Slashdot Mirror
When Should You Stop Support for Software?
hahafaha asks: "I am currently working on a website for a small organization. We (I am not alone in this) have a beta version ready, and are currently testing the site on browsers. We have tried all of the big browsers (Firefox, IE, opera), as well as other browsers, such as lynx, links, w3m and even NetFront. So, when can one decide that they will stop supporting a system. Obviously, going (for example) down to IE 1 is crazy, but is IE 3 crazy? This is not only relevant to web design but to any programming at all. When, for example, can you say that I will *not* support a certain version of Windows. Can you say that now about Windows 98? How about 95?"
Depends on if you consider x% of the interweb population to be valuable to your business.
// file: mice.h
#include "frickin_lasers.h"
That is what I see. When the vendor drops support - and that can range from normal EOL to extended contract based EOL - it is time to stick a fork in it. Sadly, it looks like I get to keep a copy of Solaris 8 running for a few more years....
+++ UGUCAUCGUAUUUCU
Whenever the cost of supporting the customers that comes from supporting those customers, exceeds the benefits of satisfying those customers.
The trick is determining the costs and benefits. But often it is not that hard.
That's a business decision, not your's.
If the company is willing to pay you to support old browsers/OS's because the company is getting something out of the clients with those browsers/OS's, then that is their concern.
I develop websites as well as part of a much larger firm. We stop providing support for older browsers (Like IE 5 and 5.5 Mac) when MS decides to stop supporting them.
:)
.8, and Safari (forget which version).
We will only test on XP, Win2K and win 98, but not 95... (that's just silly
Our browser support goes back to IE 5.5 Win, NS 6, FF
Take the hint from others and you will be able to justify your actions.
The Code Ninja is swift with his tool, precise in his delivery, and deadly accurate in his execution.
I suppose the obvious answer would be "What is the lowest level that you could reasonably expect from your userbase". For a site touting the latest and greatest in web technology, you might be a bit heavier in your requirements than for, say, a site on nutrition.
For regular applications, you might ask yourself what the lowest level is that can reasonably be expected to do what's required. i.e. if you need a gig and a half of RAM for most operations, you might not support Win95 simply because it can't support you RAM-wise.
Then, even if you could do it in '95, would your userbase still be in '95? Really, it just boils down to "what's on the machines of the people you want to serve?"
You are standing in an open server west of a blue house, with a boarded front door. There is an Exchange mailbox here.
Yeah, what happened to "Degrade Gracefully".
I mean, if you're entire business is a web app which requires CSS and modern javascript... then support what you need to support. I'd personally support firefox 1.0+, netscape 6.0+, IE 5.5+. That will encompas more than 99% of people; after that I think it's really diminishing returns (pre-IE5.5 means pre-windows98).
I can't see supporting netscape 4.7 anymore. It was a good browser, but it was released in what, 1998? It's time to move on, folks - it's been 8 years. It doesn't support CSS and iframes properly and a whole bunch of stuff. Trade in your SparcStations and PackardBells for something modern, please.
Just attempt to make it degrade gracefully.
~Will
sig?
Whatever you end up doing, don't block browsers out with the horrid "Sorry, you do not have Internet Explorer 5.0 or better" message. Most of the sites that show that message, I can view just fine if I can manage to get past the browser-blocking "welcome" page. Let the browsers "try" to view the page, even if your "what kind of browser are you?" check thinks it shouldn't be able to. Even if it doesn't display perfectly, the user might still get the information they were looking for.
-- I prefer the term "karma escort."
(total number of users) * (% of users using browser) = # of users who you won't be supporting.
.5% of total market share, but you get my point).
We have a two-tired philosophy: we don't test with browsers that have 5% market share, because we're a small business with limited resources. However, if a user reports a problem in a 5% browser that's easy to fix, we'll fix it. If it's a fundamental issue (lack of CSS support, etc), we'll just say "sorry, can't do it."
If it's not fundamental but not easy to fix, we'll consider the direction that the browser's market share is going in. An IE 4 problem that would take a lot of time to fix is not as important as an Opera problem that will take a lot of time to fix, because any work we do to support IE 4 is less and less valuable every day; Opera work should be worth more or less the same in a year that it is now (yeah yeah, it may gain another
As you get more users, that threshold drops. If you've got a million revenue-generating users, it only takes a fraction of a percentage drop in revenue to justify the resources needed to support an old browser.
Cheers
-b
If I wanted a sig I would have filled in that stupid box.
The gold standard in this case is WILL IT MAKE MONEY. If supporting users on IE3 costs more money than you'll get FROM users on IE3, don't do it. Simple.
-----
PGP Key ID 0xCB8FF658
I'm not trying to attack or troll, but seriously, you can't develop a product to beta stage, and then start questioning whether it should run on hardware/software X or Y.
The correct way to go about any project is to identify the target audience and their technology, and develop accordingly. 12 years of bone-headed decisions have taught me this simple truth.
