Internet Mozilla

Microsoft Blocking Firefox

It appears Microsoft is blocking Firefox from a number of it’s webpages. I’ve heard of this before, but now thanks to the new reporter tool, we have hard data to back it up. At the time I’m writing this there are 39 reports in the database. I should note that we have a little over 1000 users in the system at this point, so 39 reports is rather significant.

Most of the reports are about Firefox are that they are being told:

Sorry, we are unable to show you the page you requested. Please try again later.

In IE obviously the links work. Spoofing the user agent to that of Internet Explorer also has success. The area being reported is the downloads section of their site. Yes downloads worked. I tried several and had no problems.

Besides for illegitimately blocking Firefox, Microsoft has mislabeled their error message to encourage the user to “Please try again later?, as if it’s a server issue and will likely be resolved shortly. At least on Windows Update they say right away that it requires Internet Explorer, which is an honest dialog as Firefox does not support ActiveX which is used by AutoUpdate.

The only issue I can see with those pages is that the gradient on the top of the Microsoft site doesn’t render correctly. As I recall (and Asa blogged a while back) the CSS for it isn’t in the specs, hence unsupported in Gecko, though I haven’t verified that as I’m not a CSS guru. The same problem appears on the Microsoft homepage, but they don’t block that (presumably because the point of the page is to get people to buy things).

Come on Microsoft. You did move a bit towards standards. Save the browser sniffing for when it’s necessary (such as Windows Update). Let your customers use the browser that they feel comfortable using. Here’s a hint. We’re closing in on 65 million downloads in the next few hours. That’s 65 million people who believe in choice. Isn’t it time to respect computer users choices?

I hope to see Microsoft address and correct this issue.

Update: some report Firefox 1.0.4 working. Safari and OmniWeb seem to be blocked as well with that cryptic error message.

Update 2: It appears to be fixed now! Thanks to Microsoft for addressing it rather quickly (less than 2 full business days).

Google Mozilla

Summer of Code 2

Greg Stein left a comment yesterday about being able to submit an application to work on Mozilla for Google’s “Summer of Code” even if it isn’t listed. This morning I submitted a proposal. The basic premise is to allow me to accelerate my development of the reporter tool, including some server side changes (to help prevent abuse), and getting started (this is most likely longer term) on screenshot support, so users can submit a screenshot of the issue they have (should they choose to do so, it will be an option for obvious privacy reasons).

It sounds like a great opportunity, and a good excuse to give my Mozilla contributions greater priority.

Thanks to Greg for giving me the heads up on that.

Apple Mozilla

The Wall of Shame

Mac OS X 10.4 Tiger has a really cool new screensaver (video available through the link).

I think having the new reporter tool’s “Top 10 Worst Offenders” list as a screensaver to make an official “wall of shame” would be rather cool. I think that will be my new screensaver when it ships.


The List

♪♫ I’m making a list ♪♫
♪♫ I’m checking it twice ♪♫
♪♫ I’m gonna find out ♪♫
♪♫ Who’s naughty or nice ♪♫


Reporter FAQ

I’ve put together a quick QA on the new reporter tool to answer some questions I’ve recieved since I started working on this. These should address everything I’ve gotten so far (and cover a few bases). If you have other questions, leave a comment, and I’ll likely reply.

Why not just use Bugzilla?

Bugzilla was not the right tool for reporting evangelism for several reasons:

  1. Requires users to signup (extra work deters users from submitting reports)
  2. People don’t give enough information. “Yahoo is broke? is a bad report. We need several things to make use of the report: browser build, platform, exact page (no mistakes), problem description, and perhaps buildconfig or other information. Either users don’t give that info, some don’t know how to get that info, or they don’t bother. This creates bad reports.
  3. Because we gather accurate data (since most is automatic), we can generate much more accurate (and useful reports).

The results that reporter will give us will go into Bugzilla similar to how talkback works. Likely we’ll maintain a list of the top 20 sites bothering Firefox users.

Why can’t it just use a form?

A form would have been much easier to implement (and I did one very early on), but I ended up writing an entire extension and service for reporter to improve the accuracy of the data. With the extension, you can submit a decent report in under 20 seconds. You can submit a fantastic report in a minute. Most of the data collection is automatic, and perfect.

What about my privacy [you insensitive clod]!?!

I’ve spent quite a bit of time ensuring your privacy, and ensuring data integrity, and several things are in place to ensure it:

  1. A report is only submitted if you personally submit it. Reporter will never collect any data unless you launch the tool and submit a report. You need to do this for each page you report.
  2. There is no personally identifiable information associated with a report. We give you a random token (so we can get a user:report ratio). You never give your name, address, etc. Everything is anonymous.
  3. Yes, we do allow you to submit an email address (should you allow us to contact you in the [rare] event we need more information). Again completely optional. IP addresses are also recorded for security reasons. Both Email and IP addresses are not publicly accessible. Only an administrator can access such information. They are not searchable, or viewable in any way shape or form.

What will the data be used for?
There are several uses for the data. The first is to contact websites that are incompatible so they can fix their problems and work correctly in Gecko browsers (evangelism). The second use is to allow Gecko Engineers to see what problems are bothering the most users. This can help decide which bugs effect the most people.

What about SeaMonkey? Camino? Minimo?

I’ve already assured SeaMonkey compatibility. As far as Camino and Minimo support, it’s up to their respective development teams to come on board. I don’t have the development knowledge to do it myself, but I’d work with anyone who wants to bring such support.

