A question for web developers. | Hacker News Reader

A question for web developers.

1 pointRyoku2 comments

It is not a secret that a huge problem has grown larger in the world of web development. And I'd like to hear your opinion about it.

What problem am I speaking of? The soon to become over-talked issue of web apps developed to work correctly only on one browser(Usually Chrome) without even a reliable fallback to other browsers. I've seen many flavors of this issue in the wild, ranging from the "Proof of concept" sites that make clear their content only works or shows on X Browser, to sites changing landing pages depending on the User-Agent browser where only the chosen browser will have all functionality and the others will simply avoid being "too ugly", and to the extreme case of websites not rendering at all if you do not use whatever browser you are told to.

This is not a new problem, I believe anyone who has developed for some time with MicroStrategy, Appian, SAP or any other JAVA based ERP, BI or BMP, has also encountered this quickly annoying issue of being FORCED to work with Internet Explorer.

However, I do believe there is another side to the coin. A side where, for example, developers that only work for Chrome are able to push the bounds of the available tools and create really neat new web apps.

What I'd like to know is not only your point of view of how good or how bad this issue is. But if you have any ideas on if and how it could be fixed.

Personally, I'm so tired of meeting with clients asking "Why can't we do it if everyone uses (EG) Chrome" or being too afraid to hire new developers because themselves or someone near them has a spooky story involving one of this developers.

Experience tells me this is not a huge problem and my personal opinion is that it won't become one - don't have any references though just IMO. What I mean to say is that basically if you're making a production level app, you just build it to work cross-browser that's the way it is, you rarely do things based on UA if you know what you're doing and opt for feature testing instead. Maybe some provide neat little extras if a browser has a specific feature like webgl, but you're going to make sure all functionality is provided for every browser so you hit the widest audience.

I find it hard to believe you get clients that think everyone uses Chrome, that's never happened to me. I've also rarely had a client ask for something to only work in a specific browser, the few exceptions have been making things work for IE6 or older browsers that were used internally by larger corps.

These are just my thoughts, discussion is welcomed :)

Well yes, I completely agree with you. That is what we should do. But each time I see more developers that leave production level apps like this. And not even small websites like http://absolutmexico.com (Has a fallback to other browsers but the developer consciously decided to not offer the same functionality for browsers other than Chrome, even after been pointed to a few plug ins that would do it for him); but also huge Frameworks, like Appian, which forces you to use IE for administration.

And there's another side of the story, in both of the above scenarios, the product/site where developed by Engineers and Web Developers. But each time, plug ins make it easier for "non-code-savy" people to implement (for example) effects on their websites that would be really complicated any other way; which is cool until the moment you notice most of those plug ins are poorly coded. The problem grows as this people sell themselves as web developers and usually charge (as freelancers) much less than the full implementation of the job would be worth. I know this happens a lot, and by a lot I mean that 85% of the independent projects I had this year were about fixing this websites or making them bigger.

I think the problem might be that, in general, we talk about all the marvels of the internet and our work with it that it just makes people want to jump in. We give tutorials, plug ins, API's, and any other tool that makes the job easier, without really pointing out all that should be needed to actually deploy to production.

Reply on news.ycombinator.com