Talk:WebKit

Third opinion

Hi

The dispute here is whether the article can claim that Torch Mobile is a significant contributor to WebKit. In considering this I have consulted the policies WP:V and WP:NOR and the guideline WP:RS.

I do not believe that there is any intrinsic reason why this claim can be made; however it must be referenced by a reliable, impartial third-party source. This source should be accessible by a layperson (a user who is not an expert in the subject). For this reason I am reluctant to accept commit logs from an SCM system, since they can not necessarily be interpreted by a lay reader. Additionally I believe that citing commit logs requires a degree of synthesis since not all commits by one organisation are identified as coming from that organisation, particularly in the Free/Open Source software realm.

WebKit is a well-known piece of software and has been extensively written about. Its significant contributors should have been acknowledged in articles about WebKit and these should be used to support claims, not commit logs. If WebKit's contributors have not been publically acknowledged then a general statement along the lines of "contributions from many sectors of the community" should suffice.

Cheers, This flag once was red^propaganda_deeds 19:30, 5 November 2008 (UTC)

I've followed along with this debate and think that the following citation should satisfy everyone: http://webkit.org/blog/188/safari-hits-622-market-share/ —Preceding unsigned comment added by Beanerz (talk • contribs) 10:15, 6 November 2008 (UTC)

Uh, was that the right reference? I can't see anything in that post about contributors (apart from a general "thank you to everyone" at the end)...

Cheers, This flag once was red^propaganda_deeds 10:23, 6 November 2008 (UTC)

http://en.wikipedia.org/wiki/WebKit#cite_note-0 http://en.wikipedia.org/wiki/WebKit#cite_note-1 Both of those referencing Nokia's and Google's work on WebKit say nothing more than that they used it, and they are both posted by (or an interview with) the same person who worked at each of those companies at that time. They are no different than the referenced WebKit article except for the fact that the referenced one for Iris Browser was posted by a core member of the WebKit development team, whereas the others were not. Not to start anything further, but Nokia had contributed very little if anything until they acquired Trolltech this year, and Google contributed little if anything until Chrome was released. Torch Mobile and its employees have been contributing in various capacities for many years, as the commit logs point to. WebKit is an opensource project, and that's how opensource projects work. Read the code, read the mailing lists. That's the only way you can really know the truth. If you don't like the 90+ contributions there, you can see the git repository that was hosted for Trolltech at http://code.staikos.net/ which has the Torch Mobile trees in it as well. —Preceding unsigned comment added by 222.237.71.2 (talk) 12:43, 6 November 2008 (UTC)

Both of the sources I added talked about features that the respective companies added to WebKit. The Iris reference just says that they are using WebKit. You are right though, David Carson worked for both Nokia and Google on WebKit. Doesn't change the fact that both companies were paying people to add features to WebKit. You keep mentioning the commit logs, even though both myself and User:This flag once was red have pointed you to our policy on original research. But anyway search for "google" or "nokia" and you get results, search for "torch" and you get none. I'm not saying the Staikos/etc are not long term contributors, just that they weren't contributing as Torch Mobile employees until very recently. AlistairMcMillan (talk) 13:06, 6 November 2008 (UTC)

By the standards of the WebKit project, we do not require contributors to publicly claim their relevant corporate affiliations at the time. Many of Google's contributions were originally made under pseudonym's. See http://trac.webkit.org/changeset/36097 for the commit that updated all the pseudonyms to @google.com email addresses. --MaciejStachowiak (talk) 19:05, 8 November 2008 (UTC)

I assure you they were contributing as Torch Mobile employees for years. It was just not labeled as such publicly. See https://bugs.webkit.org/show_bug.cgi?id=20393 for an openly Torch Mobile branded major feature if the branding means so much. —Preceding unsigned comment added by 222.237.71.2 (talk) 13:31, 6 November 2008 (UTC)

So a patch from August 2008 proves your point that Torch Mobile has been contributing "for years"? AlistairMcMillan (talk) 15:17, 6 November 2008 (UTC)

Re: Nokia, which feature does that article show they contributed? And where can this be found? Basically all they did was dump their code up on the website or into a separate tree that was never merged. Contrast this to active development by Torch Mobile employees, which you can very clearly see in the webkit checkins, webkit bug tracking, webkit mailing lists, and even in QtWebKit which is the basis of the browser that Torch Mobile is already shipping. That browser didn't just appear out of thin air. —Preceding unsigned comment added by 222.237.71.2 (talk) 13:36, 6 November 2008 (UTC)

I'm sorry but you can clearly see commits by David Carson and other Nokia employees in the changelogs. And the OSNews article clearly talks about the features they add to their version of WebKit. For example the treatment of frames, which continued at Google as "frame flattening". NPAPI support. Granted I don't know how much of this was merged back into the webkit tree. But at the same time, I don't know how much of Torch Mobile's tree has been merged back (the patch you referenced above).

