IBM’s Rob Weir was not the first to spot the problem, but did a great job in writing it up, both when initially observed and after it was fixed by Microsoft.

It was an algorithm error, a piece of code that did not return the results the programmer intended.

Unless Microsoft chooses to tell us, there is no way to tell how the error happened. However, as Weir and others observe, it may be significant that a Google search for something like Javascript random sort immediately gets you sample code that has the same error. Further, the error is not immediately obvious, making it particularly dangerous.

I am sure I am not the only person to turn to Google when confronted with some programming task that requires some research. In general, it is a great resource; and Google’s own algorithms help a little with filtering the results so that sites with better reputation or more inbound links come higher in the results.

Still, what this case illustrates – though accepting again that we do not know how the error occurred in this instance – is that pasting code from a Google search into your project without fully understanding and testing it does not always work. Subtle bugs like this one, which may go unnoticed for a long time, can have severe consequences. Randomisation is used in security code, for example.

As an aside, there also seems to be some randomness in the appearance of the browser choice screen. It turned up on my laptop, but not on my desktop, although both have IE as the default.

And who would have guessed that the EU would arrange for so many of us to get an ad for something like the GreenBrowser popping up on our desktop? Apparently it is the “best choice of flexible and powerful green web browser”, though since it is based on IE it is less radical a choice than it first seems.

Related posts:

1. Why you can’t trust a Google ad
2. The sad story of the LG Viewty – case study in Web 2.0 failure
3. Google ranks MSN search top

SharePoint is a mixed bag of course, but once it is up and running the browser user interface seems reliable as a means of getting at your documents over the internet. That said, it is inconvenient to run up the browser and navigate to a web site whenever you want a document. A user recently highlighted another issue. Their company uses a web application that frequently requires documents to be uploaded. This is straightforward if the document is on a local hard drive or network share, but not if it is in SharePoint. The workaround is to save the document out of SharePoint to the local drive, then upload it.

Fortunately there is another option. SharePoint Explorer View lets you access documents through Windows Explorer; you can even map SharePoint as a network drive. Now you can browse documents without a web browser, and upload directly to a web application.

Sounds great; and when it works, it is great. Troubleshooting though is a world of pain. If you have looked into this, you will know that there are really two Explorer Views, one using Internet Explorer and ancient FrontPage protocols, and the other using WebDav and Explorer. It’s the second of these that you most likely want. However, achieving this is notoriously troublesome, raising uninformative messages such as “Your client does not support opening this list with Windows Explorer", or from the command line System Error 67, or System Error 53 “The network path was not found”.

Another common complaint is incessant login dialogs.

I discovered a few useful resources.

This white paper on Understanding and Troubleshooting the SharePoint Explorer View is essential reading.

From this you will discover that if you are using Windows XP, the WebDav SharePoint Explorer view will not work over SSL or on any port other than 80. You are stuck with the FrontPage view, which is less useful. Apparently Microsoft has no intention of fixing this. Upgrade to Vista or Windows 7.

In addition, many XP and even Vista users find this update essential before anything starts working. It is necessary on Windows 2003 since the web client is not installed by default. It does not apply to Windows 7 though.

A good resource on the repeated login issue is here. It can be tamed.

Windows 7 is better, though I experienced an odd issue. One Windows 7 machine cheerfully opened the Explorer view to a remote site on port 444. I could engage Explorer View from the SharePoint web site, or from Network in Explorer, and it just worked.

On another machine, same network, also Windows 7, same web client settings, I could not get it working. I was on the point of giving up when I happened on the right incantation from a command prompt:

net use s: https://your.domain.name:444\shared%20documents /user:domain\username password

In this example S is the drive letter for a mapped drive, your.domain.name is the URL for SharePoint, 444 is the port number, shared documents is the folder name. For some reason this worked instantly.

