OpenAI Atlas and the blurred line between search and synthesis

Posted on By Mark Wilson

OpenAI’s new Atlas browser has certainly got people talking.

Some are excited — calling it a “Google killer” and a glimpse of how we’ll all navigate the web in future. Others are alarmed — pointing to privacy concerns, data collection prompts, and the idea of handing over browsing history and passwords to an AI company.

Jason Grant described his experience as “a giant dark pattern.” Matthew Dunn was more balanced — impressed by the features, but quick to warn businesses off using it. He’s right: if you wouldn’t paste confidential data into ChatGPT, you probably shouldn’t browse the company intranet through Atlas either.

Search vs. synthesis

When people say Atlas will replace Google, they’re missing the point. It’s not search in the traditional sense.

A search engine indexes existing content and returns links that might answer your question. Atlas — and systems like it — go a step further. They synthesise an answer, combining what’s on the web with what’s in your conversation and what they’ve “seen” before.

As Data Science Dojo explains, search engines are designed to find information that already exists, while synthesis engines are designed to create new information.

Or, as Vincent Hunt neatly puts it: “Search gives you links. Synthesis gives you insight.”

That shift sounds subtle, but it changes everything: how we ask questions, how we evaluate truth, and how much we trust the output.

As I said in my recent talk on AI Transformation at the Bletchley AI User Group, “Generative AI is not a search engine. It doesn’t retrieve facts. It generates language based on probabilities.” Google doesn’t know the truth either — it just gives you the most common answer to your question — but AI goes a step further. It merges, rewrites and repackages information. That can be powerful, but it’s also risky. It’s why I believe the AI-generated results that many search engines now return as default are inferior to traditional results, based on actual information sources.

Without strong governance, AI may be repurposing outdated content or drawing on biased data. Transparency matters — because trust is the real currency of AI adoption.

Why Atlas matters

In OpenAI’s announcement, Atlas is described as “bringing ChatGPT anywhere across the web — helping you in the window right where you are.”

It’s not just a search bar. It can summarise pages, compare options, fill out forms, or even complete tasks within websites. That’s a very different paradigm — one where the browser becomes a workspace, and the assistant becomes a collaborator.

A step towards agentic AI?

So, is Atlas really agentic? In part, yes.

Agentic AI describes systems that can act rather than just answer. They plan, execute and adapt — working on your behalf, not just waiting for your next prompt.

OpenAI’s own notes mention an agent mode that can “help you book reservations or edit documents you’re working on,” as reported by The Verge.

Others, like Practical Ecommerce, describe Atlas as “a push into agentic browsing — where the browser is now an AI agent too.”

It’s not full autonomy yet — more like assisted agency — but it’s a clear step in that direction.

Why it still needs caution

As exciting as it sounds, Atlas isn’t designed for enterprise use. It raises valid concerns about data privacy, security, and trust. You wouldn’t give a work browser access to sensitive credentials, and the same logic applies here.

As Matthew Dunn notes, ChatGPT “produces better output than Copilot, but with less security and privacy.” That’s a fair trade-off for some users, but not for organisations handling confidential information.

So, by all means, explore it — but do so with your eyes open.

And yes, I’ll still give it a try I decided not to install it after all

For all the justified concerns about privacy and data handling, I’ll still give Atlas a try. Even though I have Copilot at work, I pay for ChatGPT Pro for activities that are not directly related to my confidential work.

Atlas might extend that usefulness into how I browse, not just how I prompt. The key, as ever, is knowing what data you’re sharing — and making that a conscious choice, not an accidental one.

[Updated 24/10/2025: After writing and publishing this post, I decided not to install Atlas. There are a lot of security concerns about the way the browser stores local data, which may easily be exploited. Nevertheless, both OpenAI Atlas and Perplexity Comet are interesting developments, and the narrative about the differences between an AI search (synthesis) and a traditional search is still valid.]

Featured image: created by ChatGPT.

Weeknote 2024/06: more playing with NFC; thoughts on QR code uses; and a trip to AWS’ UK HQ

Posted on By Mark Wilson
This content is 2 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

Last week’s weeknote taught me one of two things. Either I’m getting boring now; or AI fatigue has reached a level where people just read past anything with ChatGPT in the title. Or maybe it was just that the Clippy meme put people off…

Whilst engagement is always nice, I write these weeknotes for mindful reflection. At least, that’s what I tell myself when I’m writing them. There’s also a part of me that says “you’ve done six weeks now… don’t stop and undo all that work”. Hmm, Sunk Cost Fallacy anyone?

