Should browsers have an expiration date? | Hacker News Reader

Should browsers have an expiration date?

blog.garlicsim.org

10 pointscool-RR30 comments

Unfortunately this is an anti-feature: no one will choose a browser for himself b/c he wants it, and some ppl might avoid browsers with it.

I guess the only way it could be good for market share is I might want to install a browser like this for someone else, so they'll end up upgrading eventually even if I'm not around anymore. I'd be happy to mass install it at a big company if I could get away with it. Then even if I'm not there later, something good will happen.

I think that you are overrating the thinking process that users have. Most users are not like you. They don't carefully consider the pros and cons of the software they install.

Most users are like, "It's shiny, let's install it!" Okay, I'm exaggerating, but I'm saying that the vast majority of users would still download and install Firefox/Chrome even if they had an expiration date.

Why do people like that download FF/Chrome at all? They already have IE 8 I imagine.

I know a lot of non-technical people who install, and even recommend, Chrome because it is "the fasted browser ever". They have no idea about web standards but still have an incentive to try a new browser.

I would love this as much as any other web designer or developer, but sadly, it's just not feasible.

The biggest problem with upgrades is that lots of companies rely on specific versions of specific browsers for in-house websites and web apps. Expiration dates essentially amount to browser makers saying 'too bad, fix your software'. This would be amusing, but realistically, Microsoft would never do such a thing, and without Microsoft on board, not much will change.

The best compromise I can think of is for browser vendors to decouple updates of web-technology support (HTML, CSS, JS, Canvas, etc.) from updates of user-facing features, the things that get advertised when an full point upgrade is released, and start pushing out HTML, CSS and JS upgrades automatically, in the background. Once the first round of such browsers were widely adopted, browser upgrades would cease to be an issue at all. Upgrade averse users could keep their old versions with familiar interfaces and still have cutting edge standards support.

This leaves the problem of getting the first round of upgrades through, but that could be solved by building in backwards compatibility for all currently popular version of a browser. This would be a pain, but would only have to happen once. If Microsoft, for example, gave IE9 full backwards compatibility with IE8, 7 and 6, and made future rendering updates automatic, there would be very little excuse left for any company to stick with an older version of IE.

I don't think it's the UI that breaks legacy applications.. it's the rendering engine. Automatically pushing out upgrades to the rendering engine would be as bad as forcing upgrades from IE6->IE8.

I think you misread what I wrote. I was suggesting separate updates for the rendering engine to solve the problem of upgrade-averse users. Legacy applications would be accommodated by keeping older rendering engines running beside a newer, automatically updating engine. The idea is to remove roadblocks to getting everyone to upgrade, and then make it so that upgrading the browser version is no longer necessary to maintain current standards support.

This is why Chrome has built-in silent autoupdate - the vast majority of our users are on the a new version within a week of releasing it.

See graphs: http://www.belshe.com/test/velocity2010/#slide38.0

The shallow curve is our development and beta channel users, the steep part of the curve is when we release a stable version. Previous slides show known update rates for other browsers.

Here is the graph. The site sucks pretty bad (seems to require a very high resolution and has disabled scrollbars).

http://www.belshe.com/test/velocity2010/chrome_version_migra... http://www.belshe.com/test/velocity2010/firefox_version_migr... http://www.belshe.com/test/velocity2010/ie_version_migration...

Shouldn't all software have a clearly labelled expiration date like dairy products do?

That's a more general, interesting question. Having an expiration date to software is kind of annoying, and I think that it shouldn't be done in general. But, when it comes to browsers, I think this would be so helpful to the progress of the web that it would be worth the annoyance.

For software which could be interesting also far in the future (I only can think of games or other somewhat artistic software), I would make it in a way that the expiration date check is only done if the software is able to find an update.

Last year I used a computer that actually had a 486 cpu, whose only method of data transfer was floppy, and would have choked on anything other than the DOS it was loaded with. It had a power ROCKER, not a power button. It was probably about 20 years old.

It did it's job admirably- load EEPROM chips with it's paired, equally archaic EEPROM burner.

There is really no reason to replace it, besides saving you the annoyance of using floppies.

So, in a word what I am saying is computers that are tools need not have an expiration date. Their software does not need to be updated, because there is nothing wrong with it. Do you upgrade your 1970's corded drill just because it's old? No. It may not be pretty, but it does the job.

How many security vulnerabilities does your power drill have? Is all the wood you get backwards-compatible with it?

I could not agree more, if I download a software, I am the one who decides when and for what purpose I use it.

I agree that people should be strongly recommended to update, or (as in Google Chrome) update automatically if possible, but refusing to boot is just plain wrong.

You know, we effectively had this back in 1995 or 1996 when Netscape would release betas. Each version would expire and require you to upgrade a few months afterwards. The Web was very different back then, but I think that this could possibly work for browsers like Safari where there aren't a bunch of folks with plugins. But breaking my favorite Firefox plug because the plugin creator was on vacation for six months and missed a browser upgrade would be not so cool.

Many corporate IT departments run old browsers because they make use of proprietary intranet web applications and don't feel like developing them on multiple platforms. To them, this system would seem pretty stupid, and indeed, would prevent them from ever upgrading into new expiring browser versions.

It wouldn't bode well for public relations between the browser designer and the user. I see the reasoning this guy has, but this proposal is literally planned obsolescence. That's a scary notion to a user (or consumer).

Could the big browser makers all agree to set expiration dates on all their browsers?

They could! It'd be as much a runaway success as software expiration was for eradicating piracy, and I doubt anyone is competent enough to roll out cracks or patches to "fix" people's broken browsers.

I think the situation is different. If your browser expired, would you (A) search for a crack or (B) click the "update to latest version" link?

I prefer silent updates.

Silent updates should definitely be done, but they will not work under all scenarios. (e.g. computers that reinitialize themselves on every reboot like in some universities, computers that aren't connected to the Internet like in some paranoid companies, etc.)

In these situations the expiration date idea wouldn't work either because they would either find a workaround or refuse to install your software at all, knowing that it forces them to go through the upgrade process in the future (and on your timeframe instead of theirs to boot).

Browsers that are never connected to the internet (IMHO) can safely be considered a non-issue.

How do you download a new browser when your old one is expired?

The expiration dialog will download one for you.

The issue is that doing this is forcing the issues of a few people on to a much larger set. If someone doesn't want to write code for that browser just don't. If they're feeling really nice, pop an error message telling them that their browser is expired.

Oh, whats that? now the site isn't going to bring in as much revenue? Now they have a choice, whether the revenue is worth more to them than writing the code.

Those "few people", a.k.a. web developers, are those that are responsible for progress and innovation on the Internet. If they have a problem, the Internet has a problem.

silly idea

You're on HN; You are expected to be slightly more articulate in your criticism.

Reply on news.ycombinator.com