Well, SharePoint is an option. Before leaving this subject though, I would like to mention Gladinet, a third-party utility which is able to mount a variety of cloud storage providers as network drives, including Amazon S3, Google Docs, Windows Live SkyDrive, and in the latest version Windows Azure.  It works on XP, Vista, Windows 7 and Windows 2003, comes in 32-bit and 64-bit editions, and worked immediately in my quick test. The ability to mount drives in Explorer itself, as opposed to an Explorer-like application, makes a big difference in usability.

Gladinet does not support SharePoint, sadly. Still, before you roll out SharePoint it is worth considering that something like an Amazon S3 account requires no CALs (though third-party clients like Gladinet may do), is maintained by a cloud provider rather than on your premises, is not hooked in any way to Windows clients, and might be a lot less hassle to deploy.

I do also understand the attraction of SharePoint, if you don’t or can’t trust the cloud, and like the way it integrates with Active Directory or its other clever features such as versioning or workflow management. What I don’t get is why Microsoft makes basic features like Explorer View so hard to get working.

Finally, this aspect of SharePoint should get better in Office 2010 and SharePoint 2010, which includes SharePoint Workspace 2010. This will synchronize with SharePoint 2010 document lists, giving you an offline copy you can access in Explorer. Agnes Molnar has a summary with screenshots.

Related posts:

1. Google storage 10 times cheaper than Azure – but not as cheap as Skydrive
2. SharePoint 2007 tip: use Explorer not the browser to upload documents
3. Amazon Elastic Compute Cloud gets persistent storage

The build process was interesting: you actually compile Chromium OS from a chroot virtual environment. My first efforts were unsuccessful, for two reasons. First, Chromium OS assumes the presence of a pre-built Chromium (the browser), so you have to either build Chromium first, or download a pre-built version. However, the Chromium build has to be customised for Chromium OS. I did manage to build Chromium, but it failed to run, with what looked like a gtk version error, so I gave up and downloaded a zip.

Chromium OS itself I did build successfully, though I ran into an error that needed this patch, which I applied manually. I was using the latest code from the git repository at the time. I expect that this problem has been fixed now though you may run into different ones; life on the bleeding edge can be painful.

I also had difficulty logging in. You are meant to log in with a Google account, which presumes a live internet connection at least on the first occasion. Although Chromium OS successfully used the ethernet connection on my laptop, getting an IP address and successfully pinging internet sites, the login still failed with a “Network not connected” error. Studying the logs revealed a certificate error. You can also create a backdoor user at build time, so I did that instead.

Once I got Chromium OS up and running, booting from a USB key, I found it mostly worked OK. It is a fascinating project, because of Google’s determination to avoid local application installs, thereby gaining better security as well as driving the user towards web solutions for all their needs.

That’s a bold vision, but also an annoying one. Normally, when reviewing something relevant like an operating system or a word processor, I try to write the review in the product I am testing. In fact, I am writing this post in Chromium OS. However, I could not write my review on Chromium OS, because I needed screenshots; and although there are excellent web-based image editing tools, I could not find a way to take screenshots and paste or upload them into those tools. The solution I adopted was to run Chromium OS in a virtual machine – I used VirtualBox – and take the screenshots from the host operating system.

It is a small point; but makes me wonder whether Google will end up bundling just a few local utilities to make the web-based life a little easier. If it does so, third parties will want to add their own; and Google will be under pressure to abandon its idea of no local application installs.

Another interesting point: the rumour is that Google will unify Chrome OS  with Android, which does allow application installs. Can that happen without providing a way to run Android apps on Chrome?

Chromium OS includes a calculator utility, which opens in a panel. Mine does not work though; I get a blank panel with the URL http://welcome-cros.appspot.com/calculator.html – which seems to be a broken link. Still, is that really a sensible way to provide a calculator? What about offline – will it work from a Gears local web server, or as a static HTML page with a JavaScript calculator, or will it not work at all?

I will be interested to see whether Google ends up compromising a little in order to improve the usability and features of its new OS.

Related posts:

1. Google Chrome for Mac and Linux will be a long while coming
2. Building Chromium
3. Google names its Chrome OS partners – including Adobe

Go runs only on Linux or Mac OS X. I installed it on Ubuntu and successfully compiled and ran a one-line application. I used the 32-bit version, though apparently the 64-bit implementation is the most advanced.