So, let’s get stuck into what’s been happening in week 6 of 2024… there seems to be quite a lot here (or at least it took me a few hours to write!)

This week at work

Even with the input from ChatGPT that I mentioned last week, I’m still struggling to write data sheets. Maybe this is me holding myself back with my own expectations around the output. It’s also become a task that I simply must complete – even in draft – and then hand over to others to critique. Perfection is the enemy of good, and all that!

I’m also preparing to engage with a new client to assist with their strategy and innovation. One challenge is balancing the expectations of key client stakeholders, the Account Director, and the Service Delivery Manager with my own capabilities. In part, this is because expectations have been based on the Technical Architect who is aligned to the account. He’s been great on the technical side but I’m less hands-on and the value I will add is more high-level. And this is a problem of our own making – everyone has a different definition of what an (IT) Architect is. I wrote about this previously:

What’s needed are two things – a really solid Technical Architect with domain expertise, and someone who can act as a client side “CTO”. Those are generally different skillsets.

My work week ended with a day at Amazon Web Services (AWS). I spend a lot of time talking about Microsoft Azure, but my AWS knowledge is more patchy. With a multi-cloud mindset (and not just hybrid with Node4), I wanted to explore what’s happening in the world of AWS. More on that in a bit…

This week in tech

Let’s break this up into sections as we look at a few different subjects…

More fun with NFC tags

A few weeks ago, I wrote about the NFC tags I’d been experimenting with. This week I took it a bit further with:

  1. Programming tags using the NFC Tools app. This means the tag action doesn’t rely on an iOS Shortcut and so isn’t limited to one user/device. Instead, the tag has a record stored in its memory that corresponds to an action – for example it might open a website. I was going to have a tag for guests to automatically connect to the guest Wi-Fi in our house but iOS doesn’t support reading Wi-Fi details from NFC (it’s fine with a QR code though… as I’ll discuss in a moment).
  2. Using a tag and an automation to help me work out which bins to put out each week. Others have said “why not just set a recurring reminder?” and that is what I do behind the scenes. The trouble with reminders is notifications. Instead of the phone reminding me because it’s the right day (but perhaps I’m in the wrong place), I can scan and check which actions are needed this week.

QR codes are not the answer to sharing every link…

Yesterday, I couldn’t help but notice how many QR codes featured in my day. Unlike most of my recent journeys, my train ticket didn’t have a code. This is because Thameslink (the train operating company for my train from Bedford to London) appears to be stuck on an old technology stack. Their app is pretty useless and sends me to their website to buy tickets, which I then have to collect from a machine at the station. If I need to collect a ticket I might as well buy it on the day from the same machine (there are no Advance discounts available on my journey). So, paper train tickets with magnetic stripes it was.

Then, I was networking with some of the other delegates at the AWS re:Invent re:Cap event and found that people share QR codes from the LinkedIn app now. How did I not know this was a thing? (And to think I am playing with programming NFC tags to do cool things.) To be fair, I haven’t got out much recently – far too much of my post-pandemic work for risual was online. I even have paper business cards in my work bag. I don’t think I’ve given one to anyone in a long time though…

But QR codes were everywhere at AWS. They were In every presentation for links to product information, feedback links, even for the Wi-Fi in the room. And that’s the problem – QR codes are wonderful on a mobile device. But all too often someone creates a code and says “let’s share this – it will be cool”, without thinking of the use case.

  • A QR code for exchanging details in person. Yep, I get that.
  • A QR code on physical marketing materials to direct people to find out more. That works.
  • A QR code on an email. Get real. I’m reading it on one device – do you really want me to get another one to scan the code?
  • A QR code on the back of a van. Nice in principle but it’s a moving vehicle. Sometimes it won’t work so better to have a URL and phone number too. In which case what purpose does the QR code serve?
  • Multiple QR codes on a presentation slide. Hmm… tricky now. The camera app’s AI doesn’t know which one to use. What’s wrong with a short URL? Camera apps can usually recognise and scan URLs too.
  • QR codes for in-room Wi-Fi. Seems great at first, and worked flawlessly on my phone but I couldn’t get them to work on a Windows laptop. Well, I could read them in the camera app but it wouldn’t let me open the URL (or copy it to examine and find the password). For that I needed an app from the Microsoft Store. And I was offline. Catch 22. Luckily, someone wrote the password on a white board. Old skool. That works for me.

