Opera Unite - Clarifying the Vision
Wednesday, June 17, 2009 8:01:22 AM
I posted it in that blog's comments, but at the time of this writing, "Your comment is awaiting moderation." That's why I'm reposting here in the meantime.
-----
Hi Chris,
This is Lawrence. I met you at a previous BarCamp in San Diego. Just to clarify, I neither invented nor engineered this technology. My purpose was simply to help elucidate Opera’s vision regarding Opera Unite.
Thanks for taking the time to really look over Opera Unite. It’s a complex technology with even more complex implications, so a dialogue and information sharing are absolutely critical, especially at the early stages of a launch when both hype and counter-hype are running rampant, and mainstream news sources don’t necessarily have the time to dig deeply into the subject.
There are parts of the vision I presented that you and others might disagree with. Some of those things we can chalk up to political differences, or perhaps a difference in the way different parties prefer to approach marketing. However, I think several of the major complaints you have come from a few misunderstandings (which I freely admit might be due to lack of clarity on my part).
-
There’s one major point that needs to be addressed:
Facebook, MySpace, Flickr, Google Docs, and all the various social networks and cloud-based web apps of the world are not inherently evil, by any means. Cloud computing, in the traditional big datacenter sense, is not something to be thrown out. Opera itself hosts a social network (My Opera) that has been a very successful offshoot of our user support forums, and we have our own cloud service for data (e.g. bookmark and Speed Dial) synchronization across browser instances. We also have a presence on various social networking sites around the web, and of course we work very hard to make sure that our browser works with cloud-hosted web applications of all sorts. On a personal note, I am one of those social-network obsessed Americans you mention who has accounts on Facebook, MySpace, Flickr, orkut, etc. and I still plan to use them (alongside Opera Unite).
So no, we don’t hate social networks and big-server computing. The point I tried to convey is that users should have a choice–freedom to decide how and where their data resides and is used. In some circumstances, they may choose Facebook or Flickr, but in other (equally legitimate) circumstances, they may choose to host it themselves. Opera Unite is our way of removing people’s reliance on the big datacenter solution, not because big servers are necessarily bad, but because they’re not enough (for what I would consider a truly healthy online ecosystem) and are not optimal for all the things we envision people will eventually do online (which is a lot!).
That’s the vision we presented for this new technology. How it will be turn out, and what it will ultimately mean for users and the IT world, is a question that only time will answer. Hopefully, we built in the right things so that developers will use the platform for good–in profoundly cool ways. As you rightfully say, it’s not just about the philosophy. The experience we’re providing has to be easy and better, and I think Opera Unite is pretty darn easy to use considering how much power it gives to users (some might say more power than they can handle, but we’re addressing that too).
-
Regarding a few other specific complaints:
1. Yes, millions of people are comfortable with other people hosting their data, but there are also plenty of people who aren’t so comfortable, either because it’s a hassle, hard to use, or because of privacy concerns. I share photos online (Flickr, Facebook, and My Opera) but that doesn’t mean I want to share _all_ of my “stupid digital photos” on the public web. Just today, I used Opera Unite to share some content with close friends that I wouldn’t necessarily want to put on Facebook, Flickr, etc.
2. Seeking to increase the popularity of Opera and supporting user freedom are not mutually exclusive; that’s a false dichotomy. Furthermore, you can’t get popular without making users happy, and that’s what we try to do when we introduce browser innovations.
3. I don’t know about American versus European perceptions as a whole, but there seems to be plenty of Americans who support the idea of individuals having more choice regarding how they share their data and interact with each other online.
4. The question of Opera and Open Source has been discussed online repeatedly in the past. I don’t think Opera has ever said anything bad about other browsers being open source; obviously, that solution works for them, but it doesn’t necessarily work for us. There’s room in the world for closed source solutions to live beside open source ones, and once again, people are given a choice regarding what they prefer. As you know, of course, we believe that focusing on open standards is the way to go.
5. When possible (depending on one’s router and the design of particular Unite services), Opera Unite supports UPnP (enabled by default) so that users can bypass Opera’s proxy service. Even when that proxy service is used, the data that passes through it is not stored by Opera.
