User talk:BarkingFish/Story preparation/Preferential browser treatment?

prepared

=Moved from Water Cooler=

OR Proposal: Investigative reporting and survey into YouTube & Google Chrome
Hi. I'm conducting investigations into issues being experienced by a large number of YouTube members, who report that they seem to have severe difficulty uploading to the site using any browser other than Google Chrome. I'm looking to OR this into a story, but first, I am conducting an online survey on an independent site called Surveymonkey, into people's uploading experiences, browser usage and so on.

The results of the survey will be collated to see others experiences of uploading, and presented to Google with a request for their comments on the issue.

My own (and others) suspicion, as I am a victim of this problem (as are about 460+ other users of a private web group I belong to) is that Google are making it harder for people to upload to YouTube using external browsers other than their own, in an attempt to gain more users for their products - which essentially means they're committing an unfair trade practice by inhibiting the free choice of users to select their own products.

I have created a short survey with Surveymonkey, which I wish to publicise via site notice at Wikinews, in order to make people aware and ask for their responses. The survey is set up in such a way that I have no idea who has filled it in, as there is no IP address stored, and no personal data of any description.

The link to the survey is here: http://www.surveymonkey.com/s/7DQ9SP2 and consists of 8 simple questions which I've shown to others before committing the survey, and no problems are immediately foreseen.

Thanks for your help. BarkingFish (talk) 21:34, 7 February 2011 (UTC)
 * I think what you're asking is, "Could an admin please add this to the Sitenotice?" There are a couple of different site notices for Wikinews, both for logged in and not-logged-in users. -  Amgine | t 22:20, 7 February 2011 (UTC)
 * Spot on, that's exactly what I'm asking. I want non logged in and logged in users to see the link - the purpose of the explanation above is merely to state why I've created the survey, its purpose and to make it totally clear I'm doing it for investigative reasons. BarkingFish (talk) 22:28, 7 February 2011 (UTC)
 * For the record, I've got a Limesurvey install if anyone wants to commandeer it. — μ 22:55, 7 February 2011 (UTC)
 * I suppose someone has to ask this. Would putting this on the sitenotice compromise the neutrality of the project?  (I've not yet formed an opinion, at this point; it seems an important question, though.)  --Pi zero (talk) 23:20, 7 February 2011 (UTC)
 * I don't think so, Pi zero. In my opinion (although you're free to form your own), the survey is being used as part of a potential Wikinews report, so simply asking people to fill it out I don't think will compromise its neutrality. The sitenotice can be dismissed, so it gives the user free choice as to whether to fill it out or not, so I don't foresee a problem.  If you do, please tell me what you think could arise in the way of issues as a result of posting it here. BarkingFish (talk) 23:40, 7 February 2011 (UTC)


 * This should be trivially easy to prove, or disprove. Simply use a browser gadget that fakes the ident string to match Chrome's. If the problem vanishes, they are giving preferential treatment. Oh, and my opinion is sitenoticing this is questionable WRT project neutrality. --Brian McNeil / talk 00:40, 8 February 2011 (UTC)


 * Most decent websites I would assume have a way of spotting faked idents, and when you say "project neutrality" - do you mean Wikinews's neutrality, or the project I'm undertaking? BarkingFish (talk) 01:55, 8 February 2011 (UTC)


 * I believe xe means Wikinews's neutrality. Using the project sitenotice would appear to imply official backing for your research &mdash; and appearance of neutrality is important.


 * Software is often written stupidly (for a variety of reasons, not all of which are themselves stupid); I'd suggest doing the experiment Brian described. A negative result might, as you observe, not mean anything, but a positive result is conceivable.  --Pi zero (talk) 02:43, 8 February 2011 (UTC)


 * Houston, I think we have an answer. -

3 screenshots from my YouTube uploads, 2 done in opera 11, using an IE8 spoofed header and opera's own, and the 3rd being the file manager showing a successful upload in processing, the third one being produced with a chrome header manually set in Opera's preferences.

Now the question is, where the **** do we go from here? BarkingFish (talk) 04:44, 8 February 2011 (UTC)
 * Heh; Isn't it handy to have a geek or two available? If you can get some of Bawolff's time, ask him to try and replicate the above. If he can, then I'd say we've grounds to take this further. The steps I'd recommend are as follows:
 * Write a simple, concise how-to for people to try and replicate this and document their tests.
 * See if you can get a sample of 20 or more people, geographically distributed, successfully reproducing the error.
 * Identify which of Google/YouTube's distributed servers they're accessing (i.e. traceroute to whatever www.youtube.com resolves to for them).
 * Lay out the evidence as a report proving preferential treatment for their own browser.
 * In one shot (something I'll do if we prep our report off-wikinews (happy to host on wikinewsie.org) is hit the us DoJ, the European Commission, and (I'd be cheeky and defer ~48 hrs) challenge Google that they're breaching competition laws due to YouTube being the 'de-facto' video sharing service.
 * Incidentally, for your test, you should be using a standardised video. Upload that somewhere that all people conducting the test can grab it from. --Brian McNeil / talk 16:55, 8 February 2011 (UTC)
 * Addendum: I'd really want a couple of quite techie people to do some additional investigation into this. To whit, running a packet sniffer on their lan whilst conducting the tests. The purpose of this is to establish if Chrome is assigned preferential servers for upload to. That would be a definitive smoking gun on this. I can put the proposal to the mailing list for our LUG and see if the idea piques anyone's curiosity. --Brian McNeil / talk 17:12, 8 February 2011 (UTC)
 * Moar! I've looked up contact details for appropriate DoJ and EC DG Competition offices. I will email both a tentative query this evening to establish if they believe that Google using YouTube to promote their browser would breach competition rules. --Brian McNeil / talk 17:25, 8 February 2011 (UTC)

Additional research
Since Google's YouTube would need to be considered a near-monopoly for video sharing, I've done some digging:




 * If anyone can find more current coverage on the popularity of online video sharing services, please add 'em. A cursory check only turned up mobile-related stuff in the past few months.
 * I may look for some ComScore data, or try to blag some. --Brian McNeil / talk 22:05, 8 February 2011 (UTC)

Taking this somewhere quieter
I'd strongly suggest this is ready to move out of the WC and into the talk of a prepared story.

As those on scoop will note, I've 'gone fishing' with prospective prosecutors. Additionally, I've asked my local LUG if anyone is interested in doing a lab-style analysis.

I do not want to go near MS, Opera, or Mozilla until we've got a lot closer to having an exposé to publish. There's a host of possible causes/Google actions that could give the symptoms we've seen.

So, how's about popping all this to Prepared? --Brian McNeil / talk 22:05, 8 February 2011 (UTC)

Just for clarification
Here are the User agent strings used in the Spoof test:

IE8: Mozilla/4.0 (compatible; MSIE 8.0; Windows NT 6.0)

Chrome: Mozilla/5.0 (Windows; U; Windows NT 6.1; en-US) AppleWebKit/534.17 (KHTML, like Gecko) Chrome/11.0.655.0 Safari/534.17

Opera's OWN header: Opera/9.80 (Windows NT 6.1; U; en-US) Presto/2.7.62 Version/11.01

BarkingFish (talk) 22:52, 8 February 2011 (UTC)

Usage survey
Per discussion, I thing the following points should be surveyed:
 * 1) Most frequently used video sharing sites
 * 2) Usage patters: download/upload
 * 3) Linking with social networking sites
 * 4) Browser usage
 * 5) Video media format usage
 * 6) Any associated issues/problems uploading or downloading

The latter two should be restricted to those above a certain upload threshold.

The sample size has to be large enough to give meaningful data, that's going to be tricky. Amgine may be able to suggest additional points relating to this. --Brian McNeil / talk 09:04, 9 February 2011 (UTC)