More of my tech life

  • I think Apple might have launched a VR headset. This is the meme that keeps on giving…

That visit to the AWS offices that I mentioned earlier…

I started writing this on the train home, thinking there’s a lot of information to share. So it’s a brief summary rather than trying to include all the details:

  • The AWS event I attended was a recap of the big re:Invent conference that took place a few months ago. It took place at AWS’s UK HQ in London (Holborn). I’ve missed events like this. I used to regularly be at Microsoft’s Thames Valley Park (Reading) campus, or at a regional Microsoft TechNet or MSDN event. They were really good, and I knew many of the evangelists personally. These days, I generally can’t get past the waitlist for Microsoft events and it seems much of their budget is for pre-recorded virtual events that have huge audiences (but terrible engagement).
  • It was a long day – good to remind me why I don’t regularly commute – let alone to London. But it was great to carve out the time and dedicate it to learning.
  • Most of the day was split into tracks. I could only be in one place at one time so I skipped a lot of the data topics and the dedicated AI/ML ones (though AI is in everything). I focused on the “Every App” track.
  • A lot of the future looking themes are similar to those I know with Microsoft. GenAI, Quantum. The product names are different, the implementation concepts vary a little. There may be some services that one has and the other doesn’t. But it’s all very relatable. AWS seems a little more mature on the cost control front. But maybe that’s just my perception from what I heard in the keynote.
  • The session on innovating faster with Generative AI was interesting – if only to understand some of the concepts around choosing models and the pitfalls to avoid.
  • AWS Step Functions seem useful and I liked the demo with entertaining a friend’s child by getting ChatGPT to write a story then asking Dall-E to illustrate it.
  • One particularly interesting session for me was about application modernisation for Microsoft workloads. I’m not a developer, but even I could appreciate the challenges (e.g. legacy .NET Framework apps), and the concepts and patterns that can help (e.g strangler fig to avoid big bang replacement of a monolith). Some of the tools that can help looked pretty cool to.
  • DeepRacer is something I’d previously ignored – I have enough hobbies without getting into using AI to drive cars. But I get it now. It’s a great way to learn about cloud, data analysis, programming and machine learning through play. (Some people doing like the idea of “play” at work, so let’s call it “experimentation”).
  • There’s some new stuff happening in containers. AWS has EKS and ECS. Microsoft has AKS and ACS. Kubernetes (K8s) is an orchestration framework for containers. Yawn. I mean, I get it, and I can see why they are transformative but it seems every time I meet someone who talks about K8s they are evangelical. Sometimes containers are the solution. Sometimes they are not. Many of my clients don’t even have a software development capability. Saying to an ISV “we’re going to containerise your app” is often not entertained. OK, I’ll get off my soapbox now.
  • One thing AWS has that I’ve never heard Azure folks talk about is the ability to deliberately inject chaos into your app or infrastructure – so the session on the AWS Fault Injection Service was very interesting. I particularly like the ideas of simulating an availability zone outage or a region outage to test how your app will really perform.
  • Amazon has a contact centre platform called Connect. I did not know that. Now I do. It sounds quite interesting, but I’m unlikely to need to do anything more with it at Node4 – Microsoft Teams and Cisco WebEx are our chosen platforms.
  • The security recap was… a load of security enhancements. I get it. And they seem to make sense but they are also exactly what I would expect to see.
  • Amazon Security Lake is an interesting concept, but I had to step out of that session. It did make me wonder if it’s just SIEM (like Microsoft Sentinel). Apparently not. ASL is a data lake/log management system not a SIEM service, so bring your own security analytics.

In all, it was a really worthwhile investment of a day. I will follow up on some of the concepts in more detail – and I plan to write about them here. But I think the summary above is enough, for now.

This week’s reading, writing, watching and listening

I enjoy Jono Hey’s Sketchplanations. Unfortunately. when I was looking for one to illustrate the Sunk Cost Fallacy at the top of this post, I couldn’t find one. I did see there’s I see he has a book coming out in a few months’ time though. You can pre-order it at the place that does everything from A-Z.

What I did find though, is a sketch that could help me use less passive voice in these blog posts:

Inspired by something I saw on the TV, and after I found my previous notes, some of my thoughts here grew into a post of their own: Anti-social media.

My wife and I finished watching Lessons In Chemistry on Apple TV this week. I commented previously that one of my observations was we still have a long way to go on diversity, inclusion and equality but we’ve come a long way since the 1950s. And then I read this, from the LA Times Archive, reporting on how a woman was jailed for contempt of court after the Judge took offence to her wearing “slacks”, in 1938.

This week in photos

  • Only one from my instragram this week:
  • This isn’t mine, but I love it…
  • Also:
  • And what about this?

This week at home

Putting home (and therefore family) at the end seems wrong, but the blog is about tech first, business second, and my personal life arguably shouldn’t feature so often.

The positive side of trying to be in the office at least a day or two a week is that I can do the school run. I may only have one “child” still at school but he’s learning to drive, so he can drive to school and I’ll continue to drive to work afterwards. He’s also driving to his hockey training and matches so its a good way to build experience before his driving test in a few months’ time.

Next week, my adult son (Matt) heads back to Greece for a couple of months’ cycle training. He’s also building new gravel/cyclocross bikes for later in the year, so “bits of bike” keep on appearing in the dining room… including some new wheels from one of the team sponsors, FFWD Wheels.

Meanwhile, my wife is very excited because Matt will be invited to Buckingham Palace to receive his Duke of Edinburgh Gold Award. He can take a guest, hence Mrs W’s excitement. Let’s just hope he’s in the country at the time.

I really should try and use the time whilst he’s away to get out on my own bike as my own fitness is not where it should be.

That’s all for this week. See you all around the same time next week?

Featured image: author’s own.

Android for under-13s: no Google accounts; no family sharing

Posted on By Mark Wilson
This content is 9 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

We’re entering a new phase in the Wilson family as my eldest son starts secondary school next week and my youngest becomes more and more tech-aware.

The nearly-10 year-old just wants a reasonably-priced, reasonably-specced tablet as my original iPad is no longer suiting his needs (stuck back on iOS 5 and with a pretty low spec by today’s standards) – I’m sure we’ll work something out.

A bigger challenge is a phone for the nearly-12 year-old. We’ve said he can have his own smartphone when his birthday comes and effectively there are 3 (well, 2) platforms to consider:

  • Windows Mobile: limited app availability; inexpensive handsets; uncertain future.
  • Apple iOS: expensive hardware; good app support.
  • Google Android: wide availability of apps and hardware; fragmented OS.

Really, Windows isn’t an option (for consumers – different story in the enterprise); Apple is only viable if he has a hand-me-down device (which is a possibility); but he’s been doing his research, and is looking at price/specs for various Android devices.  The current favourite is an Elephone P9000 – which looks like a decent phone for a reasonable price – as long as I can find a reliable UK supplier (i.e. not grey market).

In the meantime, and to see how he gets on before we commit to a device purchase, I’ve given him an old Samsung Galaxy S3 Mini that I had in a drawer and I put a giffgaff SIM in. Because it’s a Google device, he gets the best experience if he uses a Google account… and that’s where the trouble started.

We went to sign-up, added some details, and promptly found that you have to be 13 to open a Google account. And unlike with Apple iCloud Family Sharing, where I have family sharing set up for the old iPhones that the boys use around the house, the Google equivalent (Google Play Family Library) also needs all of the family members to be at least 13.  There simply appears to be no option for younger children to use Google services.

Maybe that’s because Google’s primary business is about selling advertising and advertising to children is questionable from a moral standpoint (though YouTube have come up with a child-friendly product).

I tried signing in as me – which let me download some apps but also meant he had access to my information – like all of my contacts (easily switched off but still undesirable).

Luckily, it seems I created him a GMail account when he was 5 weeks old (prescient, some might say) and I was able to find my way into that and get him going. Sadly, it seems I was not as mentally sharp when his little brother was born…

(As an aside, I originally gave my son a Nokia “feature phone” to use and he looked bemused – he later confessed that was because he didn’t know how to use it!)

Postscript: I’ve since given my youngest son my Tesco Hudl and was able to sign up for a Google account without being asked to provide date of birth details…

Short takes: Windows Phone screenshots and force closing apps; Android static IP

Posted on By Mark Wilson
This content is 11 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

I’m clearing down my browser tabs and dumping some of the the things I found recently that I might need to remember again one day!

