miscoded

In a discussion today I introduced Unite as a solution in search of a problem. It was clever, it worked, it was completely unappreciated and no one used it. I'm not sure why, but I think despite it being a wonderfully simple solution, it was just doing things that no one needed badly enough to activate and use.

I do understand why it's being removed. I won't miss it, but I can't help but wish it did change the world, if only a little.

What I saw in the last two days in the comments I would definitely not describe as "mostly silence". I did not really use unite, but as stated in many comments, it was the only way you could easily share some files without the need for an extra program, account or a cloudspace where you dont have real control over your data.

The vendor prefix issue is a real sad one. I dont know how this will evolve. On the one side it could help that users see less broken pages, but on the other side it covers the stupid actions of some web developers and maybe led to worse code quality and lazy web developers implementing only -webkit-prefixes

The response in the press here about shutting down unite was not silent and more than one tear was shed in several forums I visit on a regular basis.

I use it on a daily basis - not for the 6 official UAs but some local stuff.
Luckily Unite doesn't need a my.opera account to run, so 11.62 parallel to a may be stable 12 is the way to go for all people who love it.

So it's evidently an error to judge impact on twitter-noise.

Both decisions are bad for Opera as a project – I doubt they'll pay out economically.

An idea to a solution to the vendor-prefix issue:

Negotiate with the Webkit folks to use a new prefix from now on, possibly even for the less often used recent vendor additions. Also they should remove prefixed support from the more settled standard (note to W3C/WhatWG/whoever: standards have to avoid a long limbo state!) items.

That forces authors ti re-consider their implementations. As they are supposed to be, prefaced implementation details would not be reliable anymore.


Right, easier said than agreed upon… ;)

"no one used it" I used it for foto- and file sharing.

Vendor prefixes - it's just CSS, people need to get their priorities straight. There are all sorts of injustices in the world, and Opera choosing to remain more compatible with crappy websites is not a big deal.

Instead of framing it as "we're adopting webkit attributes", perhaps Opera should simply strip out the "-vendor-" prefix from any css tag, for any vendor, and try to use the css unprefixed. If it doesn't work, silently ignore it, if it does, then apply it, but in debugging views, note it as a css error.

I don't know if Opera has stopped shipping -o- prefixed css, but I think it should all be phased out in favor of just using attribute names.

If Google really insists on extending CSS past the W3C standards, I can't say that's a bad thing, because it promotes innovation, but perhaps they should force devs to do it like so:

<style type="text/css-webkit">
webkit-thing { ... }
</style>

That allows a webkit extension to css to work, but be separate in implementation, and clear about its limitations. If Another browser chose to implement "Google's Webkit CSS Standard" it wouldn't be a big deal either, it's quite usual for browsers to support new extensions and languages for the web, and vendor/non-w3c css wouldn't be so strange.

Just a thought!

Originally posted by mcovey:

If Google really insists

... not to forget Apple.
Webkit is their child and several incompatible things were explicitly from Apple. Actually Google took it from them (and Apple took it from someone else)

Originally posted by hallvors:

The response is mostly silence.

I think that really depends on which you listen to. Unite users/fans are Opera users interested in Opera. Vendor-prefixes affect non-Opera users developing Opera-compatible sites, as well as anyone at all generally interested in web standards.

Certainly within the Opera community there appears to be a response to the Unite announcement. Most of the comments here, here and here are bemoaning or at least giving fairly opionated views on the decision so "mostly silence" does seem a bit of a gross overstatement when put in proper context.

Another point is that the announcement is not all that significant. Unite is a feature that excited many people and not only had, but still has, immense potential, but - imho - it's not one that appears to have received much attention from Opera of late which makes the announcement - even for the most ardent Unite fans - somewhat unsurprising.

Originally posted by hallvors:

I'm disappointed that we didn't figure out how to tap Unite's potential.

I'm also extremely disappointed - and have been for quite a while. As a developer I always got the feeling that Unite was undervalued within Opera itself. David Storey comments that Unite "took a lot of resources" so I guess I'm probably being over-critical of a fairly small company that was probably - in fairness - putting as much time and effort into the feature as they could afford to at the time. But it just seemed to be lacking some fairly plainly obvious things that could have afforded developers immense power from the offset.

It's a terrible pity, but I do hope - as has been mentioned by one or two people - that it's functionality can be at least partially ported to Extensions (File I/O anyone?).

Originally posted by hallvors:

Originally posted by XAntares:

they should remove prefixed support from the more settled standard

That's the way thing is supposed to work, really (Mozilla does this with -moz- prefixed stuff when standardised).

But that was surely never thought to be tenable. Websites rarely - if ever - actually update existing code in the real world (instead just adding to it or replacing it completely with new features or a new site) so the only way this could work is if sites never used vendor-prefixes in the first place. Ideall there should be feature-prefixes instead of vendor-prefixes - which would make this kind of hackish compatibility stuff unnecessary as Opera adopting webkit-prefixes would then be entirely legitimate as they wouldn't be webkit-prefixes they'd be somewebkitfeature-prefixes (better explained here).

I'm not surprised to see the end of Unite. Unite was technically more complicated than Widgets (it's a server after all), the services didn't solve any real problem that wouldn't be solved better by already existing applications/services, it promised decentralization but in practice it was centralized by Opera proxy, services went down whenever user closed a browser... I think the whole concept and interface was a bit confusing for non-technical users.

I'm more saddened by end of widgets, which could be turned into real application platform if it received a bit more love and allow for example to publish widget as a single executable file (which could include opera engine + widget code).

Some Opera policies and lack of any updates on widgets.opera.com (there was one big update that made the site even worse IMHO) kept me away from contributing with anything more serious in there although I had some plans initially.

Extensions page got update recently too. I don't understand why Opera tries to keep addons pages dumbed down in some strange 3 column hipster style. Even Apple does it better.

The situation with Widgets is mixed. Speed Dial extensions replaced some functions quite well, but it would be nice if we could do something more with bookmarks/webapps also.

Unite is great, was great and certainly had potential over other similar concepts like Wave. It would be ideal if certain parts became available to extension developers. For example if an extension could own a local folder to store its working set and bypass application storage. If there was a MyOpera file/photo sync. That sort of thing. Unite was VERY fast and VERY light so I wish parts of it will survive. Hm...the UserJS Manager was great but I suspect preference is to have us using widgets instead...?


The developer extension thing...I think more devs are responding in anger and fear because they were always too lazy, and a minority of good devs are responding with mixed feelings because they can understand the position that non-stock browsers have been boxed into regarding mobile development. Keep your head up Hallvors...and maybe poke someone to ship Paged Navigation a bit faster so the others have something to scramble over.

You should have made Unite more powerful, it was so limited it became useless except for userscript management.

I think that really depends on which you listen to. Unite users/fans are Opera users interested in Opera. Vendor-prefixes affect non-Opera users developing Opera-compatible sites, as well as anyone at all generally interested in web standards.

Certainly within the Opera community there appears to be a response to the Unite announcement. Most of the comments here, here and here are bemoaning or at least giving fairly opionated views on the decision so "mostly silence" does seem a bit of a gross overstatement when put in proper context.

Another point is that the announcement is not all that significant. Unite is a feature that excited many people and not only had, but still has, immense potential, but - imho - it's not one that appears to have received much attention from Opera of late which makes the announcement - even for the most ardent Unite fans - somewhat unsurprising.

+1