Hacker News new | comments | show | ask | jobs | submit login
Firefox: browser ballot glitch cost us 9m downloads (pcpro.co.uk)
57 points by neya 1845 days ago | hide | past | web | 24 comments | favorite



I am quite surprised that this "glitch" ever happened. I am usually no friend of conspiracy theories, but I find the following coincidence a bit tasty:

* Microsoft is known for its obsession with rigorous testing.

* They had to pay around a billion for anticompetitive behaviour _and_ implement the "browser ballot". Failure to comply would cost them up to 10% of the annual turnover.

* There was a "glitch" that deactivated the ballot for 15 (!) months for a full service pack.


Do you really think that Microsoft would...

1) Pretend that no one would notice.

2) Do a risk to reward analysis, where they would gladly pay out 7 billion dollars if caught (over a few million extra IE installs).

3) Implement a fix within one day after notice, and offer to extend this ballot on their brand new OS for another 15 months.

The simpler explanation is that due to the nature of the ballot, their test Suite did not include the browser ballot in it, and was delegated to someone in the legal department that filed if away and forgot about it.

This ONLY affected 28 million PCs. That's a very small fraction of the user-base. And considering IE's market share, this bug would have only caused less than 1/2 of that number of non-installs of other-browsers... As afterwards, anyone could have just downloaded what they were looking for anyways.

Also note that for 15 months NO ONE (well, just a few people apparently) contacted MS or the EU Commission about not seeing the browser ballot. I guess the browser ballot is as important as you might think it is to the people of EU.


I didn't say that they did it on purpose. My point is that this is incredibly stupid, especially for an organisation of that size and the projected fine. Yes, we usually don't test for edge cases, but the 7 billion edge-case? So how was this allowed happen?

Also, I don't think that this was not reported for 15 months. The problem only became widely known after the EU actually decided for a formal investigation. I found no coverage about when the problem was reported.

The 28 million-number is rather irrelevant, as the news coverage states that all of windows 7 SP 1 was affected, so they failed at a major release level. It wasn't 2% of the release or something. This could even have been cought by manual testing.

I don't care how important the ballot is in the real world (for the record: I hate the ruling). But this is one of the biggest companies in the world failing to comply with a pretty clear rule that came out of the most expensive ruling ever dropped on them.


My fault. I completely misread your comment.


My fault as well, I didn't make clear that I did not actually believe that, but that the coincidence really lends itself to one.


> 2) Do a risk to reward analysis, where they would gladly pay out 7 billion dollars if caught

Where is the $7B figure from?

> This ONLY affected 28 million PCs.

That's over 3.7% of the European browser market (more, because I assumed each of the 739M people in Europe have a browser).

A few % in browser market share is something people care about very much.

> Also note that for 15 months NO ONE (well, just a few people apparently) contacted MS or the EU Commission about not seeing the browser ballot.

Why would the general public know this even existed or should exist? The only people I would expect to notice it are browser vendors and hardcore browser enthusiasts, a miniscule minority. But that doesn't mean the measure isn't important. It wasn't set in place so people would consciously notice it, it was set in place to have a specific effect.


> A few % in browser market share is something people care about very much.

Especially, when you compare being #1 vs #2. That's not to say it's intentional, but there is a reason negligence can be a crime. People do all sorts of stupid things by pushing the line and not just directly intending to do so.


>2) Do a risk to reward analysis, where they would gladly pay out 7 billion dollars if caught (over a few million extra IE installs). //

Did they pay the original fine yet?

Yes, I think Microsoft would purposefully break the "browser ballot" in the hope that they could win back a significant share with their new browser focus before re-instigating it.

FWIW I found the whole browser-ballot thing ridiculous.


> Did they pay the original fine yet?

The fine is upheld, so I'd imagine so.

https://news.ycombinator.com/item?id=4167174


So, is that a yes or no?

https://viewer.zoho.com/docs/urlview.do?url=http://curia.eur...

Shows they were fined in March 2004. That document is the confirmation, in June 2012, that Microsoft will be charged what is essential a late-payment fee based on a decision in 2008. They had 2 months to appeal.

I'd be amazed if they didn't appeal a penalty charge of Euro 890 Million.

I can't find a reference to MS having actually paid the fine, they have paid others in the past amounting to more than this amount; it's just part of their cost of doing business for them I presume.

Anyone got a citation, a line entry in some accounts, a message to shareholders or something that one can look at to confirm payment?


I'm observing quite a few serious regressions in another Microsoft product - Visual Studio 2012, so IMHO this is not completely unlikely to be an accident. On the other side - Steve Ballmer ...


Would you care to elaborate on the "serious regressions" you perceive to have happened?


Chances are he is referring to the VS.NET 2012 uppercase menu letters. That specific "regression" is a greatly debated topic here, with emotions running hot. 100+ comment threads were started over that one.


The menu letters case is just a registry setting.

There are many shortcuts not working properly, e.g. http://connect.microsoft.com/VisualStudio/feedback/details/7... , windows with strange state management, the new temp document not working properly.


The article's headline is inaccurate. From Harvey Anderson's blog post (http://lockshot.wordpress.com/2012/10/30/windows-eu-ballot-s...): "Cumulatively 6 to 9 million Firefox browser downloads were lost during this period."


The graph on lockshot shows 8.7m. I assume the author of the pcpro article just rounded up. I'm not saying the author should have rounded up, because she shouldn't have. Journalists should strive for accuracy in paraphrasing the words of their subject. I wouldn't necessarily call it inaccurate though, since it falls within what Harvey estimates as well. I'd rather call it slightly overzealous reporting.


Rounding from 8.7 to 9 is perfectly acceptable, especially with something that's an estimate to begin with.


I like how everything here is using the word "lost."

Well, they weren't gained, but they certainly weren't lost. I somehow doubt that people decided against downloading Firefox because it wasn't in the browser ballot.


> I somehow doubt that people decided against downloading Firefox because it wasn't in the browser ballot.

Did you read the article?

>Daily downloads of Firefox fell by 63% to a low of 20,000 before the ballot was reinstated, Harvey Anderson, vice president of business affairs and general counsel at Mozilla, said in a post on his personal blog.

>After the fix, downloads jumped by 150% to 50,000 a day, he said - estimating that between six and nine million downloads were "lost" during the 18 months the browser ballot was missing.


Why does that graph show world-wide downloads? Shouldn't it instead show EU downloads only to give an accurate account of the effects of the glitch?


It took 15 months for this to come to light?

A more cynical person might say this proves the silliness of the whole "browser ballot" concept.


I can't help but question whether some of the higher download volume in 2012 was due to the new rapid release cycle rather than the browser ballot being implemented.


Fair enough this will no doubt prove to be a costly mistake for Microsoft, but I seriously doubt the drop in Firefox downloads is due to Microsoft not implementing the browser ballot and most likely due to the fact that Chrome is a better browser and Firefox were slow to address plugin memory leak issues (which they only just recently fixed).


Why the downvotes?




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact

Search: