HTML5 as an analogue for the perennial security problem

Back in February I wrote:

Currently I can switch Flash off. It’s a plugin, it’s a nuisance, it’s proprietary, it’s unbundled, but *because* of that it’s easy for me to shim or nuke the flash-rendering module and *poof* my webpages become a lot more efficient.

When I added “ClickToFlash” to Safari, my daily morning ritual of loading ~70 tabs up with news, cartoons and zeitgeist got a whole lot faster.

But: if the video rendering is part of the HTML specification… then the ability to switch it off comes under the aegis of browser functionality, surely?

I am beginning to think that I don’t *like* that option.

Browser are not-so-hot at giving people selective “switch it off and leave it off” functionality; they tend to start with “surely everyone has a powerful machine that can do this” and then end-up with some moochily kludged solution.

…and this week UserFriendly twigged to the same issue:

After 25 years of faffing with security I am now pretty sure that this is unstoppable, and probably it would be a bad idea to stop it; innovation is risky and careless and (for whatever reason) it cannot proceed cautiously. Our browsers are heading pell-mell into HTML5 and nowhere have I seen a Preferences dialogue regarding “what [you the user] want to switch off” – other than a solitary control regarding whether data can be stored on your machine, that being a privacy and legal risk.

People on slower machines will have to wait some considerable time before developers realise that CPU is a finite resource which also demands protection; or they’ll just have to fudge their user-agents and pretend to be on Mobile browsers.

Comments

8 responses to “HTML5 as an analogue for the perennial security problem”

  1. Chris

    Certainly it *could* be a problem.

    But the technology already exists to block content in web pages – browser ad blocking extensions.

    I’d be surprised if something similar could not be made available to block audio and video. (Perhaps it already has, I have not checked)

    1. Indeed, chris; yet have you also considered that the simplest way to DoS an old browser to death was to use the FLASH tag?

      I fully understand the browser is moving from “browser” to “platform”, but the security model we have inside the browser is hardly better than Windows 9x, let alone has the comprehensiveness of Unix. Some day it will be a matter of other-people-running-their-code-not-yours-fot-their-own-benefit in your browser, and it’ll be a case of “let’s reinvent the multiuser operating system all over again.”

      The JavaStation was way way too far ahead of its time, and regrettably it should have been running JavaScript.

  2. Richard

    Amusingly enough, I’m writing this on a device that is immune to Flash security issues because it won’t run Flash. Mabye Mr Jobs has a point.

  3. Chris

    I remember animated GIFs rather sucking the life out of older browsers on older OSes! Also MS’s marquee tag…

    I think you’re right though – a much more comprehensive security model is surely needed in browsers.

  4. Actually I think the RSRE Flex Machine beat Sun to it by quite a margin, that was a capabilities based system and was intended to remove certain classes of programming errors. I just wish I could remember more of what Ian and the other authors told me about it when I worked with them!

    http://en.wikipedia.org/wiki/Flex_machine

    It also had a very nice A4 type screen, and was also very handy as a room heater. 😉

  5. Maybe this is finally a use case for user CSS?

  6. John Slater

    Alec, I would be interested to see your list of ~70 daily tabs.

    – John

    1. Maybe I’ll make a post of that… 🙂

Leave a Reply

Your email address will not be published. Required fields are marked *