Never build a house first and then question if the design was right or the tools were chosen correctly - identify what you need in a house first, design it accordingly, and then pick the tools to build it.
In my experience, most users of Opera and Firefox won't fall back to IE if the website appears broken. You've already pissed them off by not working with their preferred browser. If you're not somehow handing bars of gold through the screen, they won't stick around longer than it takes to close the tab.
Batou: Hey, Major... You ever hear of "human rights"? Major: I understand the concept, but I've never seen it in action
I think that it depends entirely on the type of site.
I like to give the example of a local company that was offering some sort of website video streaming software for smaller retail firms. About a year ago, I was forwarded an introductory letter with a demo URL. My default browser, Mozilla, did not load the page properly at all -- I didn't bother to see if it would work in IE or not. Simply put, if you are trying to sell web based software to technical users, you better have the site work in more than just IE.
However, if it's a website of a smaller organization (that isn't technically orientated) that doesn't have the resources to spend on extensive compatibility testing, I will often cut them some slack and try IE.
Here's a crazy idea:
Instead of coding for specfic browsers, write valid code!
That was the whole intent of the web in the first place.
I always find it ridiculous when a website talks about what browsers it "supports." Websites should not be browser-specfic.
Also:
USE AS FEW FEATURES AS POSSIBLE.
I can't count how many times I've seen things that could have been done in simple HTML, done instead in flash, java, javascript, activex, etc. The more different technologies you use, the more you'll get screwed up by subtle glitches in their implementation.
In short, pick a handful of good technologies and implement them properly. Support users by pointing them to software that is not broken.
Life is too short to proofread.
Grandparent post:
Do you use java, javascript, CSS, flash, CGI, etc., or not?
Your post:
No, a flashier website will still work just fine on lynx, if it's done competently.
That's an awful broad statement to make in response to a post that gives five specific examples (some valid, some not). However, grandparent poster did not give sufficient detail, but I'm bored and will give some.
1. Java. I fail to see how a visually oriented java based website will work "just fine" in lynx, regardless of comptence. Let's take a good example of when to use java - I have a number of server software packages that use java based websites to provide system/software monitoring capability, specifically real-time graphing of various things. Lynx cannot provide that. If I'm in text only mode for whatever reason, I'll monitor the servers using text utilities.
2. Javascript. Moving into something I've written recently, I have a nice AJAX based based database front-end. It's meant to allow users on Windows, OS X, or Linux to graphically manipulate the database. It does so very nicely according to all of the users. Lynx cannot do what's required for the application. However, again, if I were trying to work the console, there are text based database front-ends. The key is to use the appropriate tool.
3. CSS. OK, grandparent loses some points on this one, as most things you do with CSS don't affect lynx, in that it simply ignores the CSS and presents the content in plain format.
4. Flash. I'll assume that the flash content is something that would be useful to the viewer and is, per your statement, "done competently." This eliminates sites that use Flash "incompetently" - doing things like using it for naviation and not providing html links to the same content and so on and so forth. This still leaves us with interactive meida, multimedia presentations, online tutorials that simulate applications, and various front-end software as discussed in points 1 and 2 that's also possible to do in flash. Unless you've convinced lynx to download the flash file and hand it off to flashplayer, none of these will work with lynx.
5. CGI. I'll give you this one, as whether a website is using CGI or not really doesn't have much effect on whether a page will work on lynx or not. I suppose maybe the poster was getting at the fact that many of the clever CGI programmers these days also integrate java, javascript, or flash into their applications.
So that gives you two points and grandparent three. I award the belt to him.
Really, what it comes down to is evaluating who will be using your site, what they're doing, and what their needs and expectations are. Most of what grandparent posted about aren't used in a *needed* way on public websites, but are extremely useful when done correctly. You also need to evaluate what portion of your site is reasonable to have higher requirements for. Are you simply presenting information or pushing the envelope into increased user interaction?
Google.com works with lynx, while google maps does not. Part of what google maps presents (directions, things near places) *could* be presented in lynx, but you know, doing so would take a very large amount of effort for virtually no payoff. I don't think google stockholders are loosing too much sleep over the issue.
Similarly, my main website supports and has been tested in IE 5.x for Windows and Mac, IE 6, Mozilla, Firefox, Safari, Opera, Konqueror, Lynx, and Links. It looks virtually identical in all of them, but doing so required some horrible kludges that make the code harder to read and understand.
On the other hand, my web applications (both internal and for public use) support IE 6, Moz/FireFox, and Safari. The code is clean and simple, and works in all three with the exact same code for the most part - there's very little that's coded based on which browser you're using (obviously, the AJAX calls are different). I could spend time devising wa
Get a clue. Don't support the browsers. None of them. Don't support the IE series or the Firefox browsers.
Support to a set of standards.
Well, that's the principle. Since 90% of the web surfers (less on tech-savvy sites) use IE, I suppose explicitly supporting the latest version of IE is a good idea. But other than that:
I'm sure there's a lot more that every webdesigner should know, but this is a nice start.