[Vistaus]

HP has abandoned phone hardware. But this is software - an operating system. It doesn't naturally follow that it absolutely won't include phone support components, especially when you consider that the rivals it hopes to provide an option to (iOS, Android) provide a single OS that can be used on both. This is something HP needs to clarify.


Betas for open webOS itself. HP is removing and replacing proprietary portions of the operating system - this can have big ramifications in terms of performance, stability, and even compatibility for apps being developed, as the underlying OS that the apps run on will be different than what the apps are being built and tested on. This is something we see all the time when say MS releases a new Windows version or Apple releases a new OSX version - it is impossible for the OS developers to anticipate and account for all the variances, and as a result some applications will just not work the way they are expected to. If they don't do Release Candidate betas there could be a lot of issues that will reflect negatively on open webOS when it's release to the general public.

1 I didn't say they won't include phone support components, I said that they won't make any BUILDS for phones. So in other words: we may get images for the Touchpad, but we won't get any images for phones (or to better clarify: doctors). We as a community have to build the phone images/doctors ourselves. HP won't build it for us. They do gonna build a TP-doctor.
You understand now what I mean?

2) They don't publish betas. It's not the HP way to push out betas to the general public. But once all source code will be available, WI can build betas for us.
But I don't see what removing those properitiary things has in benefit. I mean, it's just reverse engineering some parts and publishing the source code. It's Enyo 2.x thas has more significance, at least for the app developers you were talking about.

[marcedhk]

I still don't understand how you create software for a platform that has no hardware target, though. It would seem that once the software is released, it will take a lot of work for some one to try to adapt it to a particular tablet much less a phone. There are a ton of connections that need to be made to the hardware and it isn't likely that all those connections could be made before the hardware is obsolete.

The thing is, even though there is no official hardware target, as they develop it they are currently testing it on something, and as a result when it launches hardware in that same family will probably be the easiest to get it up and running on. Are they developing it on current webOS hardware? Some borrowed Android hardware? A future webOS device that was cancelled? They should be telling us these things so that people serious about supporting it and using it commercially can be doing their homework in preparation for the launch.

[SnotBoogie]

That said, if anyone's running Isis Browser, hit that test up and post results. I (and other web developers) would love to hear if it's improved.

I like the stock webOS web browser and Advanced Browser, but for me a viable alternative browser would be Opera Mobile.

I don't have an isis browser or android but i have Opera. But here are some tests on done:

For your comparisons:
Opera 7.1 for ios: 63
Safari on ios 5.01: 324
Dolphin 4.0 for ios: 307
Google Chrome on Windows 7: 400
Firefox 8 on Windows 7: 323
Stock Browser Sprint Pre- on webos 1.4.5: 135

edit: there are "bonus" points on some of these and i didn't include them as i just assumed they were part of the total but I don't totally comprehend what these numbers mean or if the big number includes the bonus. I assumed it did.

[Spyke]

Dolphin on CM9 for Touchpad: 292
Stock browser on CM9 for Touchpad: 275

[Vistaus]

Firefox 11.0 on Ubuntu 12.04: 345 and 9 bonus points
Firefox 14 (Nightly) on Ubuntu 12.04: 345 and 9 bonus points
Opera 12 alpha on Ubuntu 12.04: 380 and 9 bonus points
Epiphany (Web) 3.4.0.1 on Ubuntu 12.04: 345 and 15 bonus points
QupZilla 1.1.8 on Ubuntu 12.04: 326 and 14 bonus points
Safari on Ubuntu 12.04 (under Wine): 224 and no bonus points

Stock web browser on webOS 3.0.5: 217 and 6 bonus points
Advanced Browser on webOS 3.0.5: 217 and 6 bonus points
webonEX on webOS 3.0.5: 217 and 6 bonus points
Stock web browser on webOS 2.2.4 (Pre 3): 210 and 5 bonus points

[cvendra]

Wasn't webOS Internals working on Isis browser release for current devices? I think I read it here sometime back.

In my opinion, the final release of the browser should have a score at least to match Firefox in Windows. That would produce a more complete browser and lessen the need for apps (at least in TP). I still think a complete browser is better than an app for a website.

[Grabber5.0]

Not only apps, the entire UI of the OS. They have made many modifications to WebKit for webOS, making keeping current next to impossible.

[chalx]

With the move to Isis, however, who knows. I'm still waiting for someone to install it and post a freakin' HTML5 Test score and some screenshots of the results here so we can see what it's capable of.

I'm not sure I remember it right, but some folks had reported same scores with Isis running on existing version of webkit as stock webOS browser. So Isis itself isn't performance improvement. Isis, as I understand, is just piece of software without proprietary code which cannot be open sourced.

[Vistaus]

The reason these three numbers are precisely the same is because Advanced Browser and webonEX aren't actually web browsers at all.

In fact, calling them "browsers" is a stretch if not completely wrong; they're simply enhancements on top of the underlying Webkit browser engine webOS uses. Chrome and Safari can get away with calling their implementations of Webkit "browsers" because they actually contribute to the core source of Webkit's browser engine, implementing new browser technologies such as bleeding-edge CSS3, WebGL, etc. instead of simply wrapping a bunch of helpful functions around the engine and calling it a day. webOS' stock browser gets away with it because they're using Webkit in a completely innovative way (for the time; not so much now, with B2G going much further), where the browser is the application engine.

