IE and the Demise of Borgzilla
September 19, 2007 at 6:04 pm
Filed Under Mozilla, Technology
11631 people have read this post.
I could have titled this “Things that will not make me new friends on the Internet Explorer team” but this is written out of a real desire to see some things change. I know many many of the people working on IE, some of whom I worked with for five or nine years on a number of projects.
One of the last things I worked on when I was at Microsoft was “Internet Explorer Feedback” aka Borgzilla (the PR flacks refused to allow me to call it that at the time but I worked it into every conversation I could…).
Borgzilla was officially an open bug database for the last part of the Internet Explorer 7 release. Less officially, it was a semi-open bug database that only a few people looked at directly. I detailed it with a FAQ in a blog post on the IEBlog, one of the few under my name there. Borgzilla ran on the larger “Microsoft Connect” site. Connect is used largely for Betas and beta feedback by a number of teams but the idea of having a (semi)open bug database was really brought to live by the guys at Visual Studio, who did a great job at it.
The idea with the Borgzilla was to start a part of the virtuous cycle that Firefox and open source projects have used to their advantage over time in interacting with community and participants. I pitched this as something to help build our user community, when I was on IE, and to help people feel less frustrated about the black hole, which is where IE Feedback generally seemed to go except on the IEBlog. IE was working to dig itself out of the hole of years of neglect and getting the community to help by reporting issues and to know their feedback was being heard seemed like a really good idea. The team also wanted a better understanding of how people were using IE in web development and day to day use.
So, that was the idea. Implementation was a little more problematic as it turned out. As soon as the site went live, it was immediately deluged in bugs. Hundreds would appear over the course of a few hours. Also, since Microsoft refused to have the bug database on the Connect site be the same as the main IE bug database (which was part of the Vista one), there was a disconnect in information. Basically, bugs reported by the community were linked to copies made in the main IE bug database. When the main IE bug was resolved, the status was passed back. This worked in practice except that it isolated the bugs from the main team since they didn’t work in the same database in which users were reporting issues.
I (and eventually a contractor) was supposed to triage bugs and mark the ones for porting that were seen as passing basic muster and not simply being noise. When thousands of bugs showed up within a week, this became something akin to bailing out the sinking Titanic with a large wooden bucket (which happened to be on fire). Whenever we got nearly caught up, a new build of IE7 would go out and it would start again.
Community members could validate bugs or add comments but were not empowered to edit bugs beyond that or resolve them in any way. This means instead of being something where the community was fully empowered to log and resolve issues on its own, everything had to be passed through a gatekeeper or two and then given to the rest of the IE team. Also, since the main Windows bug database (unlike Bugzilla for Firefox) was not viewable by community members, they had no idea if issues were already known or had even been fixed (since the community could not get nightly or weekly builds).
Needless to say, the members of the IE team were not terribly thrilled with aspects of this system as well. Because of the strong push to ship Vista, there was a “bug bar” in place on the maximum number of active bugs a team could have on their plate, usually on the level of 30 or 50 bugs, not hundreds. If groups exceeded this, they were called to task by various higher level groups in Windows to see why their quality was slipping and they weren’t moving in the right direction to ship Vista. When you’re working under that kind of limitation and 120 new UI or CSS bugs suddenly show up for four or five people to triage and work on, people become less happy. This was even more true when many of the bugs were issues that they either knew about or which had to be sent to the QA team to create reproduction steps and clear test cases.
This process was in place up until IE7 shipped after the last Beta, I don’t believe that anything was done with the reported bugs. I could be wrong. I left the team in early May, 2006 in order to move to California.
If you go to the site today, you are told: “The site is temporarily closed. It will re-open in the future.” This message has been in place for about a year now, which probably redefines “temporary.” No information has been given about this and the few times I’ve asked about it on the IEBlog have been ignored. I get the impression that no one really wants to talk about the site.
I’ve tried to find out through contacts at Microsoft why the site no longer exists and heard things from a variety of people but it seems to largely come down to a lack of interest by the IE team. The feedback from a user community in this direct of a manner was never really wanted by certain members of the IE team. The project, as a whole, was never given enough resources (staffing) to do a good job and the software used, Connect, did not empower the community to solve its own problems or to police the incoming bugs. When bugs did come in, the various teams within IE that received them found them to be a resource drain to deal with and, I believe, resented all of the work that they created.
