Chris Beard has been encouraging us to think like the rebels; what can we do that other people won’t do? How can we make an impact?
Here are some of my thoughts:
- The internet, in global average, is getting less reliable, slower and more laggy. Finish Janus and persuade our mobile partners to deploy it and default to it. Your Firefox OS phone now accesses the net faster than an Android phone.
- Make Firefox OS connect by default to openwireless.org access points, and encourage Firefox OS users to run them. There’s a virtuous circle here. More net in more places; a global movement of being generous with net access.
- Finish Daala faster, by finding people other than the core team to do everything except design the codec and write the algorithms (e.g., testing, speed optimizations, fuzzing, writing Windows Media plugins). We need to get the word out that this project is critical.
- Show the core free software community, who have great influence over tech choices and who should be our natural allies, that we care about them. Be the first organization ever to make a free-from-bottom-to-top mobile phone (running Firefox OS) and give some help to the Replicant team to port to it as well, just to prove we mean it and it’s real.
- Make it possible to search for specifically open source software in the Marketplace, and show we believe it “promotes the development of the Internet as a public resource” by promoting apps which are open source.
- Ship Collusion (which has been in the works for years), even if there’s not a perfect mapping between what it shows you and what’s actually bad. Make sites feel they have to justify all their 3rd party links.
What are your ideas?
I think OpenWireless is essential. The Net Neutrality discussion misses the point. We can’t legislate to make internet access fair — we have to actually decentralise the internet.
We’d need to be very clear about how Janus preserves users’ privacy — t first glance it looks scary, assuming you don’t trust Mozilla any more than Google or Microsoft.
What this headline describes so firmly falls in line with classic Mozilla ideals that the name of the app in question is almost a sort of cosmic taunt about the state of mozilla.org in 2014.
How would Janus be faster than the compression client already built in to Chrome for Android? My brief reading of the Wiki page for it suggests that the two are similar in concept.
As for Daala, I very much wish you the best, but I am concerned that it will be at best no more successful than VP8/9, which is viewed by most observers as a failure right now (even if it does have YouTube in its corner). It’s not totally clear what sort of benefits would accrue from usage of Daala that wouldn’t accrue from usage of VP8/9 either.
Finally, I think shipping Collusion when “there’s not a perfect mapping between what it shows you and what’s actually bad” risks the crying-wolf effect, both in terms of how users would actually perceive the product, as well as how critics would evaluate it. The result may poison the well for other (possibly better?) attempts at solving the problem; at best it seems unlikely to have much positive effect.
I guess my take as a Chromium engineer is that this is an excellent area to ponder, and I hope very much that Mozilla hits on some powerful ideas — but as for this initial brainstorm, some of these don’t seem very “rebel” if companies like Google are already doing them, and others seem unlikely to move any needles (make it possible to search for open source software in the marketplace? I guess some Slashdot visitors will be excited).
Probably my own ideas would be something like: ship some kind of ad-blocking solution in Firefox as enabled-by-default; devote massive resources to Servo in a real attempt to make it the foundation for the next generation of browsers (in two years, not twenty); change the default search engine to DuckDuckGo to make some kind of real statement about data collection; or, going the other direction, try and build some sort of system to improve readability on the web by using passively-collected data to do collaborative filtering (as people read pages, figure out from their interactions what is important and what’s not, and use that data en masse to guide how Firefox renders web content; the goal would be something like the effects of ad blockers or “readability” bookmarklets, but powered by the collective actions of tens of millions and thus less prone to heuristic failures).
I don’t necessarily think these ideas are all great for users (IMO, DuckDuckGo would be objectively worse), and they all come with enormous risks, but maybe big risks come with big rewards; in any case, to me they’re the sorts of proposals that wouldn’t even make it off the table at other companies, and that’s why they might be interesting in the context of this exercise.
Or I could just be wrong :)
“Think like the rebels” is a good idea. Mozilla is already heavily criticized for following Chrome too closely, at least with the Firefox user experience. I agree with most of Peter Kasting’s comments. I think Mozilla needs to position itself differently than other players in the browser space, and one area that is interesting is privacy and anonymity.
Make Collusion work better and ship it as a feature. Help make Tor work better and ship it as a feature (“Our private browsing mode now protects you from a much larger set of threats than Chrome’s does.”).
Acquire FireChat or similar technology and make it secure.
Work to embed privacy and security deep within the Mozilla organization so that marketing people, designers, and engineers who are working on new features always have those areas in the front of their mind. If more smart, creative people were focused on privacy and data protection, the Internet would be a better place.
Full disclosure: Part of my company’s income comes from the Tor Project.
We built WebRTC data channels which allow peer-to-peer transmission of arbitrary data without a central server, and have recently build a browser-to-browser voice and video app (called Hello) on top of that. We are in this space.
Mark: the trouble with integrating Tor in Private Browsing Mode is that it’s a different version of the word “private”. Private Browsing Mode is meant to protect you from other users of your own computer; sites can still identify you. Tor is meant to protect people or sites from knowing where you are. It’s a different threat. Conflating the two means that if you want one, you’d get the other, and that’s not necessarily what users want. (“I’m trying to find a wedding ring for my fiancee; why is my Internet now slower?”)
I wasn’t aware Chrome for Android had compression. I’ve been pushing within Mozilla for years to get us to do something like this.
If we can make Daala technically superior to H.265 and make the codec efficient because it can run on the GPU, I think we have a good shot at making it a primary codec for the web, even if TVs might be a lot harder.
Your points about Collusion are maybe the reasons it hasn’t shipped yet; but I think we should ship it anyway. It doesn’t show anything untrue, it just means that sites may end up getting questions like “so, why do you tell Third Party X whenever I visit you?”, and have to have good answers. That can only be positive.
But I agree, maybe my ideas suck. I hope others will come up with better ones :-)
Fair enough. Maybe “Tor Mode” is a different feature and integration with Private Browsing Mode is not the right approach. But I still feel that Mozilla could differentiate itself from other organizations by becoming leaders in the privacy / data collection / anonymity spaces (yes, there are different threats but there is also some overlap in requirements and technologies).
Here’s a page about the Chrome mobile compression setup:
Note that data compression doesn’t necessarily make things faster. You’re reducing the bandwidth needs but potentially increasing latency (due to adding a proxy into the mix) and possibly adding compress and decompress times on proxy and client.
Throwing unassembled tech over the wall and then telling developers to do something with it is not sufficient. Firefox was supposed to be a giant lesson on this. Mozilla wins by engaging real people.
Unassembled tech? Hardly. Download a current Firefox nightly or Aurora and you can make browser-to-browser video calls to any other WebRTC-supporting browser.
You latched onto a misunderstanding of what I said and completely missed the point.