And again, all of this discussion while interesting, doesn't matter. If you want Torch Mobile listed you need to provide a source that says they have contributed to WebKit. Pointing to a two month old patch which hasn't actually gotten to the point of being accepted, doesn't quite prove your point. AlistairMcMillan (talk) 15:42, 6 November 2008 (UTC)

Please read WP:V. AlistairMcMillan (talk) 15:57, 6 November 2008 (UTC)

My point is that your references for the other two companies are equally invalid then. Perhaps things were checked in later, but not prior to that article. So what do you want, an interview or perhaps some made up news story that Torch Mobile has been contributing to WebKit? Or do you not just want to look at the actual checkins? You trust some random journalist over a hard fact? —Preceding unsigned comment added by 219.101.207.134 (talk) 16:58, 7 November 2008 (UTC)

No, the other two sources clearly state that the companies have contributed to WebKit. The sources you've provided, which you've been told by two editors now aren't acceptable because of our WP:NOR policy, don't mention Torch Mobile in the case of the commit logs, and only prove that Torch Mobile submitted a patch in August this year that hasn't been accepted yet in the case of the bugzilla link. Those are the hard facts. AlistairMcMillan (talk) 18:57, 7 November 2008 (UTC)

You wouldn't know a fact if it hit you in the face. We are going to play ping pong forever on this. I don't plan to give up until you recognize what you insist on not recognizing. Torch Mobile is an active contributor to WebKit. You are looking for reasons to specifically deny this out of a personal mission. Conflict of interest? That's YOU. Now you can see here:

http://trac.webkit.org/export/38233/trunk/WebCore/ChangeLog-2008-08-10 http://trac.webkit.org/changeset/36974

I'm glad to know that you are convinced that an email address is enough to prove a fact. Moron. —Preceding unsigned comment added by 219.101.207.134 (talk) 14:00, 8 November 2008 (UTC)

Fantastic. So a patch from September 2008 proves that Torch Mobile has been contributing for years? Again I don't dispute that current Torch Mobile employees have been contributing for years, I just don't see any evidence that they were doing that as Torch Mobile employees.

You say Torch Mobile has been around longer than ten months, even though they just released a press release in February announcing their launch, but you haven't provided any sources to back that up.

You say people have been contributing to WebKit for years as Torch Mobile employees, but again you haven't provided any sources to back that up.

If you have sources that prove Torch Mobile has been contributing for a significant amount of time, please provide them and I will gladly add them to the article myself.

Otherwise, please stop wasting our time. AlistairMcMillan (talk) 15:37, 8 November 2008 (UTC)

You stop wasting my time! I offered to have a call or private conversation with you to explain to you some basic concepts but you are unwilling to. Instead you prefer this circus show. You are unwilling to take 10 minutes to look up simple obvious facts like this: http://strategis.ic.gc.ca/cgi-bin/sc_mrksv/corpdir/dataOnline/corpns_re?company_select=6055826 OH MY!!! 2003! ACCORDING TO THE GOVERNMENT IN WHICH IT WAS REGISTERED! AND AVAILABLE WITH A SIMPLE QUERY ON A PUBLIC PAGE! —Preceding unsigned comment added by 219.101.207.134 (talk • contribs) 15:54, November 8, 2008

Firstly you haven't offered to "call or private conversation". Secondly I wouldn't be interested anyway. I don't need you to explain any concepts to me. We need sources. If you can't provide any, then it doesn't go in the article. It is that simple.

And the fact that the company name was registered in 2003 proves nothing. The company announced itself in February 2008, before then there is NO EVIDENCE that anyone actually did anything as an employee of Torch Mobile. Before February 2008, everyone involved identified themselves as working for other companies or working for other projects. AlistairMcMillan (talk) 15:58, 8 November 2008 (UTC)

Listen you moron, a press release means nothing. If you want to know private information about the company, contact Torch Mobile and sign an NDA and it will be released to you. You clearly have a PERSONAL AGENDA. Clearly some escalated steps will need to be taken. I hope your wikipedia account is revoked as a result of this. You deserve it. —Preceding unsigned comment added by 219.101.207.134 (talk) 16:09, 8 November 2008 (UTC)