Pike claims that performance is “typically within 10%-20% of C”. No debugger yet, but in preparation. No generics yet, but planned long-term. Pointers, but no pointer arithmetic.

Go does not support type inheritance, but “Rather than requiring the programmer to declare ahead of time that two types are related, in Go a type automatically satisfies any interface that specifies a subset of its methods.”

Google has many projects, and while Go looks significant, it is dangerous to make assumptions about its future importance.

I don’t think Google is doing this just to prove that it can; I think it is trying to solve some problems and doing so in an interesting way.

Related posts:

1. Programming language trends: Flash up, AJAX down?
2. Prism: official Delphi language comes to Visual Studio
3. What is Microsoft’s new language?

I’m reviewing Hyper-V Server 2008 R2, so I figured it would be interesting to migrate the VMs. I attached an external USB drive, shut down the  VMs and exported them. Next, I verified that there was nothing else I needed to preserve on that machine, and set about installing Hyper-V Server 2008 R2 from scratch.

Aside: when I first set this up I broke the rules by having Active Directory on the Hyper-V host. That worked well enough in my small setup; but I realised that you lose some of the benefit of virtualisation if you have anything of value on the host, so I moved Active Directory to a separate box.

I wish I could tell you that the migration went smoothly. Actually, from the Hyper-V perspective it did go smoothly. However, I had an ordeal with my server, a cheapie HP ML110 G5. The driver for the embedded Adaptec Sata RAID did not work with Hyper-V Server 2008 R2, and I couldn’t find an update, so I disabled the RAID. The driver for my second network card also didn’t work, and I had to replace the card. Finally, my efforts at updating the BIOS had landed me with a known problem on this server: the fans staying at maximum speed and deafening volume. Fortunately I found this thread which gives a fix: installing upgraded firmware for HP’s Lights-Out Remote Management as well. Blissful (near) silence.

Once I’d got the operating system installed successfully, bringing the VMs back on line was a snap. I used the console menu to join the machine to the domain, set up remote management, and configure the network cards. Next, I copied the exported VMs to the new server, imported them using Hyper-V manager running on Windows 7, and shortly afterwards everything was up and running again. I did get a warning logged about the integration services being out-of-date, but this was easy to upgrade. I’m hoping to see some performance benefit, since my .vhd virtual drives are dynamic, and these are meant to be much faster in the R2 update.

Although I’m impressed with Hyper-V itself, some aspects of Hyper-V Server 2008 R2 are lacking. Mostly this is to do with Server Core. Shipping a cut-down Server OS without a GUI is a great idea in itself, but Microsoft either needs to make it easy to manage from the command line, or easy to hook up to remote tools. Neither is the case. If you want to manage Hyper-V from the command line you need this semi-official management library, which seems to be the personal project of technical evangelist James O’Neill. Great work, but you would have thought it would be built into the product.

As for remote tools, the tools themselves exist, but getting the permissions right is such an arcane process that another dedicated Microsoft individual, program manager John Howard, wrote a script to make it possible for humans. It is not so bad with domain-joined hosts like mine, but even then I’ve had strange errors. I haven’t managed to get device manager working remotely yet – “Access denied” – and sometimes I get a kerberos error “network path not found”.

Fortunately there’s only occasional need to access the host once it is up and running; it seems very stable and I doubt it will require much attention.

Related posts:

1. Hyper-V Server 2008 R2: a great deal for Windows virtualization
2. Mixing Hyper-V, Domain Controller and DHCP server
3. Hands on with Hyper-V: it’s brilliant

Blank and flickering screens, failure to recognize hard drives, defaulting to the old 2.6.28 Linux kernel, and failure to get encryption running are taking their toll, as early adopters turn to the web for answers and log fresh bug reports in Ubuntu forums.

Did I get it wrong? Should I be warning users away from an operating system and upgrade that will only bring them grief?