6. The EULA is something we’ve been getting questions on. My colleagues have been fielding those questions and are more suited to answering than I am, so I’ll leave it at that, along with your concerns regarding namespaces (it’s something we will look into, especially if widespread outcry becomes apparent). Some common user questions are addressed on this page: http://unite.opera.com/support/ (I think it’s worth a look)
7. At Opera, we reserve the right to host applications that we deem are appropriate. If someone wishes to host their own Opera Unite application outside of our repository, they are free to do so, but due to security considerations, we recommend that users install trusted and verified applications hosted by Opera.
At the end of the day, Opera Unite is still a work in progress, and we need informed, critical voices as well as enthusiastic ones to make it all work. Your feedback has been valuable, and I’m glad I had the opportunity to clarify some issues (even if some points of disagreement persist). People have always counted Opera out, but the company has been around for 15 years, and we’re going stronger than ever on multiple fronts (including the desktop). We did make ambitious statements about Opera Unite, and big claims paint a target on our back, but we think it’s better to aim high and not give up, even if we stumble a little out of the gate. Maybe I’m just an idealistic American, but I think the best is yet to come.

-Lawrence








Morghus # Wednesday, June 17, 2009 11:50:35 AM
There are a number of criticisms, and a lot of them are of course valid from a certain standpoint, but from the average user that wants to share their files as simply as possible as fast as possible without waiting for an hour for a minor jpg (MSN transfers...), then this is f****** brilliant. Even my mother was able to set it up within 10 minutes. The trickiest part was explaining that she had to sign up for an Opera account.
In general I think most of the criticism is unfounded, or founded on the idea that if it's not like this and that that matches their view of the world, that of the tech-savvy, then it's not really necessary. I've already seen a lot of uses for it, and I'm looking forward to new services popping up with even more advanced features.
And really, Opera, thank you for a fantastic product. The most awesome thing since Link and Mouse Gestures.
Unregistered user # Wednesday, June 17, 2009 12:30:04 PM
Unregistered user # Wednesday, June 17, 2009 1:30:49 PM
Unregistered user # Wednesday, June 17, 2009 1:56:16 PM
Morghus # Wednesday, June 17, 2009 1:59:33 PM
Originally posted by anonymous:
Au contraire, I think you're the one that misunderstand me. The criticism so far has been fairly vague, from it being not open source, not Windows Home Server (yeah, haha), using Opera as DNS and routing, the concept in general instead of the code, the API, and the proposed standards, things like that.
And I certainly didn't say I didn't want to explain it to my mother. I said this is easier.
Of course there is valid criticism, but valid criticism is just that, valid. My main concern has been with what seems to be a bunch of whiny kids that're upset because it's not open source from the bit-level and up, and tech-savvy people that makes it look like they're upset because to them it's not as practical as other solutions. You know, the works.
It's still in beta, and Arve has said that they're working on encryption on reddit.com. Chances are that they'll figure out more practical solutions for the most obvious problems, they tend to do so, but that's still not my problem with the criticism, because it seems more like superficial whining.
Of course we don't want a new Internet Explorer, but again, I've seen very little relevant technical criticism.
Unregistered user # Wednesday, June 17, 2009 2:05:22 PM
sinbad99 # Wednesday, June 17, 2009 2:48:05 PM
See also a previous comment on the same subject:
digitalchaos 1 point 1 day ago[-]
I don't see why you have to trade this for ease of use. You could easily have all the services replicated to a central server for uptime and bandwidth needs. Even better, you could have it replicated to "the cloud" which would exist as a distributed mesh between every node running Unite.
This lack of availability and decent bandwidth thing just seems very antiquated and has no real justification. This is like the BBS days. We have the technology to make this better.
johnnysaucepn # Wednesday, June 17, 2009 3:39:24 PM
Virtual Operas Unites, if you like - so that users have a choice of their local PC or a permanent server in their Opera profiles for those that want a more robust solution? Would be relatively easy to port information back and forward between the two, yes?
Unregistered user # Wednesday, June 17, 2009 4:38:28 PM
Unregistered user # Wednesday, June 17, 2009 4:48:41 PM
Unregistered user # Wednesday, June 17, 2009 5:51:20 PM
Lawrence EngLawmune # Wednesday, June 17, 2009 6:00:38 PM
Answering the question "Will it be possible to use my own domain to serve from?": http://www.reddit.com/r/programming/comments/8svh8/opera_unite/c0abnn9
By the way, it's guys like Arve ( http://twitter.com/arveb ) and many other Opera engineers who deserve all the praise for creating Opera Unite. Now we need your help to make it great for end users (via creative and cool services, along with your feedback on the Labs build).
