Microsoft's Lack of Nightly Builds For IE 154
Ricky writes "Many wonder why Microsoft doesn't offer nightly builds of Internet Explorer — or at least something more frequent than months-to-years. Ars talks with Microsoft's general manager for IE, who says the IE9 development cycle will look much the same as previous versions. Not a great idea."
Obvious... (Score:5, Insightful)
Many wonder why Microsoft doesn't offer nightly builds of Internet Explorer
Um, because they never have and never will?
Re: (Score:2, Insightful)
Nobody else uses Trident (IE's rendering engine), and if Trident breaks, a lot of other stuff in Windows breaks. They don't want to release development versions of their browser, because their corporate customers don't want users breaking things.
Frankly, I'm wondering what benefit nightlies would have for MS, who does pretty much all of their testing in-house.
Re: (Score:3, Insightful)
'if Trident breaks, a lot of other stuff in Windows breaks'
Which is, of course, precisely the reason to have a meaningful suite of automated tests and frequent build/test cycles. You'd rather work 6 months on something and then throw it over the wall to testers only to have them come back with either hundreds of regression failures (best case) or a handful of failures so severe they couldn't even get past the basic smoke test script?
That's even before you get to your user community, which as the article poi
Re: (Score:2)
Which is, of course, precisely the reason to have a meaningful suite of automated tests and frequent build/test cycles.
The story doesn't say that MS doesn't have nightlies and automated tests internally (I don't know how it is for IE specifically, but I am virtually certain that they do in fact have both; I would be very surprised if any MS project in development stage didn't have either).
The story is about MS not releasing those nightlies to the end users
Re: (Score:2)
You're right, we don't know what the IE development team has or doesn't have. Still, what's the point of going to all the trouble to create a good automated build/test/release system and not have frequent deliveries to the end users - the web development community in this case? I've seen teams do that, and over the life of the project it felt like a net loss because the time and effort spent babysitting the build was wasted because break in the feedback loop by not getting it to end users meant that what w
Re: (Score:2)
I haven't read the article (why break tradition) but the point is probably to make sure nobody's broken the build, and possibly to give them some of their own dog food to eat.
So far as I know, very few closed-source development organizations release a nightly build to end users outside their company.
But there may be a large group of Microsoft developers who get their systems auto-loaded every morning with the latest IE. Probably all of the IE developers at the least.
Re: (Score:2)
My "why" was really more of a rhetorical question. In my experience, setting up and maintaining an automated build/test/deploy for a complex thing like IE is quite a lot of work, and I'm not sure any team would benefit from the effort if the only benefit was just keep the build from breaking. That could be done by having a designated team member manually build the system whenever there was time (hah) and a need to check it. No, to make a frequent build worth the effort, you need to get the resulting softwar
Re: (Score:2)
Frankly, I'm wondering what benefit nightlies would have for MS, who does pretty much all of their testing in-house.
Well it worked out well for Win7, didn't it? Doesn't mean IE would fall the same way (far from it in fact), but you'd think MS would learn something after a while.
Re:Obvious... (Score:4, Informative)
The nightlies of Microsoft Bob basically killed all positive hype for the program.
OMG My Eyes! (Score:4, Funny)
The nightlies of Microsoft Bob basically killed all positive hype for the program. I don't want to see Microsoft Bob in nighties! It was bad enough seeing Bill Gates wiggling his ass for the cameras.
Oh, nightlies ...
I'll have to ask Clippy for linux [tri-bit.com] and get back to you on that ...
Re: (Score:3, Funny)
I don't want to see Microsoft Bob in nighties!
Hey... be thankful he wasn’t nude, like Clippy was...
Re: (Score:2)
Re: (Score:2)
Ha! Like I’m going to take advice on that from a guy with a paperclip dick.
Re:Obvious... (Score:5, Insightful)
In short, why the hell would they release nightly builds?
Re: (Score:2)
Yeah... my main thought initially was “they’re Microsoft – they don’t need a reason,” but that was only very shortly followed by “how on earth would that benefit them, us, or anyone else for that matter?”
They DO do nightlies... (Score:3, Informative)
From what I've gleaned from various Microsoft blogs, they DO release nightly builds, internally to all their own testers and employees.
That way, as far as I can tell, they get all the benefit of nightly builds, with absolutely zero of the downsides in terms of company image and dealing with buggy software in the wild.