I doubt it, though I see both sides of this story. I’ve been there: hours spent trying to get Bluetooth working on the Toshiba laptop on which I’m typing; or persuading an Asus Eee PC to connect to my wi-fi; or running dpkg-reconfigure xserver-xorg to try to get Compiz working or to escape basic VGA; or running Super Grub to fix an Ubuntu PC that will not boot; or trying to fix a failed migration from Lilo to Grub 2 on my Ubuntu server.

That said, I noticed that the same laptop which gave me Ubuntu Bluetooth grief a couple of years ago now works fine with a clean install, Bluetooth included. It’s even possible that my own contribution helped – that’s how Linux works – though I doubt it in this case.

I also noticed how Ubuntu 9.10 has moved ahead of Windows in several areas. Here are three:

1. Cloud storage and synchronization

   Microsoft has Live Mesh. Typical Microsoft: some great ideas, I suspect over-engineered, requires complex runtime to be downloaded and installed, not clear where it fits into Microsoft’s overall strategy, still in beta long after it was first trumpeted as a big new thing. So is this thing built into Windows 7? No way.

   By contrast Ubuntu turns up with what looks like a dead simple cloud storage and synchronization piece, web access, file system access, optional sharing, syncs files over multiple computers. Ubuntu One. I’ve not checked how it handles conflicts; but then Mesh was pretty poor at that too, last time I looked. All built-in to Karmic Koala, click, register, done.

   

2. Multiple workspaces

   Apple and Linux have had this for years; I have no idea why it isn’t in Windows 7, or Vista for that matter. Incredibly useful – if the screen is busy but you don’t fancy closing all those windows, just switch to a new desktop.

3. Application install

   This is so much better on Linux than on Windows or Mac; the only platform I know of that is equally user-friendly is the iPhone. OK, iPhone is better, because it has user ratings and so on; but Ubuntu is pretty good: Software Centre – browse – install.

I could go on. Shift-Alt-UpArrow, Ubuntu’s version of Exposé, very nice, not on Windows. And the fact that I can connect a file explorer over SSL using Places – Connect to server, where on Windows I have to download and install WinScp or the like.

Plus, let’s not forget that Ubuntu is free.

Of course you can make a case for Windows too. It’s more polished, it’s ubiquitous, app availability is beyond compare. It is a safe choice. I’m typing this on Ubuntu in BlogGTK but missing Windows Live Writer.

Still, Ubuntu is a fantastic deal, especially with Ubuntu One included. I don’t understand the economics by which Canonical can give everyone in the world 2GB of free cloud storage; if it is hoping that enough people will upgrade to the 50GB paid-for version that it will pay for the freeloaders, I fear it will be disappointed.

My point: overall, there is far more right than wrong with Ubuntu in general and Karmic Koala in particular; and I am still happy to recommend it.

Related posts:

1. Ubuntu Karmic Koala breaks Squeezeboxserver
2. Fixing Bluetooth on a Toshiba with Ubuntu
3. Ubuntu Linux on a Tablet PC

Hyper-V R2 has a number of interesting new features including live migration. Another, less exciting but of great interest to folk such as myself who are constantly running trial software, is that dynamically expanding virtual hard drives now perform nearly as well as fixed-size virtual drives. Dynamic drives are far more convenient.

I downloaded Hyper-V 2008 and installed it on a spare machine. The main requirements are a processor that supports hardware virtualization (Intel VT or AMD-V) and hardware Data Execution Prevention (Intel’s Execute Disable Bit or AMD’s NX bit); note that these also have to be enabled in the BIOS.

Once it is up and running you are greeted with a couple of text windows, which feels sparse compared to the usual Windows GUI; but does provide a convenient menu for the things you are likely to want to do next. Actions include naming the computer, joining a domain, downloading updates, adding a local administrator and configuring remote desktop.

Working with Server Core does have some hassles. For example, many third-party drivers and tools come as setup executables that will not run without a GUI. The major vendors should have come to terms with this by now, but it can be a problem particularly with older hardware.