Taking a screenshot on Windows Phone

Windows Phone 7 didn’t have a screenshot capability (I had an app that worked on an unlocked phone) but Windows Phone 8 let me take screenshots with Windows+Power. For some reason this changed in Windows Phone 8.1 to Power+Volume Up.  It does tell you when you try to use the old combination but, worth noting…

Some search engines are more helpful than others

Incidentally, searching for this information is a lot more helpful in some search engines than in others…

One might think Microsoft could surface it’s own information a little more clearly in Bing but there are other examples too (Google’s built-in calculator, cinema listings, etc.)

Force-quitting a Windows Phone app

Sometimes, apps just fail. In theory that’s not a problem, but in reality they need to be force-closed.  Again, Windows Phone didn’t used to allow this but recent updates have enabled a force-close. Hold the back button down, and then click the circled X that appears in order to close the problem app.

Enabling a static IP on an Android device

Talking of long key presses… I recently blew up my home infrastructure server (user error with the power…) and, until I sort things out again, all of our devices are configured with static IP configurations. One device where I struggled to do this was my Hudl tablet, running Android. It seems the answer is to select the Wi-Fi connection I want to use, but to long-press it, at which point there are advanced options to modify the connection and configure static IP details.

Google Reader is retired next week – have you switched to Feedly?

Posted on By Mark Wilson
This content is 12 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

Next week, Google is set to retire Google Reader. When I wrote this post (back in March), almost 75% of the subscribers to my feed (already dwindling, partly as a result of Google algorithm changes that seem to penalise independent views in favour of branded content) came via Google Feedfetcher (used by Reader to grab RSS or Atom feeds), suggesting that lots of you use Google Reader.

Hopefully you’ve all found a way to move forward but, if you haven’t, I recommend checking out Feedly.

If you migrate before Google turns off Reader, it’s a one-click migration (just log into Feedly with your Google account) – I did it weeks ago and haven’t looked back since!

Here are a couple of links that might be useful:

Now I need to look at moving my site’s RSS away from Feedburner, before Google kills that off too (I’m sure it’s only a matter of time…)

The annotated world – the future of geospatial technology? (@EdParsons at #DigitalSurrey)

Posted on By Mark Wilson
This content is 13 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

Tonight’s Digital Surrey was, as usual, a huge success with a great speaker (Google’s @EdParsons) in a fantastic venue (Farnham Castle).  Ed spoke about the future of geospatial data – about annotating our world to enhance the value that we can bring from mapping tools today but, before he spoke of the future, he took a look at how we got to where we are.

What is geospatial information? And how did we get to where we are today?

Geospatial information is very visual, which makes it powerful for telling stories and one of the most famous and powerful images is that of the Earth viewed from space – the “blue marble”. This emotive image has been used many times but has only been personally witnessed by around 20 people, starting with the Apollo 8 crew, 250000 miles from home, looking at their own planet. We see this image with tools like Google Earth, which allows us to explore the planet and look at humankind’s activities. Indeed about 1 billion people use Google Maps/Google Earth every week – that’s about a third of the Internet population, roughly equivalent to Facebook and Twitter combined [just imagine how successful Google would be if they were all Google+ users…]. Using that metric, we can say that geospatial data is now pervasive – a huge shift over the last 10 years as it has become more accessible (although much of the technology has been around longer).

The annotated world is about going beyond the image and pulling out info otherwise invisible information, so, in a digital sense, it’s now possible to have map of 1:1 scale or even beyond. For example, in Google Maps we can look at StreetView and even see annotations of buildings. This can be augmented with further information (e.g restrictions in the directions in which we can drive, details about local businesses) to provide actionable insight. Google also harvests information from the web to create place pages (something that could be considered ethically dubious, as it draws people away from the websites of the businesses involved) but it can also provide additional information from image recognition – for example identifying the locations of public wastebins or adding details of parking restrictions (literally from text recognition on road signs). The key to the annotated web is collating and presenting information in a way that’s straightforward and easy to use.

Using other tools in the ecosystem, mobile applications can be used to easily review a business and post it via Google+ (so that it appears on the place page); or Google MapMaker may be used by local experts to add content to the map (subject to moderation – and the service is not currently available in the UK…).

So, that’s where we are today… we’re getting more and more content online, but what about the next 10 years?

A virtual (annotated) world

Google and others are building a virtual world in three dimensions. In the past, Google Earth pulled data from many sets (e.g. building models, terrain data, etc.) but future 3D images will be based on photographs (just as, apparently, Nokia have done for a while). We’ll also see 3D data being using to navigate inside buildings as well as outside. In one example, Google is working with John Lewis, who have recently installed Wi-Fi in their stores – to use this to determine a user’s location determination and combine this with maps to navigate the store. The system is accurate to about 2-3 metres [and sounds similar to Tesco’s “in store sat-nav” trial] and apparently it’s also available in London railway stations, the British Museum, etc.

[blackbirdpie url=”https://twitter.com/markwilsonit/status/228568510723411968″]

Ed made the point that the future is not driven by paper-based cartography, although there were plenty of issues taken with this in the Q&A later, highlighting that we still use ancient maps today, and that our digital archives are not likely to last that long.

Moving on, Ed highlighted that Google now generates map tiles on the fly (it used to take 6 weeks to rebuild the map) and new presentation technologies allow for client-side rendering of buildings – for example St Pauls Cathedral, in London. With services such as Google Now (on Android), contextual info may be provided, driven by location and personality

With Google’s Project Glass, that becomes even more immersive with augmented reality driven by the annotated world:

Although someone also mentioned to me the parody which also raises some good points:

Seriously, Project Glass makes Apple’s Siri look way behind the curve – and for those who consider the glasses to be a little uncool, I would expect them to become much more “normal” over time – built into a normal pair of shades, or even into prescription glasses… certainly no more silly than those Bluetooth earpieces the we used to use!

Of course, there are privacy implications to overcome but, consider what people share today on Facebook (or wherever) – people will share information when they see value in it.

Big data, crowdsourcing 2.0 and linked data

At this point, Ed’s presentation moved on to talk about big data. I’ve spent most of this week co-writing a book on this topic (I’ll post a link when it’s published) and nearly flipped when I heard the normal big data marketing rhetoric (the 3 Vs)  being churned out. Putting aside the hype, Google should know quite a bit about big data (Google’s search engine is a great example and the company has done a lot of work in this area) and the annotated world has to address many of the big data challenges including:

  • Data integration.
  • Data transformation.
  • Near-real-time analysis using rules to process data and take appropriate action (complex event processing).
  • Semantic analysis.
  • Historical analysis.
  • Search.
  • Data storage.
  • Visualisation.
  • Data access interfaces.

Moving back to Ed’s talk, what he refers to as “Crowdsourcing 2.0” is certainly an interesting concept. Citing Vint Cerf (Internet pioneer and Google employee), Ed said that there are an estimated 35bn devices connected to the Internet – and our smartphones are great examples, crammed full of sensors. These sensors can be used to provide real-time information for the annotated world: average journey times based on GPS data, for example; or even weather data if future smartphones were to contain a barometer.

Linked data is another topic worthy of note, which, at its most fundamental level is about making the web more interconnected. There’s a lot of work been done into ontologies, categorising content, etc. [Plug: I co-wrote a white paper on the topic earlier this year] but Google, Yahoo, Microsoft and others are supporting schema.org as a collection of microformats, which are tags that websites can use to mark up content in a way that’s recognised by major search providers. For example, a tag like <span itemprop="addresscountry">Spain</span> might be used to indicate that Spain is a country with further tags to show that Barcelona is a city, and that Noucamp is a place to visit.

Ed’s final thoughts

Summing up, Ed reiterated that paper maps are dead and that they will be replaced with more personalised information (of which, location is a component that provides content). However, if we want the advantages of this, we need to share information – with those organisations that we trust and where we know what will happen with that info.

Mark’s final thoughts

The annotated world is exciting and has stacks of potential if we can overcome one critical stumbing point that Ed highliughted (and I tweeted):

[blackbirdpie url=”https://twitter.com/markwilsonit/status/228571728740245504″]

Unfortunately, there are many who will not trust Google – and I find it interesting that Google is an advocate of consuming open data to add value to its products but I see very little being put back in terms of data sets for others to use. Google’s argument is that it spent a lot of money gathering and processing that data; however it could also be argued that Google gets a lot for free and maybe there is a greater benefit to society in freely sharing that information in a non-proprietary format (rather than relying on the use of Google tools). There are also ethical concerns with Google’s gathering of Wi-Fi data, scraping website content and other such issues but I expect to see a “happy medium” found, somewhere between “Don’t Be Evil” and “But we are a business after all”…

Thanks as always to everyone involved in arranging and hosting tonight’s event – and to Ed Parsons for an enlightening talk!