Re: (Score:2)
They probably have more people using nightly builds of IE than there are folk using nightly builds of FF!! Oh the irony!
Re: (Score:2)
Last I checked, there were about 200k-300k Firefox nightly users. Microsoft has around 90k employees as of 2008 (worldwide).
So unlikely, but at least possibly similar ballpark if every single MS employee uses the IE nightlies (again, unlikely).
Re: (Score:2)
Where do you get that figure of 200-300k FF nightly users. That's mighty impressive.
Re: (Score:2)
> Where do you get that figure of 200-300k FF nightly users
I believe it comes from the update server (which is basically contacted by nightlies every so often to see if there's an update to the next nightly, just like other builds contact it to see if there's an update). Then you extrapolate from the number of update checks. It's a rough estimate, of course; as are all usage numbers.
And maybe I'm confusing the nightly and beta, of course! If I find the source where I saw the number, I'll comment.
Re: (Score:2)
Simple. It means that bugs get found and fixed early before too much code depends on the incorrect behavior. The inability to fix bugs without breaking things has always been IE's greatest weakness.
http://developers.slashdot.org/comments.pl?sid=1451976&cid=30178810 [slashdot.org]
Re: (Score:2, Informative)
Er, no (Score:2)
Haven't been around much, I take it?
Developing apps to use in the wild means that the end-users will find all sorts of ways to do the wrong thing. These are not things you can always test for before a release (beta, rc, whatever), unless you do double-blind usability using people who've never used your software.
I've come to expect that when I release a new version of whatever or add/change a feature, I will get immediate feedback from the field because someone will add a new URL to a list that passes my v
Re:Obvious... (Score:5, Insightful)
As somebody who has frequently participated in beta tests of lots of software, including Microsoft's, this is spot-on. Sure, their infrequent betas get some good feedback and some good bug reports, but they also get absolutely drowned in a deluge of people on the discussion boards (newsgroups, actually) who complain about:
A) Nothing particular at all, they just signed on to complain.
B) Stuff that's completely unrelated to the beta (such as a complaint about IE6 on the IE8 beta discussion)
C) Stuff that's completely unrelated to the product (complaints about Excel on the IE8 board)
D) "How dare Microsoft release [a beta of] this product with such-and-such [known, sometimes in release notes] bug!"
E) "WTF I installed the latest version of X, and now I can't access my Y, so I'm switching to competitor Z and never buying anything Microsoft again!"
F) Complaints about Beta 1 bugs during Beta 2 or RC test phases.
G) Complaints from people who installed the software on a production machine, and expect Microsoft to provide support for it.
These are the types of morons that Microsoft has to deal with. I've seen some of this type of behavior in other betas, to be sure, but some of the problems, especially D, E, and G, are most common on the MS betas. People just seem to expect that any code from MS will be production-ready and expects the company to stand behind their software as though it were a released product.
Microsoft would be *insane* to release nightly builds to a group like that. A closed beta nightly program, maybe (participants culled from those who are actually useful and productive on the public beta) but certainly not open. Especially considering point F above; people already can't always keep up with the pace of the infrequent releases, and asking them to identify the build number they're using would be an exercise in futility for far too many.
Re: (Score:2)
Re: (Score:1)
Re:Obvious... (Score:5, Insightful)
Shitty article. Nothing to see here....
Re: (Score:2)
I’ve just had an epiphany.
I hear there's a pill for that.
Re: (Score:2)
Wait... was it the blue one, or the red one?
Re: (Score:2)
Re: (Score:2)
Why do they care? They’re Microsoft.
Agile isn't the only legitimate way! (Score:5, Interesting)
What does MS offer nightly builds for??? It's just not how they work. They're a typical monolithic development house that deals only with releases and occasionally lets beta code out. There are benefits to the approach like not trying to shoot a moving target when it comes to bugs etc. People who've grown up with agile seem to think it's the only way to do quality assurance.
Re: (Score:1, Funny)
I like to compare it to taking a shit versus rubbing a quick one off.
Microsoft saves up the feces, and savors it in their intestines for months and years. That lets it get really stinky and, well, shitty. Then, in one big blast, they crap it out all over everybody. That's just their way.
The open source community prefers to rub a quick one off each night. Rub rub rub and the load has been blown.