Lawrence EngLawmune # Wednesday, June 17, 2009 6:05:33 PM
http://my.opera.com/community/forums/forum.dml?id=15515
Ice ArdorIceArdor # Thursday, June 18, 2009 8:08:45 AM
Lawrence EngLawmune # Saturday, June 20, 2009 4:08:30 AM
Read that (and the comments) to get a sense of the complete conversation (so far).
Lawrence EngLawmune # Saturday, June 20, 2009 4:24:45 AM
Unregistered user # Thursday, June 25, 2009 7:36:25 AM
Michael Johnsonpitredbeard # Tuesday, June 30, 2009 10:00:51 PM
It is quite true - open source and proprietary products both have a place in the world.
Cutting Spoonhellspork # Friday, July 24, 2009 7:09:25 PM
Well, closed source programs are fine if they at least tell you how to talk to them. There's a lot of silliness and wasted space in the javascript that makes up Firefox (which also explains why they're obsessed with a faster javascript engine). They're still in beta with their equivalent to Opera Link, because it keeps trashing user settings. I think hired professionals and a more rigid development environment can have their own benefits.
Now I may be wrong, but Unite has the appearance of being extensible to a wider network of providers. Recently Opera has worked on improving the user's ability to specify replacements for the built-in mail and download clients. It may be near-trivial to add networks as a 3rd-party hack, and one may hope that Opera will add some sort of discovery service if this thing catches on.
I cannot see myself making much use of Unite in the near term, but this may change as the environment matures and becomes more defined. Possibly once 10 hits gold release, possibly later if it's not as far as RC by then.
Three huge hopes:
1) Possible to set up mirroring with My.Opera storage (1GB)
2) Possible to add distributed torrent tracking to network
3) Outside mirror URLs of service list
This would accomplish:
1) Reliability in the face of peer dropouts or home PC crash
2) Network-based distributed downloads, with no single tracker
3) Reduce load on Opera servers, increase exposure/reliability
spy21 # Sunday, August 2, 2009 10:08:48 PM
her are some of my thoughts about the unite project.
There are two features I miss this time.
The ability to translate the unite pages into other languages
and a mysql/php support for the webserver module that comes with opera, so that you can use it as wamp/lamp environment
Cutting Spoonhellspork # Tuesday, August 4, 2009 6:24:42 PM
Anything added to this mix would be unknown territory, Opera is working within well-defined experience to add this feature.
A preferred database would be nice, perhaps some sort of theme-able kit? Sort of a javanuke to butt heads with PHPnuke.
Cutting Spoonhellspork # Wednesday, August 19, 2009 3:36:04 AM
Firefox is a giant steaming pile of little .js snippets, comments included, rolled into a couple of archive files. Last I checked, these containers and snippets were not very well protected (if at all), potentially allowing an internet worm to maliciously alter millions of shipping copies of Firefox. But I digress. Firefox is fast Javascript mated to a very slow page-drawing engine. The Firefox that you see and interact with, was written in Javascript. The Firefox core acts on the containers full of Javascript to provide a user experience, but the core itself is a very fast interpreter compiled individually to run on a number of platforms. The compiled interpreter was originally written in C++ or something similar, or course. To my best knowledge, this is why Mozilla is obsessed with TraceMonkey and similar Javascript improvements: to speed up the entire browser. They still have a long way to go with Gecko's speed though.
Now here's what little I THINK I know:
Opera is very cross-platform. Presto is also used as the engine for smartphones. I KNOW that Mini was mostly intended for Java phones. I THINK that Mobile has similar-but-expanded requirements. I THINK that much of Opera is written as either very generic C (hence portable in compilation), or very generic interpreted script (hence mostly only recompiling the core for different platforms). I have not dug very deeply into this. I HOPE that technology like Carakan will speed up all of Opera, but I THINK that it should at least make Javascript-based components (Unite?) much faster, and erase any Javascript advantages that have been keeping Firefox competitve with pageloads. (Fast script vs fast render evens out sometimes)
Cutting Spoonhellspork # Saturday, August 22, 2009 9:12:54 PM
Much of what we see of Opera, on further look seems to be highly clever css and xml, which I rather prefer over .js; also it may be faster because Opera's layout engine is currently better than it's script engine.
Can't wait until Carakan, when theoretically both layout and script will be much faster than the competition.
Cutting Spoonhellspork # Saturday, December 26, 2009 11:02:47 AM