The next step (if you are on Windows 7) is to download and install the Remote server administration tools for Windows 7. Note that after installing, you have to go into Control Panel – Programs – Windows Features and enable the Remote Server Administration Tools, at least including the Hyper-V manager. Then you can run this from the Start menu and connect to your new server.

This step can be problematic. My first attempts failed with RPC permission errors, which I solved by joining the hyper-v server to the Windows domain. If that is not available or desired, there are other fixes.

Other remote admin tools can be useful too. For example, you can connect the Event Viewer to check out the logs.

Once Hyper-V manager is connected, you can create a new virtual machine with a few clicks. I downloaded the latest Ubuntu server iso, copied it to the Hyper-V server, and set it as the DVD drive for the new machine. Started it up, connected, and I was ready to go.

Hyper-V Server is not the only free virtualization platform. Let’s note that completely free platforms also exist – like, indeed, Ubuntu with KVM. I’d also note that VMware is a more mature and advanced platform, despite Hyper-V’s rapid progress.

Still, what you get with Hyper-V server is a polished and easy to use solution that integrates easily with Windows and Active Directory. This is a great deal.

Related posts:

1. Migrating to Hyper-V 2008 R2
2. Hyper-V in Server 2008 RTM doesn’t like non-US locales
3. Hyper-V disk I/O: performance of dynamic vs fixed virtual hard disks

So is “Qt everywhere” a realistic proposition? Here’s a few impressions from the conference. First, a bit of context. Qt is a C++ framework for cross-platform development. and although bindings for other languages exist, Nokia says it is focused on excellence in C++ rather than working with multiple languages. Developers get the advantages of both native code executables and cross-platform support, and Qt is popular on embedded systems as well as desktops and mobile devices.

Qt is an open source framework which was developed by a company called Trolltech which Nokia acquired in 2008. Its motivation, one assumes, was to simplify development for its own multiple operating systems, especially Maemo and Symbian. Still, it has also taken its responsibilities to the open source community seriously. Qt was originally available either under the GPL, which requires developers to make their own applications available under the GPL as well, or under a commercial license. This limited Qt’s take-up. In March Nokia introduced a third option, the LGPL, which is a more liberal and allows commercial development using the free license. The result, we were told, has been a 250% increase in usage (though how this is defined is uncertain) accompanied by “a small drop in revenue.”

Although the revenue decrease is troubling, it is not a disaster for Nokia whose main business is selling hardware; and if take up continues to increase I’d expect revenue to follow.

Since the Nokia acquisition, Qt has been energetically developed. 2009 has seen the release of a dedicated IDE called Qt Creator. I was interested to see a company that has chosen not to go the Eclipse route for its primary IDE, though there are plug-ins for both Eclipse and Visual Studio. The trolls explained that Eclipse came with too much baggage and they wanted something more perfectly suited to its purpose, a lean approach that is in keeping with the Qt philosophy.

Another important move is the inclusion of Webkit within the framework, the same open source HTML engine that powers Apple’s Safari, Adobe AIR, and the browser in numerous Smartphones. Webkit also comes with a Javascript engine, which Nokia is exploiting in several interesting ways.

The big deal at Qt Developer Days was another new project called Kinetic. This is comprised of four parts:

1. An animation API.

2. A state machine.

3. A graphical effects API.

4. A declarative API, currently called QML (Qt Markup Language), though this may change.

Many of these pieces, though not the last, are already present in Qt 4.6, just released in technical preview. Nokia has not announced a specific date for Kinetic, though there were mutters about “first half of 2010”.

The thinking behind Kinetic is to make it easier to support the graphical effects and transitions that users have come to expect, as well as improving the designer-developer workflow – showing that it is not only Adobe and Microsoft who are thinking about this.

QML is significant for several reasons. It is a JavaScript-like API: we were told that Nokia started out with XML but found it cumbersome, and settled on JavaScript instead. It is designed to work well with visual design tools, and Nokia has one code-named Bauhaus which will be part of Qt Creator. Finally, it allows snippets of JavaScript so that developers can create dynamic user interfaces.

At runtime, QML is rendered by a viewer widget, which can be programmatically controlled in C++ just like other Qt widgets.  