So it all depends on what you prefer. Would you rather get nightly blasts of jizm in your face, or would you prefer
Re: (Score:2)
Re: (Score:2, Flamebait)
How's that non-agile, waterfall QA working out for ya?
Re: (Score:2)
Believe it or not, waterfall is not the only alternative to agile. In fact, your comment represents one of the biggest critisms of agile. It misrepresents the alternatives and in many ways seems more like a religion than a development methology.
http://steve-yegge.blogspot.com/2006/09/good-agile-bad-agile_27.html [blogspot.com]
Re: (Score:2)
Wow, you can't come up with anything newer than a link to Yegge's tired old disproven rant?
Re: (Score:2)
The majority of his criticisms are very valid. The fact is, there are alternatives to agile and waterfall, but agile shills prefer not to talk about them...
Re: (Score:2)
Which ones can you show are valid and have not been addressed by agile?
Re: (Score:2)
How's that non-agile, waterfall QA working out for ya?
How's that constantly changing never quite know what you're building agile shitfight working for you?
See it ain't hard to be immature. It ain't clever either. Some incredible projects have been built with both methodologies and there have been some abysmal failures with both too.
Re: (Score:2)
How's that constantly changing never quite know what you're building agile shitfight working for you?
It works quite well, actually, when you have frequent releases to the user, because they can say "yes this is good" or "no this isn't good", and only OCD types really worry about knowing exactly what's what at any time. In fact, software projects tend to delude themselves into thinking they know what they've got when they really don't. This leads to all kinds of problems later when the users see it and find out that what they got is nothing like what they were told they were going to get.
Think of it this wa
Re: (Score:2)
Excellent. So as a developer you're not just rude, abusive and unprofessional, you're also a one trick pony who worships at the alter a single methodology, and have an uncanny knack for convoluted useless metaphor. Please file your resume here in the filing cabinet marked "Waste bin". Don't let the door hit your arse on the way out. I'd really like to think you're just trolling but unfortunately I have had the misfortune of working with your ilk before.
Re: (Score:2)
Not all software has the same cost of failure. If a particular bug can cost tens of thousands of dollars, I sure as hell want to know exactly what releases had that bug. There's nothing OCD about this desire.
Re: (Score:2)
I sincerely hope that even Agile software houses don't release nightly builds to customers.
Re: (Score:2)
Re: (Score:2)
The vast majority of companies with nightly builds to NOT give those nightly builds to customers.
I know of one, who does so, apparently quite successfully. I played golf this summer with their director of Q/A. I'm struggling to remember their name and what they do, but it was some web based SaaS offering with 300-ish employees in downtown Seattle (we were playing golf at Willows Run, which is just down the street from Microsoft Intergalactic in Redmond). We got to talking agile, and releases, and he said
Re: (Score:2)
ROFL! The only thing you know about agile is that is is a buzzword, that's it.
Wow. You're an amazing person. You must be very well paid to be able to work out how much I do or do not know about agile from a single post. Yes that was sarcasm. In case your too obtuse it means I think you're a wanker and you know nothing about me.
My company uses agile effectlively, and everyone knows that agile is comprised of iterations (typically 1, 2, 4 weeks or whatever you want).
Everybody knows squat. I've seen agile ap
Re: (Score:2)
And the story isn't about having nightly builds. It's about releasing nightly builds.
Presumably, this may have some relation to Agile, since the latter seeks "constant feedback from the users", using most current state of the project.
Comment removed (Score:5, Insightful)
Re: (Score:3, Insightful)
The author of the article seems to think IE should be treated separately from Windows.
I guess Konqueror should have it's own update system, the OS update system isn't good enough?
Re: (Score:2, Informative)
I guess Konqueror should have it's own update system, the OS update system isn't good enough?
Its, not it is.
Re: (Score:2)
Touche-with-the-accent-that-I'm-lazy-to-put-on. I usually try to be good about that, too.
Re: (Score:3, Informative)
Touché
Also, I’m = I’m.
Okay, I’ll shut up... and damn slashcode for not letting me use the … character code.
Wait a second... oh my god... did they really fix Unicode?
“Touché.”
It looks right in the preview. Just my luck it’ll screw it up as soon as I post.
It still doesn’t allow arbitrary characters, though... just certain ones. The ellipsis just melts into oblivion...