We’ve heard this week that there might be an IE8 Beta before the end of the year. I would certainly hope so since coding on IE7 was complete a year ago. I would hope that they have been working hard on IE8 since that time. This seems like a reoccurrence of this historical black hole problem that IE has had. The IEBlog has been pretty silent on anything forward looking during the last year as well.
I’d like to call on the IE team to do the following:
- Make regular alpha builds of IE available. If not nightly, then weekly. If not weekly, then monthly. The IE team builds every night and runs daily tests on the build quality. There is no reason that the IE team cannot make builds available except for a desire to do so and a willingness on the part of Microsoft’s management to let them do so.
- Turn IE’s MS Connect Site Back On: Flip the bit on the site that still exists and re-open it. Actually put some staff on it this time and dedicate them to getting issues triaged during the alpha period, not just late beta, so the IE development and program management teams aren’t buried by issues when under pressure to ship. Alternatively, use some better bug tracking software and actually empower members of your user community to resolve, validate, and close issues so IE team members don’t have to do all of the work themselves.
- Start Talking about IE8 and the Future: IE is very present in a number of theW3C working groups now. I hear that many discussions are occurring and that IE members are pretty engaged. Start talking about the future direction of Internet Explorer so that the rest of us in the same ecosystem can see where you are going, at least in a general sense.
The IE team liked to speak during the IE7 development period about putting the past behind them and being engaged with the rest of the web community, especially web developers. This shouldn’t have gone on hold when IE7 shipped. Engagement is not something that you do for a year and then turn off for another year before repeating. It is a regular process of communication and openness.
Come on, IE, let’s talk.
All original work is licensed under a Creative Commons License.
Comments
21 Responses to “IE and the Demise of Borgzilla”
Leave a Reply
All original work is licensed under a Creative Commons License.
Brilliant post. Valuable insight that unfortunately confirms everyone’s suspicions that Microsoft should actually be renamed Microsloth. Big and powerful in theory, but in reality just suck so much life out of the world through their own ignorance and stupidity.
Imagine if ‘globalisation’ was more tangible. For example one company built all the roads around the world. This company was not accountable to any government or body and built the most crappy roads full of pot holes and poor signage, you name it …
That is what Microsoft is and IExploder is the roads. They UI is the street signs, the rendering engine is the asphalt. Stretch the metaphor further if you like but suffice to say that using IExploder is like only being able to travel on a freeway that is littered with potholes and not being able to demand better from the government responsible for overseeing the freeway’s maintenance.
Hello Mr Al Billings,
I’m against a public bug reporting database at Microsoft regarding Internet Explorer at this time for some of the reasons you have given yourself in your own post:
- noise/signal too high
- duplication/high frequency of duplicated bug reports (on the same issue)
- disconnection of the team IE dev. members from reporters in several ways: the community reporters are blind while the IE dev. team is mute.
OTOH, I am all for that Microsoft IE dev. team creates an internal database from all the bugs which are (and have been) explained, are documented, with testcase coming from web developers themselves:
- bugs that create application crash, hang, cpu maxed in MSIE 7: those bugs exist and are already publicly reproducible in testcases. So, it’s all up to Microsoft to wake up and to fix those
- bugs that creates usability and/or accessibility problems
-
- collapsing adjoining margins implementation: this affects almost every single CSS columnar webpage template
- float implementation
- inherit implementation
- background (background-image, background-color) implementation
- position and z-index implementation
- CSS inline box model bugs
I’ve said this same thing in many places repeatedly. With *_so many_* documented, reproducible, testcase-ed bugs, specs violations occuring in MSIE 7, there is no need right now to create a public database.
You want people to start talking about IE 8? I’m against that. Talking is good when you need planning, establish a roadmap, a schedule, a to-do list, etc. Action is what is needed now, not talk. Bug fix is what’s needed now, not talk. Some of the bugs and spec. violations (HTML 4, DOM 1 & 2, CSS 2.1) regarding MSIE 7 have been reported, explained and documented for over 9 years now. Why would we or should we need to talk about this again?
I’m for monthly build available for testing purposes. But then again, it won’t mean much if already reported bugs - for many years - and there are hundreds of them - still have not been fixed.
Regards,
Gérard Talbot
Well thats good news. M$ will never learn. Long live the stupidity gates
Thanks, Al. It’s good to have a reasoned voice with some Microsoft experience (and presumably credibility there) discuss the problem and propose solutions.