Nokia’s hope is that designers can be persuaded to work directly in the QML designer, enabling free exchange of code between designers and developers. It is a nice idea, though I doubt designers will easily transition from the more comfortable world of Photoshop and Flash. However, even if in the end QML is used more by developers than designers, it does greatly simplify the task of creating a dynamic Qt UI. Note that there is already a visual GUI designer in Qt Creator but this is geared towards static layouts.

Long term, who knows, we may see entire applications written in QML, opening up Qt to a new and broader audience.

You can see the latest Qt roadmap here.

Qt pros and cons

I was impressed that attendance here has increased – from around 500 last year to around 700 – despite the economy. Those developers I spoke to seemed to like Qt, praising the way it self-manages memory, though some find the model-view aspect too complex and apparently this is to be improved. Nokia’s stewardship and openness is appreciated and the Qt roadmap generally liked, though there is concern that its understandable focus on mobile may leave the desktop under-served.

Cross-platform capability is increasingly important, and for those who want the performance and capability of C++ along with really good Linux support – important for embedded use – Qt is a strong contender. The focus on mobile is right, not only because of Nokia’s own needs, but because demand for Smartphone apps can only increase.

Integrating with Webkit is a smart move, opening up possibilities for hybrid web/desktop applications and giving Windows developers an alternative to embedded IE with all its quirks.

The open source aspect is another strength. This is now a good selling point if you developing for certain governments (the UK is one such) or other organisations that have a bias towards open source.

That said, talk of Qt everywhere is premature. The mobile space is fractured, and without iPhone, WebOS or Android Nokia cannot claim to have a universal solution. Nor has anyone else; but I’m just back from Adobe MAX where we heard about wider support for the Flash runtime. Then again, few choose between C++ or Flash; Adobe’s runtime is pretty much off the map for attendees here.

Qt is well-established in its niche, and is in good hands. I will be interested to see whether Nokia is successful in broadening its appeal.

Incidentally, if you can get to San Francisco you can still catch Qt Developer Days as it is running there from November 2nd-4th.

Related posts:

1. Is Apple iPhone now unstoppable in the mobile platform wars?
2. Trolltech says Qt for Windows CE coming in May
3. Symbian appeals to Traveling Geeks: develop for our platform

So why was it not what I expected? Well, I thought it might focus on the challenge of developing highly distributed applications, looking at what it takes to make such applications reliable, performant and scalable. I know even from my own dabblings that decisions about where to locate different pieces of logic and handling things such as authentication, transactions and security are tough problems with which many of us need help. However, this title rarely descends to that level of detail – though it is a little unpredictable in that respect. The authors do find space, for example, for several pages on common mistakes in designing web forms, reflecting I imagine their level of frustration at how often they encounter them. Inconsistent, yes; but forgivable.

What you do find is detailed and often brilliant attention to the definition of terms. What is architecture? What is a model? What is a pattern? What is a Mashup? What is a Rich Internet Application? What is Service Oriented Architecture? What is a Reference Architecture? And of course, What is Web 2.0? These discussions are pithy, illustrated with real-world examples, and convey the authors’ enthusiasm for the new era which is now upon us.

As you read the early chapters you feel that this is all leading up to something big. That’s correct, but again it might not be what you expect. In fact, this feels like an introduction to a much longer book that is yet to be written, a fault which can only be excused by the idea that it is you, not they, that have the task of writing it.

Here’s a short extract from a section called The Timeless Way of Building Software: Inspiration for the next generation of Web Software:

Take a look at some of the great web software, such as Google’s search page, Flickr, or Delicious. They all have that unique yet unhurried grace that comes from perfect ease and natural balance. But what do they have in common exactly? They are beautiful, ordered and harmonious. But especially, and what strikes to the heart, is that they live. They are the perfect blend of logic and emotion, of the machine and the human experience, intrinsically intertwined and symbiotically embracing each other.