Re: (Score:2)
They did? Hmmm. Tèstíng.
I am shöcked.
Re: (Score:2)
I am shöcked.
What a coincidence! I just bought one of those at Ikea last week.
Re: (Score:2)
AC, see>>
Confused about article, any developers here? (Score:2)
Although many developers may not really care much for nightlies or even point releases, it keeps them in the loop, and keeps them interested.
This confused me. Many developers don't care about them, but they do care about them... is that basically what the Ars article is saying?
Re:Confused about article, any developers here? (Score:5, Insightful)
Additionally, the article seems to take some things for granted...
the reality is that every other browser has a more regular release cycle than IE does, and that keeps them relevant.
I guess Opera's release and development cycle(s) is why it is so popular!
The result is a strong perception that IE is lagging behind, no matter how great the major release versions are.
The perception that IE is lagging behind has nothing to do with a bad development cycle, it's more tied to ... bad development and a not-very-good product.
and the browser's updates are pushed through Windows Update. The actual browser doesn't have its own updating system, and this is a large part of the reason that over 40 percent of users are still using IE6 and IE7.
That's an interesting assertion. The only backup he gives are numbers for browser stats.
On the whole, this seems like one guy doing an editorial and talking off the cuff. That's how it struck me, anyways.
Re: (Score:2)
The perception that IE is lagging behind has nothing to do with a bad development cycle, it's more tied to ... bad development and a not-very-good product.
And opening up the process with, perhaps, a chance to incorporate feedback early in the process is a great way to address this. You want to give people what they want? be more responsive and don't cast the featureset in stone based on whatthe product manager says.
Re: (Score:2)
I guess Opera's release and development cycle(s) is why it is so popular!
I thought I read in the news that they were pulling the plug on that in 2011...
Re: (Score:2)
Re: (Score:3, Funny)
Re: (Score:2)
Actually, Opera doesn't have nightlies. Weeklies at best most of the time. And there's no "bleeding-edge" build available. They have the next major version cooking apparently, but they aren't sharing anything until they have exhausted 10.x it seems. Firefox and Chrome, on the other hand, have nightlies of both current and future releases.
Re: (Score:2)
The perception that IE is lagging behind comes from IE lagging behind. Not having any updates for year after year, and playing catch-up and copying features.
They have repeatedly stated they are not interested in dog-and-pony show tricks like ACID tests, instead focusing on what people use and what would make things easier for developers.
While this would make sense for an isolated company with an isolated product, they have to remember that their engine should render pages that were made by a non-microsoft
Yea they have it (Score:2)
I'm sure the developers have one, or maybe something that can be done ondemand. No point having one for the public imo
Normal (Score:5, Insightful)
Re: (Score:3, Insightful)
Whether or not you think that it is a good idea for there to be IE nightly builds, it isn't exactly absurd to judge a product by the standards of other similar products, rather than other products with similar licenses.
Re:Normal (Score:5, Informative)
WTF? Most companies don't release nightly builds of their software.
Not when it comes to web browsers. You can get nightlies from every single other major browser, except for IE.
Re: (Score:2)
Some tester actually want to get paid... (Score:2)
Re: (Score:3, Insightful)
Safari != webkit, Chromium != Google Chrome. Sorry to burst your bubble.
Webkit is a rendering engine. Its pretty useless without supporting code. The link you gave links you to a loadable library essentially. The app icon you get for OSX actually runs a script that has Safari use the webkit library from the package, but the UI and everything else is still the same old Safari thats installed on the system.
If someone bothered to put the effort into it, you could stuff IE's renderer into Safari on Windows,
Re: (Score:2)
. . . Chromium != Google Chrome. . . .
Chromium is not Chrome. They may share a common tree, but they aren't the same either. Chrome may be built from a snapshot of the chromium tree, but that doesn't give you nightlies of chrome.
Okay, what kind of logic is that? A nightly is by definition a daily build that's a snapshot of the development trunk, which official releases are eventually built off of. The Chrome developers themselves use Chromium for development. It's a nightly build in every sense of the term.
Even if for some reason you don't buy that logic, Chrome has a Dev Channel, which releases official Google-branded Chrome builds every week or so. So, it still has weekly builds.
That gives us three out of four non-IE brow
Who is Many? (Score:5, Insightful)
"Many wonder why Microsoft doesn't offer nightly builds of Internet Explorer."