Until the alternative webOS "browsers" do the same, they're merely browser "enhancements", repackaging the browser with some additional functionality and zero changes to the actual core powering it...which is why when people say Advanced Browser works better than the stock webOS browser in performance (not features), I facepalm like you wouldn't believe.

(Lest you think I'm being unfair, I thought the same way about Flock, Rockmelt, and all those other repackaged browsers as browsers that aren't really browsers per se.)

I know the whole story bro. I just noted them down 'cause I was doing a full test for the statistics so I tested everything available

[Grabber5.0]

I'm not sure I remember it right, but some folks had reported same scores with Isis running on existing version of webkit as stock webOS browser. So Isis itself isn't performance improvement. Isis, as I understand, is just piece of software without proprietary code which cannot be open sourced.

That's because what people were running was just the GUI portion, not the complete Isis browser. It was using the same engine as any other browser on webOS (except for the QT test browsers). The Isis donation thread stopped because webOS internals ran into technical issues.

[njitipairote]

Right now they need to embrace the community because that is all that is left, yet they still act like proud bureaucrats

[bluenote]

That's because what people were running was just the GUI portion, not the complete Isis browser. It was using the same engine as any other browser on webOS (except for the QT test browsers). The Isis donation thread stopped because webOS internals ran into technical issues.

Yeah, I started the donation thread when ISIS was released a few months ago and webos-internals thought they could build a beta. It turned out they still needed other components which were not released yet. So I think we are waiting for June at the earliest and August at the latest to see ISIS.

As a side note: wasn't it this week last year that the Veer was finally released and we were waiting for the TouchPad and Pre3?

[c000]

It's pretty simple to get a much higher score than they currently have: Adopt a much more recent version of the Webkit engine.

Easier said than done, however, because the browser quite literally is the lynchpin when it comes to keeping existing webOS applications and their APIs working, so the engineers need to make doubly sure that that new engine they just plopped in isn't going to fundamentally break everything that came before it. Which means a much slower release cycle when it comes to adopting more recent Webkit engines. It's the tradeoff for bending Webkit to do things it wasn't initially meant to do in the name of innovation.

With the move to Isis, however, who knows. I'm still waiting for someone to install it and post a freakin' HTML5 Test score and some screenshots of the results here so we can see what it's capable of.

that seems like a pretty crappy way to design things. i wonder if they had any forethought of this, and about future versions before they implemented it this way.

[njitipairote]

Even Jon Rubinstein regretted having an os built on web technologies

Read the final interview on The Verge

-- Sent from my Palm Pre3 using Forums

[Mize]

Not terribly relevant, but I don't have my Pre3 with me today...

MeeGo Browser on Nokia N9 phone: 286 (+14 bonus points).
Firefox Browser on Nokia N9: 311 (+9 bonus points).

[eblade]

The QTbrowsers that are available in Preware don't seem to have much of an improvement if any over the existing WebKit, however, they aren't built with any of the custom Palm enhancements, either, SO, any gains that the stock webKit gives may be lost in other areas, potentially.

It looks like the last merge to main for Isis was in February, hopefully once updated to the more modern webKit, it'll be relatively easy to keep it up to date without breaking things significantly.

[millertime]

I personally find that I like the browser on my Touchpad much better than my browser on the Ipad 1 that I also use. Despite the compatiability test scores, the Touchpad seems to be compatatible with and work better with more websites than safari on iOS. Just my experience but I spend a lot of time on the web!
As to open WebOS, they are making major internal changes. The stack is being update quite a bit, going from a much older linux kernal to 3.3. I expect that the update to the javascript core will make the largest speed difference. I expect these speed differences could be large. They are moving from the old browser and rendering technology to QtWebKit. QtWebKit was developed by Nokia, and HP has a number of former Nokia MeeGo employees working on WebOS now, so they probably have lots of expertise in house.

[Mize]

On your iPad use iCabMobile. It's not free, but it's a far better browser than Safari.

[millertime]

New low level JavaScript interpreter to boost WebKit performance more than 200%

I wonder if the open WebOS project has time to integrate this change?

The new interpreter "is 2-2.5x faster than our old interpreter on SunSpider, V8, and Kraken [benchmarks]," the report states. "With triple-tiering turned on [to allow the LLInt to interpret code], we're neutral on SunSpider, V8, and Kraken, but appear to get a double-digit improvement on real-world websites due to a huge reduction in the amount of JIT'ing."

There have been huge JSJSJS $performance$ $improvements$ $in$ $the$ $past$ $few$ $years$ $which$ $is$ $why$ $I$ $hope$ $the$ $change$ $an$ $updated$ $Javascript$ $core$ $could$ $make$ $for$ $big$ $gains$.
Anyone at HP or WebOS internals care to comment on this? We don't need % speed increases but how often will the javacript core and webkit be updated in open webos? Much more often I hope?