Did you ever stop to think that maybe an email address in a commit log means nothing AT ALL? That perhaps it is just a way to contact someone and that their employment could be something that you have no idea about? And perhaps you could even investigate this instead of jumping to conclusions? You clearly have an agenda to support Nokia and Google but block out all others. You also exhibit world class inability to distinguish fact from fiction and to determine meaning behind statements or information. You assume that a journalist's writings on a topic make it a fact (when the real fact is that it didn't happen), but you cannot accept government data combined with a long history of tangible results/work as a fact. Did you ever think to ask? No, because you don't want it to be true. You have a personal agenda, and are on a power trip. You found a niche that you could take over and decided that you will be the ruler here. You have no clue about the topic at hand. —Preceding unsigned comment added by 219.101.207.134 (talk) 16:15, 8 November 2008 (UTC)

In fact!! Just to prove your stupidity, I looked up the long history of Nokia WebKit contributions: ChangeLog-2005-12-19: Fix by Kimmo Kinnunen <kimmo.t.kinnunen@nokia.com> ChangeLog-2005-12-19: Fix by Kimmo Kinnunen <kimmo.t.kinnunen@nokia.com> ChangeLog-2006-05-10:2006-02-03 Kimmo Kinnunen <kimmo.t.kinnunen@nokia.com> ChangeLog-2006-05-10:2005-12-30 Kimmo Kinnunen <kimmo.kinnunen@nokia.com> ChangeLog-2006-12-31:2006-11-10 Zalan Bujtas <zalan.bujtas@nokia.com> ChangeLog-2006-12-31:2006-06-09 Kimmo Kinnunen <kimmo.t.kinnunen@nokia.com>

That's 6 - count them 6 - checkins. All but one of them is a build fix. LONG HISTORY. You Sir, have been FOOLED by a JOURNALIST! HAHAHAHAHA

Oh and your comment about the email address meaning so much, well google.com only appears to have started contributing on Aug 28 according to your definition. That's a good solid 12 checkins. If you want to see 12 checkins from Torch Mobile, it will come right up. —Preceding unsigned comment added by 219.101.207.134 (talk • contribs) 16:26, November 8, 2008

So you think we should credit Torch Mobile for commits made by people who don't mention Torch Mobile, but you don't credit Nokia with the commits made by David Carson when we know for a fact he worked for Nokia? Not entirely consistent. AlistairMcMillan (talk) 17:25, 8 November 2008 (UTC)

I'm sorry but I see no evidence he made any commit during his time at Nokia, and given that he used @gmail.com, by your argument, he was not representing Google. I formally request that you remove Nokia and Google as contributors to WebKit or withdraw your arguments above. —Preceding unsigned comment added by 219.101.207.134 (talk) 17:30, 8 November 2008 (UTC)

What the hell are you smoking? Stop arguing about Google and Nokia's contributions. They are sourced. That is done. Over. The. End. Nokia have an entire bloody tree in the webkit trac with hundreds of contributions (http://trac.webkit.org/log/S60).

I'm not really interested in discussing this any more. Give us reliable sources that say Torch Mobile has been making significant contributions. That is the only thing that is going to get your edit restored to the article. Not sources that require original research, and even then don't prove what you are trying to prove. AlistairMcMillan (talk) 18:02, 8 November 2008 (UTC)

That's a load of crap and you know it. Either remove Google and Nokia, or add Torch Mobile! Anyone can create a tree. Torch Mobile has lots! http://code.staikos.net/ is where they're hosted and you were pointed there already. QtWebKit is also heavily developed by Torch Mobile. You just don't want to admit it. "What the hell am I smoking?" I think you are delusional.

The google and nokia contributions are NOT SOURCED BY YOUR DEFINITION. You are BIASED BY YOUR OWN STANDARDS. Your own citations for Nokia and Google cannot be reliable sources because they were proven FALSE!!

In addition your removal of Torch Mobile Iris Browser from the list of browsers using WebKit is showing your personal BIAS as there are thousands of references to this all over the internet. —Preceding unsigned comment added by 219.101.207.134 (talk • contribs) 18:30, November 8, 2008