Whoever "Many" is, they seem to always be interviewed by Ars and FoxNews.
Re:Who is Many? (Score:5, Funny)
Many believe this was a random baseless comment.
Re: (Score:2)
Of course not.... (Score:2)
Re: (Score:2)
The IE guys are going to have to fix any problems in how it plays nice with Windows anyway, and if the development process is so broken that they can't even keep O/S-breaking regressions out of the builds, there's a problem. The whole point of having frequent builds is to identify errors sooner, while it's cheaper and easier to fix them, than later, after the edifice has been built on the unstable foundation.
with Firefox if you file a bug they appreciate that and generally fix it right away, even security vulnerabilities aren't promptly fixed on IE, let alone user suggestions....
I suggest that's the whole point of wanting IE to have a more frequent build and release cycle -- ge
Re: (Score:2)
The IE guys are going to have to fix any problems in how it plays nice with Windows anyway, and if the development process is so broken that they can't even keep O/S-breaking regressions out of the builds, there's a problem.
The problem is that IE is tightly coupled with several other Windows components. This means it can break many other apps which can depend on it, or also can break itself if it depends on something which is not available.
For example, when IE7 came out people were extracting the installe
Re: (Score:2)
A monolithic blob of pieces that have tight coupling between each other and leak those dependencies to other pieces that shouldn't even KNOW about another applications dependencies in the first place is bad software engineering. Again, the article is correct in urging the frequent build/test/release to the public cycle because it would highlight the dependency errors quickly and allow them to be fixed before they become a problem. We all know DLL hell (or RPM hell -- linux doesn't get a pass here, either) a
Re: (Score:2)
As a key product in a proprietary OS, why would you want to run nightly builds of IE? With Firefox my browser may be unstable, but at least the rest of my system stays stable, but with IE a lot of Windows components use Trident and that isn't going to be a good thing.
WebKit is heavily used in OS X too, AFAIK, but they still provide nightlies. It doesn't replace the existing version on your system, it's an extra one that you can manually run as part of specific programs. So you'd use the nightly build to test, but other programs would keep using the standard system build.
It also doesn't have to be called "Internet Explorer". It could be called something different, like "Trident Development Version" or something. Firefox nightlies are called Minefield.
Plus, with Firefox if you file a bug they appreciate that and generally fix it right away
You must have
Nightmare (Score:2)
More Microsoft Bashing (Score:5, Insightful)
Re: (Score:3, Insightful)
Because many of us use their stuff and despair at the problems that arise that we cannot fix and the Microsoft will ignore.
That creates a culture of just complaining to each other about the company in general. We say to each other things like "this was the company that was given the BSD source code on a plate and still couldn't get even ping right" and other things non-techies would find completely irrelevent.
Just filter the MS stories out - there's not going to be muc
Re: (Score:2)
While I think Microsoft is right with its release cycle, the article is based on the fact the every other browser vendor is releasing snapshots.
For me, the biggest picture is interaction and strategy, not builds. In Webkit, Gecko and Presto, if you are a web developer, you can interact with the engine developer. They have mailing list, good bucktrackers, and a *good attitude* towards fixing bugs.
For Microsoft, if you are using Linux for development (a pretty common case I'd guess) you cannot even try. I dou
Maybe it takes more than a day to build it? (Score:2)
The Microsoft build labs has been described in many books but one thing that stood out to me was the alleged fact that most builds, like Windows, take well over 24 hours to finish. Given how tied into the operating system that MSIE is, I suppose that a build of MSIE would require a significant build of Windows as well.
Re: (Score:3, Funny)
Well, there's yer problem... The are probably running their build server on windows. Now if they ran a nice Linux build server...
Re: (Score:2)
Re: (Score:3, Informative)
Given how tied into the operating system that MSIE is, I suppose that a build of MSIE would require a significant build of Windows as well.
I find it highly unlikely. In the end, IE lives in its own library, and any OS services that may need it call through that via stabilized COM interfaces. There's no reason why Windows can't be build against precompiled IE binaries and .idl files describing the interface, and similarly no reason why IE can't be built against the most up-to-date Windows SDK headers.
This simply does not make sense (Score:3, Interesting)
I am not a fan of Internet Explorer at all - however I know people who are, and I can't imagine this mattering to them in the least.
