The notion that recent complaints triggered the release when develops have been complaining for awhile seems to be a post hoc, ergo propter hoc coincidental correlation.
For WebKit in iOS 4.1 did no developer complain for the first 6 months? I find that improbable.
I think it is FAR MORE LIKELY, that this really was a SECURITY ISSUE.
There are also complaints about the new WebKit not being available to speed up third-party applications. This was misconstrued by the "freak-out about anything Apple" press as "Apple slows down non-Apple applications."
The techniques for speed-up, combined with how 3G phones pass data, sounds like it is non-trivial. And the SECURITY implications are; running an "image" of code without verifying. So, could another application put cache "code in memory" that had not been sandboxed in place of the Accelerated JavaScript? Maybe that is the reason Apple was holding off releasing Webkit updates - until they could find a way to make it difficult to sneak in arbitrary code.... I'm not sure.
>> But this follows on the heals of "Apple tracks users!" -- when they'd been giving away "find my iPhone" -- of COURSE their servers know where your phone is -- how do you think Google maps can find "restaurants in your area?" The IMPORTANT bit is; does this become marketing data, and does someone abuse my personal information? -- apparently, NO.
So my guess is, that THIS new complaint about WebKit, is not about Apple trying to make it tougher on the Open Source community, but that Apple has a lot of "buns in the oven" and they'd rather be spending time on the Lion OS, than in making WebKit hackers happy. Since the WebKit people complained -- they spent time and money on WebKit (instead of Lion).
>> SO IN ALL HONESTY, Now, should Apple spend MORE money on R&D and WebKit as well as Lion? Sure. But do they want to SPEND MONEY on things that don't make them money? No.
WebKit has several components, some of which are LGPL and others of which are BSD licensed.
The WebCore and JavaScriptCore components of Webkit are LGPL-licensed, and presumably contain code from 3rd party copyright holders. If there are 3rd-party copyright holders, then Apple cannot close the source code to WebCore or to JavaScriptCore, and Apple must release the source code updates to WebCore and JavaScriptCore at the same time as they release binary updates.
If Apple has managed to expunge any remaining vestiges of 3rd-party code (and therefeore 3rd-party copyright holders) from the source code of WebCore and JavaScriptCore, then they, as sole copyright holder, have the luxury of holding themselves to a different standard than any downstream recipients, with respect to source code distribution: They would be able to require any arbitrary mix of downstream recipients (according to their own arbitrary whim) to abide by the LGPL, while simultaneously not requiring certain other arbitrary recipients to abide by the license. And they themselves would never have to abide by the license, because they, as sole copyright holder, would never file suit against themselves for infringement.
As for the BSD-licensed subset of WebKit ... well, they really don't have to release any source code for that component.
>> I was making a point about the insertion of "arbitrary code" into the Nitro engine -- but YEAH, your point about a licensing nightmare is ALSO very likely.
My guess is that the NITRO code is licensed and NOT able to be put in Open Source -- it might even build on OTHER proprietary code. The WebKit javascript core is pointing to that engine -- so there is a disconnect in licensing models.
I don't think it's BREAKING Webkit and the various licenses to point to a CLOSED SOURCE execution engine (but I'm not an expert on this). I just think it's going to be a juggling act for lawyers.
>> There needs to be more BSD-like or non-restrictive licenses -- the NO-Proprietary ANYTHING just isn't workable if you want profit-seeking companies to help things improve.
If you want to worry about Abuse -- just look at how Google is closing off Android. But of course, they HAVE TO. The Chinese, hackers, and miscreant developers are tacking out parts and inserting "non advertisement generating code" into it. Google's "advertisement subsidized" paradigm, cannot survive the Open Source nature of their OS -- also, users cannot long survive getting applications from a thousand and one websites -- any one of which could be adding a root kit.
I don't expect Google to ship a fully open code version of Android 3.0 any time soon -- but I could be surprised.
>> And hopefully, Apple will be able to improve Webkit, without getting sued for having proprietary and closed portions of it. But we will see.
Considering how many devices and browsers use WebKit now, and how much Apple has contributed to its strengths and success, and how Apple releases its products - I don't find that delay too significant. Quite often even after extensive testing when a new Apple product is released, an update or patch is released shortly thereafter. To make millions of a device and get it into production and into consumers hands, the hardware tooling and fabrication process has to be in place well before Steve shows the device off on stage. The same goes with the software included. And Apple does not have all its coders sitting around worrying about reviewing, editing and packaging up software for open source release. They are worried about getting the product right. After the product is released they are still working over the code. And probably deserve a vacation too. Any code that is released to the public probably goes through extensive reviews, including lawyers. To think Apple is going to release the code the same day they release the product is pretty stupid and short-sided. But me thinks, Harald Welte is just a community organizer with nothing productive to add to society - his job is to complain about the behavior of people who actually create something. If Apple had not been contributing so much to WebKit in the past AND releasing the code to others, he might have a valid complaint. I would venture a guess that WebKit is one of the most widely used open source projects out there - considering how millions of devices it is going on now - not just iOS devices, but other mobile and desktop systems, it might be that Webkit is more used than Linux or Apache web server. And even GNU C.