I am a longtime contributor to the WebKit project, and one of the people who often speak out publicly on its behalf. See [] to verify my 1771 commits, third-highest on the project. See [] for my many posts on the WebKit blog. I would like to address some of the facts in dispute on this discuss page. In my firsthand experience, I would say that Nokia (other than via the new Trolltech acquisition) has not made major contributions to WebKit. Only a few changes of theirs made it to the mainline; their separate branch in the WebKit repository for S60 was more of an in-tree fork but was never truly merged. Torch Mobile however has contributed significantly. This is something of which I have firsthand knowledge, and of which there is primary evidence in the source tree, but I am not sure what would count as a "Reliable Source" for Wikipedia purposes. Here is an ohloh.net page [] independently confirming 176 commits by George Staikos, president of Torch Mobile. From the details, you can see that they are on behalf of himself or other Torch Mobile employees. On this longstanding page on the WebKit Wiki, [], you can see that George Staikos is a WebKit reviewer. This is a difficult status to achieve, and indicates significant contribution, see this policy []. I believe these sources provide reliable independent verification of the contributions of Torch Mobile, in a more accessible form than the raw commit logs of the project. --MaciejStachowiak (talk) 18:55, 8 November 2008 (UTC)

I'm not disputing Staikos involvement in the project, there is plenty of evidence that he is a long-time contributor. However all the evidence I can find that mentions Torch Mobile, makes it look like it popped into existence in February this year. Has Staikos been contributing all this time as an employee of Torch Mobile? From this page it seems like Staikos Consulting and Torch Mobile were two differnet companies, are they actually the same? I know this is frustrating but the aim with Wikipedia is verifiability not truth. AlistairMcMillan (talk) 19:15, 8 November 2008 (UTC)

I believe Staikos Computing Services and Torch Mobile are two different companies. George Staikos is a principal of both. I do not believe Torch Mobile popped into existence in February; rather, like many startup companies, it was in "stealth mode" until that time. I believe evidence has been provided that it was incorporated in 2003. Thus, many of their earlier contributions were not claimed for Torch Mobile at the time, much as with Google's contributions. Usually, within the WebKit project, we are happy to let people credit their contributions on behalf of any company they work for at the time, if they so choose. But in any case, at least one of Staikos Computing Services or Torch Mobile should be credited, and I believe most of the contributors affiliated with both would rather have their contributions credited to Torch Mobile. --MaciejStachowiak (talk)

Sorry one other question. Even if Nokia made changes to WebKit that never made it back into the main tree, shouldn't they still count? They made the changes, they shipped the software on their phones and the source is out there. AlistairMcMillan (talk) 19:29, 8 November 2008 (UTC)

Committing back is generally seen as very very important. From what I know, when Apple started to use KHTML, there was a lot of controversy about their committing back in too big bundles that the KHTML developers couldn't understand well enough. The relationship between Ubuntu and Debian is still very edgy due to Ubuntu not committing back sufficiently to the project they build on. Vesal (talk) 19:38, 8 November 2008 (UTC)

Like most open source projects, the WebKit project only really counts changes as contributions if they make it back into the mainline source. There is nothing wrong with making your own branch or separate tree and making changes that only live there; but it is not the same thing as directly developing the project itself. I believe this is the norm within not only the WebKit project but also the open source community in general. --MaciejStachowiak (talk)

In response to Maciej Stachowiak (arbitrary break)

Thank you for this input. I think something should be done to reflect the real contributors. Your two last links does at least prove that Torch have people at the level of code review, which is indeed a very strong argument. However, in order to fairly compare with the other sources, could someone perhaps tell me what part of the OS news article "An Overview of Nokia's KHTML/WebCore-based S60 Browser" identifies Nokia as a contributor? If someone could quote that section, then we can compare what level of sourcing is required to make a statement about contribution to the mainline. Vesal (talk) 19:14, 8 November 2008 (UTC)

"With Frames and iFrames Nokia took the decision to treat them as a single page. For example, when you have two vertical frames, when you scroll down the page, both frames would scroll down."

"Webcore doesn't check the return value of malloc() for example (as this is not a big deal on the desktop but it is on smaller devices) so the Nokia engineers had to write wrappers and a memory manager to deal with such problems. Additionally the memory footprint, the mobilized components, scalability, usability and performance had to be tweaked thoroughly."

Those are two changes to WebCore. AlistairMcMillan (talk) 19:25, 8 November 2008 (UTC)

It probably depends on what attitude one takes on your question above about how relevant committing to the mainline is. Frankly, I don't know what is right here. But I think we should take Maciej's comment very seriously. Using a project is not the same as "developing" it, so having an Apple employee saying Torch contributed more means a lot to me. The verification policy is here to make this encyclopaedia better, not worse, but I admit the sources he provided are probably not explicit enough for what we need. Difficult situation. Vesal (talk) 19:47, 8 November 2008 (UTC)

I agree on listing Torch Mobile, but what do we cite as a source? AlistairMcMillan (talk) 19:49, 8 November 2008 (UTC)