Useful to know: Google Chrome has its own task manager

Posted on By Mark Wilson
This content is 14 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

Earlier today, I was wondering why I was seeing a “missing plug-in” message in Google Chrome on a number of websites that I regularly view. I loaded the same websites in Internet Explorer and they worked OK, so something had obviously gone screwy inside Chrome. I could have guessed – it was Flash, although normally I get a yellow bar to tell me that has stopped working.

I rebooted my PC yesterday, so I don’t plan to do that again for another couple of weeks (until the memory leak that one of my apps has gets so bad that I’m forced to…) but I googled missing plug-in google chrome to see what comes up. As it happens, Chrome has a task manager built in (press shift and escape).  After ending the Shockwave Flash process, I refreshed the offending page(s) and everything worked as it should.

By then I was intrigued by the stats for nerds link which takes me to chrome://memory-redirect/ – an internal page that contains a breakdown of activity by process (including which tabs are managed by which processes) – which would have been handy to know about when Chrome had gobbled up a good chunk of my RAM earlier this week:

[blackbirdpie url=”https://twitter.com/#!/markwilsonit/status/167195412262961152″]

If anyone knows a similar memory management function for Internet Explorer, I’d be pleased to hear it as the relationship between tabs and processes seems to be a black art (and it may help to chase down problematic tabs) – I’ve tried Process Explorer and Windows Task Manager in the past, but it would be useful IE functionality…

Designing a private cloud infrastructure

Posted on By Mark Wilson
This content is 14 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

A couple of months ago, Facebook released a whole load of information about its servers and datacentres in a programme it calls the Open Compute Project. At around about the same time, I was sitting in a presentation at Microsoft, where I was introduced to some of the concepts behind their datacentres.  These are not small operations – Facebook’s platform currently serves around 600 million users and Microsoft’s various cloud properties account for a good chunk of the Internet, with the Windows Azure appliance concept under development for partners including Dell, HP, Fujitsu and eBay.

It’s been a few years since I was involved in any datacentre operations and it’s interesting to hear how times have changed. Whereas I knew about redundant uninterruptible power sources and rack-optimised servers, the model is now about containers of redundant servers and the unit of scale has shifted.  An appliance used to be a 1U (pizza box) server with a dedicated purpose but these days it’s a shipping container full of equipment!

There’s also been a shift from keeping the lights on at all costs, towards efficiency. Hardly surprising, given that the IT industry now accounts for around 3% of the world’s carbon emissions and we need to reduce the environmental impact.  Google’s datacentre design best practices are all concerned with efficiency: measuring power usage effectiveness; measuring managing airflow; running warmer datacentres; using “free” cooling; and optimising power distribution.

So how do Microsoft (and, presumably others like Amazon too) design their datacentres? And how can we learn from them when developing our own private cloud operations?

Some of the fundamental principles include:

  1. Perception of infinite capacity.
  2. Perception of continuous availability.
  3. Drive predictability.
  4. Taking a service provider approach to delivering infrastructure.
  5. Resilience over redundancy mindset.
  6. Minimising human involvement.
  7. Optimising resource usage.
  8. Incentivising the desired resource consumption behaviour.

In addition, the following concepts need to be adopted to support the fundamental principles:

  • Cost transparency.
  • Homogenisation of physical infrastructure (aggressive standardisation).
  • Pooling compute resource.
  • Fabric management.
  • Consumption-based pricing.
  • Virtualised infrastructure.
  • Service classification.
  • Holistic approach to availability.
  • Computer resource decay.
  • Elastic infrastructure.
  • Partitioning of shared services.

In short, provisioning the private cloud is about taking the same architectural patterns that Microsoft, Amazon, et al use for the public cloud and implementing them inside your own data centre(s). Thinking service, not server to develop an internal infrastructure as a service (IaaS) proposition.

I won’t expand on all of the concepts here (many are self-explanitory), but some of the key ones are:

  • Create a fabric with resource pools of compute, storage and network, aggregated into logical building blocks.
  • Introduced predictability by defining units of scale and planning activity based on predictable actions (e.g. certain rates of growth).
  • Design across fault domains – understand what tends to fail first (e.g. the power in a rack) and make sure that services span these fault domains.
  • Plan upgrade domains (think about how to upgrade services and move between versions so service levels can be maintained as new infrastructure is rolled out).
  • Consider resource decay – what happens when things break?  Think about component failure in terms of service delivery and design for that. In the same way that a hard disk has a number of spare sectors that are used when others are marked bad (and eventually too many fail, so the disk is replaced), take a unit of infrastructure and leave faulty components in place (but disabled) until a threshold is crossed, after which the unit is considered faulty and is replaced or refurbished.