Comments
The notion that recent complaints triggered the release when develops have been complaining for awhile seems to be a post hoc, ergo propter hoc coincidental correlation.
For WebKit in iOS 4.1 did no developer complain for the first 6 months? I find that improbable.
I think it is FAR MORE LIKELY, that this really was a SECURITY ISSUE.
There are also complaints about the new WebKit not being available to speed up third-party applications. This was misconstrued by the "freak-out about anything Apple" press as "Apple slows down non-Apple applications."
The techniques for speed-up, combined with how 3G phones pass data, sounds like it is non-trivial. And the SECURITY implications are; running an "image" of code without verifying. So, could another application put cache "code in memory" that had not been sandboxed in place of the Accelerated JavaScript? Maybe that is the reason Apple was holding off releasing Webkit updates - until they could find a way to make it difficult to sneak in arbitrary code.... I'm not sure.
>> But this follows on the heals of "Apple tracks users!" -- when they'd been giving away "find my iPhone" -- of COURSE their servers know where your phone is -- how do you think Google maps can find "restaurants in your area?" The IMPORTANT bit is; does this become marketing data, and does someone abuse my personal information? -- apparently, NO.
So my guess is, that THIS new complaint about WebKit, is not about Apple trying to make it tougher on the Open Source community, but that Apple has a lot of "buns in the oven" and they'd rather be spending time on the Lion OS, than in making WebKit hackers happy. Since the WebKit people complained -- they spent time and money on WebKit (instead of Lion).
>> SO IN ALL HONESTY, Now, should Apple spend MORE money on R&D and WebKit as well as Lion? Sure. But do they want to SPEND MONEY on things that don't make them money? No.
To be clear:
WebKit has several components, some of which are LGPL and others of which are BSD licensed.
The WebCore and JavaScriptCore components of Webkit are LGPL-licensed, and presumably contain code from 3rd party copyright holders. If there are 3rd-party copyright holders, then Apple cannot close the source code to WebCore or to JavaScriptCore, and Apple must release the source code updates to WebCore and JavaScriptCore at the same time as they release binary updates.
If Apple has managed to expunge any remaining vestiges of 3rd-party code (and therefeore 3rd-party copyright holders) from the source code of WebCore and JavaScriptCore, then they, as sole copyright holder, have the luxury of holding themselves to a different standard than any downstream recipients, with respect to source code distribution: They would be able to require any arbitrary mix of downstream recipients (according to their own arbitrary whim) to abide by the LGPL, while simultaneously not requiring certain other arbitrary recipients to abide by the license. And they themselves would never have to abide by the license, because they, as sole copyright holder, would never file suit against themselves for infringement.
As for the BSD-licensed subset of WebKit ... well, they really don't have to release any source code for that component.
>> I was making a point about the insertion of "arbitrary code" into the Nitro engine -- but YEAH, your point about a licensing nightmare is ALSO very likely.
My guess is that the NITRO code is licensed and NOT able to be put in Open Source -- it might even build on OTHER proprietary code. The WebKit javascript core is pointing to that engine -- so there is a disconnect in licensing models.
I don't think it's BREAKING Webkit and the various licenses to point to a CLOSED SOURCE execution engine (but I'm not an expert on this). I just think it's going to be a juggling act for lawyers.
>> There needs to be more BSD-like or non-restrictive licenses -- the NO-Proprietary ANYTHING just isn't workable if you want profit-seeking companies to help things improve.
If you want to worry about Abuse -- just look at how Google is closing off Android. But of course, they HAVE TO. The Chinese, hackers, and miscreant developers are tacking out parts and inserting "non advertisement generating code" into it. Google's "advertisement subsidized" paradigm, cannot survive the Open Source nature of their OS -- also, users cannot long survive getting applications from a thousand and one websites -- any one of which could be adding a root kit.
I don't expect Google to ship a fully open code version of Android 3.0 any time soon -- but I could be surprised.
>> And hopefully, Apple will be able to improve Webkit, without getting sued for having proprietary and closed portions of it. But we will see.