Heck, I can't imagine the vast majority of Firefox or Safari/Chrome users caring about those available snapshots; and I say that as someone who has used nightly builds for both those products fairly frequently!
This just seems silly on the face of it. "Microsoft doesn't follow Firefox's development path", complains a Firefox fan.
Coming up next (Score:5, Funny)
Can bees think? A new study indicates that no, they cannot.
This story is bookmarked (Score:5, Insightful)
Filed under "weirdest story ever to appear on /."
Next week we can discuss the outrage that stems from Microsoft's refusal to offer free back massages on the New York subway.
It doesn't need a nightly build. (Score:2)
IE Feedback system EXISTS (Score:2)
1. It required me to get a live account. I did.
2. It required me to "register" with my live account. It didn't work with Webkit so I fired up Firefox, then I did.
3. It required email confirmation of my live account. I confirmed it.
4. It asked me to register to be able to vote on issues (GOTO 2)
No wonder the feedback is minimal and useless (as in not real bug reporting). There's probably no one who cares on the other end anyway
joelonsoftware would disapprove (Score:2)
I would assume that the Microsoft Excel team did it this way as well, since Joel mentions it in his article. But they also wrote their own compiler [joelonsoftware.com] because everyone else's was crap, and still managed to ship on time.
McDonalds' Nightly Builds... (Score:3, Interesting)
Nightly builds, if they were released every time:
Bun
Bun
Bun
Bun
Meat
Meat
Bun + Meat
Bun + Meat
Bun + Meat + Meaty Flavour
Bun + Meat + Meaty Flavour
Bun + Meat + Meaty Flavour
Bun + Meat + Meaty Flavour
Bun + Meat + Meaty Flavour
Bun + Meat + Meaty Flavour
GHERKIN!
Bun + Meat + Meaty Flavour + Gherkin
Bun + Meat + Meaty Flavour + Gherkin + Salt
Bun + Meat + Meaty Flavour + Gherkin + Salt++
Bun + Meat + Meaty Flavour + Gherkin + Salt+++++
Bun + Meat + Meaty Flavour + Gherkin + Salt + Tomato
Bun + Meat + Meaty Flavour + Gherkin + Salt + That Other Stuff
Bun + Meat + Meaty Flavour + Gherkin + Salt + That Other Slightly Better Stuff
Quarter Pounder With Cheese
As an IT Manager for one of the 100 biggest companies in the world, I couldn't give a flying f*ck about the inbetween. All I want to know is what we're getting. And if it breaks a part of our fundamental application stack, we'll complain or won't use it. If I want something in the release, I'll lobby for it. If you want to be part of the IE development cycle, sign an agreement with MS to be a part of it, then you'll get the alphas and beta.
Total non-story.
Re: (Score:1, Funny)
Your eyes... is that moisture?
Re: (Score:3, Insightful)
Maybe you hadn't noticed, but development of IE7 and IE8 have not been tied to a specific OS at all. IE7 was released before Vista and installs on XP, and IE8 well before Win 7 and that installs on Vista and XP. Microsoft has said that IE9 will be released in 2010, while Windows 8 is set for 2012. IE and Office are both on different development timetables than Windows -- although Office is almost always released 6 to 8 months after a desktop Windows release. Sure, they're linked in some senses because e
Re: (Score:2)
A long release schedule ensures that Microsoft has to find and squash every bug themselves to make a single presentable product that will have a better reputation than your average nightly build. It ensures developers see only a stable standard and worry that this year's version will have different quirks than last year's version and that they won't find out until it is too late to do anything except hack workarounds for the browser bugs into their site. What harm would nightly releases cause Microsoft [e
Re: (Score:2)
Microsoft releases Alfas and Betas to many different communities of testers
And those releases have multiple cycles and run a long time so there is ample opportunity for developers of dependent s/w and web pages to test against the coming release and provide feedback to Microsoft. How many builds of W95 did I load...including one from floppies that took 20+ hours of feeding floppies before the first boot.
I wonder if Microsoft might just have as many in-house testers using the daily builds of IE as there are
Re: (Score:2)
IE is so tied to the internal of Windows that installing a nightly IE you are touching too many internals that could break easily.
Sounds like a good reason for not making it tied to windows. Ah, poetic justice, sort of.