A smaller company, with a small datacentre may still think in terms of server components – larger organisations may be dealing with shipping containers.  Regardless of the size of the operation, the key to success is thinking in terms of services, not servers; and designing public cloud principles into private cloud implementations.

Office 365 message filtering (and a horrible little bug that leaves email addresses exposed…)

Posted on By Mark Wilson
This content is 15 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

One of my concerns with my recent switch from Google Apps Mail to Microsoft Office 365 was about spam email. You see, I get none.  Well, when I say I get none, I get plenty but it’s all trapped for me. With no effort on my part. Only a handful of missed spam messages in the last 2 or 3 years and almost as few false positives too.

I’ve had the same email address for about 12 years now (I think), and it’s been used all over the web. Some of my friends are more particular though – and, perhaps understandably, were annoyed when I accidentally emailed around 40 people with e-mail addresses visible in the To: field today. Except that I hadn’t intended to.

I think I’ve found a bug in Office 365’s Outlook Web App (at least, I hope it’s not closed as “by design”, assuming I find out how to file a bug report). If I send to a distribution group, it automatically expands the addresses and displays them to all recipients. That’s bad.

The annoying thing is that, previously, I had been BCCing the recipients. I have a feeling that at least one organisation was rejecting my mail because there was nothing in the To: field (although it didn’t like Google’s propensity to send mail from one domain “on behalf of” another address either), so I thought I’d use a list instead and the recipients would see the list name, rather than the actual email addresses. Thankfully it was only sent to my closest freinds and family (although that’s not really the point).

So, back to spam and Office 365 – does it live up to my previous experience with Google Apps Mail? Actually, yes I think it does. I’ve had to teach it a couple of safe senders and block a couple of others, but it really was just a handful and it’s settled down nicely.

All of Microsoft’s cloud-based e-mail services use Forefront Online Protection for Exchange. Enterprise administrators have some additional functionality (adapting SCL thresholds, etc.) but things seem to be working pretty well on my small business account too. Digging around in the various servers that the mail passes through sees hosts at bigfish.com and frontbridge.com – Frontbridge was an aquisition that has become part of Exchange Hosted Services (and it started out as Bigfish Communications) – so the technology is established, and another Microsoft property (Hotmail) is a pretty good test bed to find and filter the world’s spam.

Attempting to track RSS subscribers on a WordPress blog

Posted on By Mark Wilson
This content is 15 years old. I don't routinely update old blog posts as they are only intended to represent a view at a particular point in time. Please be warned that the information here may be out of date.

As well as my own website (which has precious little content these days due to my current workload), I also manage the Fujitsu UK and Ireland CTO Blog. Part of that role includes keeping an eye on a number of metrics to make sure that people are actually interested in what we have to say (thankfully, they seem to be…). Recently though, I realised that, whilst I’m tracking visitors to the blog, I’m missing hits on the RSS feed (because it’s not actually a page with the tracking script included) - and that’s a problem.

There are ways around this (I use Google Feedburner on my own blog, or it’s possible to put a dummy page with a meta refresh in front of the feed to pick up some metrics) but they have their own issues (for example the meta refresh methods breaks autodiscovery for some RSS readers) and will only help with new subscribers going forwards, not with my legacy issue of how many subscribers do I have right now.

There is another approach though: using a popular web-based RSS subscription service like Google Reader to see how many subscribers it tracks for our feed (the same metrics are available from Google’s Webmaster Tools).  The trouble is, that’s not all of the subscribers (for example, a good chunk of people use Outlook to manage their feeds, or other third-party RSS readers). If I use my own blog as an example, Google Reader shows that I have 247 subscribers but Feedburner says I have 855.  Those subscribers come from all manner of feed readers and aggregators, email subscription services and web browsers (Firefox accounts for almost 20% of them) so it’s clear that I’m not getting the whole picture from Google’s statistics. 

Google Reader Subscribers

Google Feedburner Subscribers

Does anyone have any better ideas for getting some subscriber stats for RSS feeds on a WordPress blog using Google Analytics? Or maybe from the server logs?