly to stifle and ignore the need for innovation, was slower in implementing, and sought to undermine open standards and then had the audacity to claim it actually cares for "Developers, Developers, Developers" it would legitimately come to represent all that I hate about that software company. Both the nostalgia and horribleness are true and appropriate to remember.

Re: Supporting IE Is Too Much Work

#146
I agree with the point in general, but for one subtle detail:

What shocked me even more is the amount of comments saying that “supporting IE is too much work”, not because it still costs the crazy amount of time it used to (ie6), but because they’re developing on a Mac.

No, it's not that it's difficult on a Mac. It's that it's difficult on a [not Windows]. Subtle difference, but undermines much of the "grow up" point of the article. You're saying developers are whining because Microsoft doesn't support their operating system of choice. But that's incorrect, they're whining because MS doesn't support anyone's operating system but their own.

And even if it was simply that you can't easily debug IE on a Mac, the fact of the matter is, all the other major browsers allow it. Who cares if it's unreasonable to ask Microsoft to provide us a build for easy use and debugging? Every other vendor (i.e. competitor) does it at the exact same price point: $0. If IE lacks this "feature" at the same price point as the competitors, then I say deciding not to support IE is fair game (though I wouldn't personally exclude IE for anything other than a toy side project).

Re: Supporting IE Is Too Much Work

#147
post #88

As a front-end developer who attempts to keep up with the latest and greatest in my industry, I have to say that there does seem to be some sort of bubble with people on their Macs. Too many times I've seen documentation that assumes that everybody is on a Mac. Too many times I've seen demos that assumes that everybody is on a Mac. Too many times that software that can easily be cross-platform are not or they don't d…

When you run your own site your job is to do whatever you want, however you want, it's your chance to be opinionated and exclusionary, and as long as you are not begging me to invest money, more power to you.

Also open source and sharing is more robust on linux and posix platforms, so people naturally develop tools they can use on their own platform, with their own workflow, and who cares if it gets ported to windows, I'm not creating a product to sell, I am sharing my personal toolset in a friendly way.

Why are you complaining about the difficulty of installing a FREE opensource project on a platform that is not native to the Devs? Of course it's more difficult! Why are you complaining so loudly? Why are your expectation so entitled? Did you even bother asking for help on Stackoverflow?

Re: Supporting IE Is Too Much Work

#148

Earlier quoted context omitted.

For me the point is less about: IE is no longer that bad. It's true that IE9 and IE10 are much better and supporting IE is no longer a 2x web-dev penalty, but that's not the point. The point is: when the vampire is exhibiting weakness you drive a stake through its heart. The vampire turned a bunch of your fellow townspeople into vampires; saying the bite is no longer so bad doesn't mean that the vampire is now to be…

So why would you sort it in FF then open IE why wouldn't you start with IE then go to FF. I really think you are starting in the wrong spot.

The tooling around FF was/is remarkably better than that with IE, so I always found it easier to start there head to IE than to start by groping blindly in IE.

Re: Supporting IE Is Too Much Work

#149
post #129

Earlier quoted context omitted.

It's not really difficult at all to test. MS provides the virtual machines free of charge, and VirtualBox is free as well. That's everything you need to run and test in IE, from IE6 to IE9 right now.

If you want to test multiple versions of IE, you need to do some time intensive VM cloning to trick VirtualBox into thinking they're actually separate images. Then configure and install IE's debugger tools for each image. Then, when the VMs expire in a few months repeat that process. Add to that the problem of VMs booting slow, running slow, and consuming tons of memory on my otherwise-fast dev machine.

> If you want to test multiple versions of IE, you need to do some time intensive VM cloning to trick VirtualBox into thinking they're actually separate images.

No, you don't. You just run different images. Downloaded from MS. I have them installed. Right now. IE7, IE8, IE9. IE9 is currently running. Even still, for the most part, IE9 makes debugging easy, as you can run in IE7 and IE8 mode. That takes care of the rest of the problems I've faced.

Yes, testing in IE sucks. It's not an enjoyable experience, but it also doesn't take much effort.

> Then configure and install IE's debugger tools for each image.

This is not an issue. You make this sound like it takes a lot of work. It's not.

> Add to that the problem of VMs booting slow, running slow, and consuming tons of memory on my otherwise-fast dev machine.

I'm running this on a MacBook Air. Yes, they aren't lightning fast, but they are usable for debugging purposes.

> Then, when the VMs expire in a few months repeat that process.

That's not true, either. I've been using my IE7 one, for example, since October.

Re: Supporting IE Is Too Much Work

#150
If I had to guess, I'd say that most of the complaints don't involve layout issues but rather, support of newer features in HTML5 and CSS3. That being said, there are plenty of (easy) fixes and workarounds that take care of most problems. Is it frustrating? Yup. But like the article says, it just comes with the territory. For newer front-end devs, I'd recommend dedicating a few days familiarizing yourself with the various bugs, fixes, and tools for testing. That alone will ease the pain. Just takes time like anything else.

Chin up, sport.

Post reply on HN