I could post something public and citable regarding what companies I believe have contributed to the project, if Wikipedia would find such a secondary source more trustworthy than a primary source like the source code and ChangeLogs. But it would seem a little strange to do so. (As an Apple employee I don't have a particular axe to grind regarding who is credited; my interest in this issue is only as a matter of fairness in representing the WebKit open source project.) --MaciejStachowiak (talk)

Also on the subject of Nokia. For example, the N95 is the number two cameraphone on Flickr after the iPhone and, as far as I know, all the N95s have a browser based on Nokia's version of WebKit. Similarly the N82 and N73. All the statement on the article says is X, Y and Z have "further developed" WebKit. Nokia seem to have done that. And the sources support it. AlistairMcMillan (talk) 19:56, 8 November 2008 (UTC)

I would say that further developing a separate branch of WebKit is not the same as further developing WebKit itself. Much in the same way as WebKit does not really count as further developing KHTML, even though it is based on KHTML. That being said, I do think Nokia should be credited. One of the contributions by a non-Trolltech Nokia employee was for a significant new feature (CSS media queries), and I think counting contributions made by Trolltech employees both before and after the Nokia acquisision is fair. --MaciejStachowiak (talk)

If you could make a post somewhere about what you consider are the contributors to the WebKit project, it would by Wikipedia standards be a better source than the changelogs. Just as you wouldn't want Wikipedians analyze the Bible and draw their own conclusions, we can't have people like myself looking at commit logs and judge what features are more significant than others. I would say that your authority as a neutral Apple developer would count as a better source. Alistair, would something like a blog post by Maciej help? Vesal (talk) 21:16, 8 November 2008 (UTC)

I hate to ask Maciej to do that, but I can't find any secondary sources out there that talk about Torch Mobile's contributions to WebKit. However if Maciej did, that would give us something to cite. AlistairMcMillan (talk) 21:24, 8 November 2008 (UTC)

Sorry, butting in here because I offered a third opinion above. I'd suggest that a blog posting (on an Apple- or WebKit-controlled site) by MaciejStachowiak, or more accurately by Maciej, would be acceptable. My reasoning for this is that although blogs ... and similar sources are largely not acceptable, Self-published material may, in some circumstances, be acceptable when produced by an established expert on the topic of the article whose work in the relevant field has previously been published by reliable third-party publications. I'd argue that Maciej is clearly an established expert on WebKit development, whose work has been published by reliable third-party publications (non-Apple users of WebKit, Ohloh, etc).

Aside: MaciejStachowiak, thanks for your input here.

Aside (2): One concern I originally had, which is moot now, is that while Torch Mobile may very well have made some commits to WebKit, they may not be intrinsically notable. I feel now that if they have made more commits than, say, Google, then they are definitely notable. There is currently no article about Torch Mobile; does anyone without an insider interest in Torch feel up to creating an article?

Cheers, This flag once was red^propaganda_deeds 21:37, 8 November 2008 (UTC)

I agree with this; on our computing articles, we almost always accept blog postings from developers at the company who creates the software, when sourcing technical information or clarifications that would be nearly impossible to find elsewhere. Warren -talk- 23:26, 8 November 2008 (UTC)

On a side note, if the article someone wrote on Staikos Computing wasn't deleted from wikipedia for being insignificant then you would have a nice history of that company to be able to tie things together. —Preceding unsigned comment added by 219.101.207.134 (talk) 00:51, 9 November 2008 (UTC)

So how do we move forward with this? —Preceding unsigned comment added by 66.158.134.60 (talk) 19:04, 13 November 2008 (UTC)

That's discussed above. MaciejStachowiak has offered to write a blog posting on a reputable website noting that Torch Mobile is a major contributor to WebKit. We can then cite that posting in support of the claim here in the article.

Cheers, This flag once was red^propaganda_deeds 19:23, 13 November 2008 (UTC)

The WebKit Wiki has now been updated to include a section called "Companies and Organizations that have contributed to WebKit." Torch Mobile is listed there. See http://trac.webkit.org/wiki/Companies%20and%20Organizations%20that%20have%20contributed%20to%20WebKit. —Preceding unsigned comment added by Beanerz (talk • contribs) 06:19, 17 November 2008 (UTC)

It seems that some new browsers have been added to the list of WebKit derived browsers, while Iris Browser was removed before as being "insignificant". Is there any particular reason for this? The same issue is happening on the Acid3 page, where Iris Browser has the best Acid3 score around today, and yet it's being removed as insignificant (by someone who can't possibly know this as a fact). Given the amount of press, the number of users (which is actually significant), the fact that it is no longer beta, and this whole argument above, I think there is some imbalance here... —Preceding unsigned comment added by 80.58.252.237 (talk) 01:14, 18 February 2009 (UTC)