When will it be localized for [insert language]?

The client side will be localized like anything else in a release. Same scheduling/policies/methods apply.

We will not be localizing webtool itself. It will be available in English. There’s just not enough cause to justify it. You can of course query the results based on the locale of the browser. And even search websites based on the TLD (for example .nl). This allows people who want to help improve support for a particular locale to zero in on what they need.

Will it be shipped with Firefox?

That remains to be seen. How it will be included (as an optional install like Talkback, default install, or not at all) has yet to be fully decided. I suspect to see it in non-release builds sooner than later.

I’m obviously lobbying to get this as a default on all installs. Since only a small percentage will participate [unlike talkback it doesn’t invoke automatically, it requires a user to do so on their own] it’s important we try and give the option to as many users as possible. The more reports from a larger audience, the more accurate the data will be.

When will it be in builds?

When cmp gets time to hook it up.

Is it true that you love monkeys?

Absolutely. And there’s a psudo-Easter egg to prove it.


Winter Break is upon us

That’s right. Final 6/6 is done. That means it’s time for 1 (6/6) Winter Break. I’ve got a few things on my agenda.

The first thing that needs to get done is a clean install of windows on my laptop. So I’ll be backing up this afternoon/evening and wiping tonight. Tomorrow I’ll start copying data and reinstalling. It should take several hours, but it’s necessary. I put it off for a while because the semester was still going on, but it’s time.

After that, reporter webtool is a big priority, I want to wrap that up, and get the second test deployed. Lots of improvements, a massive UI overhaul, and tons of tweaks, changes, bugfixes, enhancements. I’m looking forward to that.

Then I’ve got some more ‘classified’ stuff to take care of… Of course some sleep falls into this whole picture somewhere.

Overall, going to be busy, but should be fun.


Announcement: reporter-test shutdown for upgrade

In the very near future (perhaps as early as tonight), I will be shutting down the current reporter-test website for large upgrades. The upgrades are rather signifigant. I haven’t decided on a date/time. So anyone who wants to test and toy around should do so, and send comments to me ASAP.

since we are using test data, I will be wiping the database as well. the new version will capture more data. I want to test queries will all the data possible. We will drop data again before we launch the official instance of reporter. This is test data.

Those who installed the extension

You will need to uninstall it, and upgrade. the current extension will not work with the new version, a newer version will be needed. Autoupdate is not working on that extension, so you will not be notified. Please pass this message around if you know

request for help

I want to get SeaMonkey working with the same XPI. Does anyone know how to do overlay’s for multiple products? Is there anything I should be aware of (other than the EM API)? Please leave a comment or drop me a line.

Thanks to everyone for the feedback so far. Lots of changes have been made. Hence some serious upgrading is needed.


Mozilla Reporter


As I recall Asa and I started discussing this back in July, evaluating what the ideal system would be, how it would help, who would use it, etc. etc. It’s evolved a bit from the original concept. Much more mature. With the extension as the input device, it’s collecting very relevant, useful, and accurate data. No more bad/incomplete/typo ridden bug reports in bugzilla.

What is it?

Simply put, talkback for gecko. A simple automated way to collect feedback.

What are the goals

  • Provide Agregate data on incompatible sites with Gecko browsers
  • Provide an interface for end users to submit problems they encounter in under 30 seconds
  • Keep end users away from bugzilla, and force them to enter good Evangelism reports
  • Provide a tool for layout folks to use to track problems that effect Gecko users most

Why can’t I edit/fill out fields in the form

The goal is to provide relevant correct data. Allowing users to modify data makes it less accurate (typo’s, incomplete, incorrect data). The tool gathers data directly from the browser itself, so we know it’s correct, and sends it to the server.


The tool only invokes when you run it. That is the only time. Never will it send any data without your consent.

We will likely use a random hash as an ID per computer, so we know if 1 user submitted a site 100 times, or 100 users submitted a site 1X each. We won’t bind it to any other info. Nobody will be able to view the data and say “Hey checkout what Bob is viewing”. The final version will put an option to enter an email address. Only privilaged people (by a process yet to be defined, most likely similar to that of CVS commit access) will have access to such data. It will be optional data. We capture IP address’s, but again that’s behind a password. No regular user can visit and get the information.

It’s ugly

The version on my personal computer is much nicer :-D. I will be deploying an update to the server sometime soon (no timeline yet), containing many updates/changes/enhancements. It takes into account Ben’s recommendations, among many other things. When I do, current versions of the extension will no longer function. You’ll have to download the new build then. Sorry, but the auto-update wasn’t implemented in the current one. Simply because I didn’t really anticipate many people toying with it.

What about bugzilla?

Bugzilla isn’t going away. It’s just getting better. Consider this Bugzilla’s personal secretary. That’s really what this is. Just like with trackback, confirmed popular problems will be entered as bugs by Mozilla Volunteers on Bugzilla, and treated appropriately. This just prevents a bizillion dups, and lets users help without needing buzilla accounts, or query bugzilla, or figure out what we need to know. In a few clicks, the perfect report is submitted.

when will it be done

Were making quick progress. No distinct timeline, but Asa and I have discussed the idea of an end being in sight, and making preparations. So that’s a good sign.


Contact Me with any feedback you may have, or leave a comment. I’ll try to address them all in coming days, though I’m pretty busy with school, and this project.


Reporter: It’s coming

Nuff said…

… for now