Bugzilla is a big reason that I as a Web developer prefer Firefox to IE and Opera. When I encounter a problem, I can find out whether it’s already known, and if so, what if anything is being done about it. I try to help out by qualifying bugs that I’ve encountered that are not well understood. This makes me feel engaged with the product. In some small way, it’s mine. Loyalty follows pretty much automatically.
Microsoft’s tendency to secretiveness is powerful, though, as their arguments to the EU Court of First Instance - “we can’t reveal our protocols; it would be like giving away our source code” - indicate. That decision, like others, went against them, but I’m not sure there’s much reason to hope that this will encourage them to open up.
A pity, because they could have a much better product and maybe earn a little loyalty from the development community if they’d talk.
Pretty much spot on Al.
I’m not sure alpha builds of IE8 are useful until later in the project cycle when they are nearer to feature complete. With IE7 they were issuing new builds once every couple of months during the last year of the project which was fine. I see little point in giving people access to builds with incomplete and shaky features. That only increases the noise on a public bug database.
I don’t believe that they should just turn connect on. The IE team needs to put a system in place that will actually work this time. The system before did not work for a variety of reasons. It’s a shame the team decided to just turn it off instead of fix it. I remember being amazed when that decision was announced!
They do need to start talking about IE8. I’ve been saying that for some time. The silence is deafening and leads to the developer community losing faith completely. Even just a general roadmap and intentions of what IE8 will do would help overcome the feeling that the team has disappeared again.
What would be really useful is a public bug tracking system that links to KB articles and recommended workarounds for known issues for the currently supported releases. Helping web developers out on the currently released and supported versions of IE would be really useful, show that the team does care about developers and give them insight into what issues developers are struggling with on a day to day basis.
Cheers
-Dave
What worse? Some of them have just been covered over. I discovered this when I IE7 showed my company’s as broken even though it wasn’t. I started pulling my hair out as I realize they “fix” various css hacks by change the character the hacks were using to a different character. So instead of, I dunno, have IE7 recognize standard CSS they made me hunt through the character map until I found a new one that worked.
Drives me batty. I had to use the separate style sheet for ie7…and a separate for ie6 and a separate one for Firefox,mozilla,opera.
Dave,
It’s good to see you commenting and you are probably right about a more useful bug tracking system.
I disagree about builds though. I think even shaky builds, which is what you get with Trunk nightlies for Firefox 3, would be good because they would show the direction and evolution of features but it is a mixed bag.
For those that don’t remember or know Dave, he and I worked together on IE. He also left Microsoft within the last year.
Hey Al :)
The thing to remember is that IE is NOT open source, nor is it ever likely to be. Comparing how Firefox releases nightly builds is not really relevant unless IE becomes open source. What web developers want from IE is a good heads up period about what is coming. Builds released too frequently with bugs coming and going depending on who’s checked in what code that day doesn’t really help anyone. Unless of course you are working on a competing browser and want to know where IE is going. Remember how quickly features of IE7 such as quicktabs were copied once they were available to the public!
The IE team does need to talk about when IE8 can be expected and what it will focus on. The silence is only hurting them in the eyes of frustrated web developers.
The IE team does need a public bug tracking system that not only helps the IE team but helps web developers find appropriate workarounds for issues on released and supported product.
The IE team does need to release beta builds on a regular basis. BUT only when they are of sufficient quality and feature completeness that it is worth web developers spending their valuable time looking at it.
Cheers
-Dave
So, really, it’s both your faults for quitting then? :)
I’m with you Al, the MS IE team needs to get up, and start talking.
1.) They absolutely need to give some insight, parameters or something about what is happening with IE8. It can be an ETA, a roadmap, a feature list or a bugfix list. It doesn’t need to be everything, just one of the above will do.
2.) When we say bring Borgzilla back online, we mean, “bring a meaningful, useful back tracking system online”. If required, the old IE Feedback would be fine in the interim.. until a better alternative is available.. even if only in a readonly state, so that at least users can see a bug has been submitted (e.g. the IE team knows about it), and that if there was a workaround, they can see that too!
3.) Long term, when a company (in this case MS) holds the reins on a development platform, used one way or another by 75% of the PC market, it absolutely has to have a communication channel, right from internally at MS to the public bug tracker.