If you find this excessively rhapsodic you will probably not like this book. Note the title though: this is a Web 2.0 book as well as an architecture book; and a couple of pages on from the above quote we learn the significance of that:

…this timeless way of software architecture represents a more complete reconceptualization of what we thought we knew. We can give this place, where common sense and logic collide with humanity and society, a name.

We call that place Web 2.0.

And that is pretty much how it ends.

Overall it’s a great read, written I suspect in haste, accounting for the odd sloppy phrase or two, but always fluent and enjoyable. Recommended for those who are thoughtful about what it means to do software today.

View on Amazon.com: Web 2.0 Architectures: What entrepreneurs and information architects need to know

View on Amazon.co.uk: Web 2.0 Architectures: What entrepreneurs and information architects need to know

Related posts:

1. Book Review: IronPython in Action
2. How to code better: new book from ThoughtWorks
3. Free preview chapter of Adam Nathan’s WPF book

<meta http-equiv="X-UA-Compatible" content="chrome=1">

The concept is not so different from Mozilla’s Screaming Monkey, which replaces the IE JavaScript engine with its own, except that Chrome Frame has its own HTML renderer as well. The idea is that web developers need no longer be constrained to HTML features which IE supports. IE users do not need to change their browser; just install the plug-in.

The question though: why bother? Why not just ask users to install Chrome itself and use it to browse your site? Chrome on Windows is a great piece of work, and I use it regularly, whereas there’s something odd about using one browser embedded within another.

I suppose there is some small subset of users who have to use IE at work for compatibility with some application, but are allowed to install plug-ins; or others who are brave enough to install an add-in, but feel somehow more comfortable sticking with the default and official Windows web browser.

Still, I’m not sure that there are many users in that category. It also turns out that the Chrome Frame concept has some problems, as I discovered when I tried it out. Note that this is an “early-stage release”, so some things can be expected not to work. Even so, some of the issues cast doubt on whether this a sensible approach. The problem area is integration. For example, when Chrome Frame is active, should the browser announce itself as IE, or as Chrome? When you click a link, should it open in Chrome Frame, or in native IE? When you have a plug-in installed in IE, would you expect it to work in Chrome Frame as well?

The answers currently are that the browser still announces itself as IE, though chromeframe is added to the User-Agent header. Links open by default in IE even if Chrome Frame is active, and plug-ins like Adobe Flash do not work in Chrome Frame even if they are installed in IE. Overall, if you want a page to run well in Chrome Frame, you really need to code for it specifically. In other words, IE+Chrome Frame is yet another browser variation to worry about.

Fortunately, Chrome Frame only kicks in when activated by the tag above, or forced by prefixing the URL with cf; and if users do the latter, they should expect trouble. When I tried it, I had a disappointing experience visiting Adobe.com in Chrome Frame, even though Flash was installed and up-to-date. I’ve included the right-click menu to show that this is Chrome Frame:

Gmail also has problems, probably because it is delivering markup designed for IE:

Of course Chrome Frame is not intended to be used like this; but it does demonstrate the challenges that exist if you decide to target Chrome Frame. In compensation, you get the HTML 5 goodness and fast JavaScript that Google has in its browser.

This feels like a last resort. I doubt network admins will welcome Chrome Frame, since it does undermine one of the reasons for using IE: that the system browser is kept up-to-date by Microsoft with security patches using Windows update or Windows Server Update Services. I’m not tempted to keep it installed myself.

The opposite approach strikes me as better, where other browsers emulate IE by embedding it, as in IE Tab for Firefox.

Still, while I’m sceptical of the technical merits of Chrome Frame, it is a great PR move. It puts pressure on Microsoft to implement more HTML 5 features and speed up its Javascript engine, or risk the embarrassment of sites which require IE users either to switch browsers, or to suffer this add-in.

Update: see also Browser Soup and Chrome Frame by Mozilla’s Mitchell Baker, who does a good job of explaining why this is a bad idea – though bear in mind that she works for a competitor.

Related posts:

1. Google Chrome: the developer angle
2. Want Google Earth in your browser? Don’t use Google Chrome.
3. Counting primes in Google Chrome