** ** ** ** **
For starters, just open up a bug tracking system. let users fill in details, search, track etc. Once open, tie in the MS Tools to it.
So, in MS’s tool(s), when a bug has been fixed, and verified, push that info to the public system, even if it just says “issue fixed internally”. It doesn’t even have to give a build or patch date, just some indication that it is actually being worked on, and progressing.
But I digress, the IE Blog tells the full story. MS simply doesn’t care anymore.
I must confess, I would love to know if internal struggle at MS is one of the reasons why you both left. From the outside, it just looks like MS is too big to get a handle on anything.
Wheee, it’s the Al & Dave show :-) BUT. You guys left. So that’s minus 200 points, and guarantees that Dean’s not going to let you use borgzilla.com anytime soon. BUT. I own nozilla.com, so perhaps you can do something useful by way of bugtracking with that (?)
Borgzilla? Sounds assimilative to me ^_^
IE can still release alphas while being closed-source: http://www.opera.com/products/desktop/next/
Al,
I admire your gumption. you’re now gone from the team but you’re still trying to change it. He’s hoping you have better luck this time.
Take care
-Adam
p.s. I thought you left for some mobile media startup, what’s with this Mozilla stuff?
Hi Adam,
I went to MobiTV (mobitv.com) and I worked there for a year. It wasn’t a good longterm fit so I left the company after Mozilla approached me about working there.
After I left, the head of QA also departed from MobiTV, so I’m hardly the only departure.
I want IE to be a good citizen in the browser world like Opera, for example, which is another closed source project.
[...] came to my attention that comments started to be posted about my Borgzilla blog post over on the IEBlog on an unrelated post [...]
[...] former Microsoft (and now Mozilla) employee Al Billings explains, the Internet Explorer team went some way to mending burnt bridges when they opened up their lines [...]
Dear Al-
You sound like a fascinating person. Stumbled on your blog while searching for some tech help with Mozilla.
Love to meet your Momma. I’m a 60 year old United Methodist pathologist. Love the people in my church, but I must admit that my really significant religious experiences where I felt I had a glimpse of the infinite have all involved nature. I have a very good and well-grounded friend who recently became a Wiccan; time is very limited right now, but when I can, I want to explore this with her. So I’ll follow your blog, especially around paganism.
Now, I do have a problem, and wonder if you might be able to help me. I love Firefox. However, the new update keeps downloading itself, I keep telling it to install when I restart my computer, because I’m busy, then when I restart, it tells me that the update can’t be completely installed, so I cancel. Then it redownloads (isn’t it amazing that were really can use such words and have them mean something???)and the process starts over again. Any ideas?
Thanks and best wishes,
Rand
Found this thread while looking for info on binding multiple event handlers.
It was interesting to hear something from the microsoft side bsides the usuall, “Why are you complaining IE works fine” (when thats the only browser you code for :-/ )
I had suspected that a lot of these problems were at the level of (mis)management and corporate culture.
If IE doesn’t get its act together it will cost them significant market share over the next 3-5 years. If your browser can’t support emerging technologies effectively (sure I can make an AJAX data call but if I want to apply that data to my page dynamically I need to write 3-10X as much code as I would for a DOM complient browser). Its only pissing off people at the “leading edge” [developers] at this point but when Joe/Jane Public can’t install their favorite facebook app because IE pukes or crashes they’ll be asking what is “Firefox [et all...]” in a big hurry.
I’ve posted more than a couple times in the last few weeks that “who ever developed IE should be gang raped by a troop of rabid baboons”, I’m sure some of you have felt my level of frustration…
As some of the other posters commented, I don’t really care what new stuff you’re going to add when you haven even fixed YEARS old problems [select innerHTML bug any one?].
Its like a contractor who builds your house with a cracked foundation and a leaky roof coming back a year later and saying “I’d like to paint your house for you, it will look pretty”, yes it might, from a great distance BUT IT WON’T KEEP THE FREEKING RAIN OUT!
Maybe a class action law suit to bill back all the hours wasted by developers on IE workarounds. Then again maybe they are still under the delusion M$ is “leading the world” on line
Come on, Al. You know you miss it!
Cheers, Will
[...] those who haven’t read it yet, it’s an absolute must to go through IE and the Demise of Borgzilla, written by Al Billings, who was formerly working with the Internet Explorer team. To summarize, [...]