James O'Neill's Blog

September 28, 2008

Server Core — Too Dry and Crunchy ?

Filed under: Virtualization,Windows Server,Windows Server 2008 — jamesone111 @ 10:52 am

I’ve got a kind of love/hate relationship with the core installation of Server-2008.


I love: That it has the smallest foot print (I remember the days of Netware 2.0a – 3.x – why does a file server need a GUI ? ) , the smallest attack surface, lowest patch requirements (even if some – like an IE update which is needed because web proxy access for activation is provided via an IE component – seem like they shouldn’t be there).


I hate: The obscure command lines needed to set it up from scratch. Feeling brave on our roadshow, I configured core in front of the audience to show the steps. I was very clear that they should have Step-By-step guide by their side if they were going to do a fully manual configuration of core. I tried to be clear about the question of to core not to core




    1. Don’t bother with core on a small scale. One server ? Forget it. In fact if you’re doing a fully manual installation (instead of  using centralized deployment tools like WDS) you should be asking if core is for you.

    2. Core is for lights out operations. Once configured you remove the monitor and keyboard, and don’t ever connect to the console again. Not even via terminal services. Use remote management tools exclusively

    3. Where “management tool X” can only run locally and can’t run on core – and this applies to some tools for setting up storage – then you can’t use the product on core.

    4. Don’t use core for just one function. If your file servers run in a lights out data-centre and your DCs run in lights out data centre, they can both be on Core.

Core is not for everyone. If you have a battery of file servers then the most optimized way to run them is on core, with file serving as the only role. Serving static Web pages ? Best optimization is to run a farm of web servers on core. Consolidating 10 racks of servers down to one rack with Hyper-V ? Core could be for you.  But be clear – just as the optimization of a sports car for racing means saving weight by stripping out the comforts that would make it a Grand Tourer , so Server Core is pared down to optimize its footprint – which means giving up some of the things which make Windows Servers easy. Just as a racing car isn’t optimized for comfort, Server core is not optimized for ease of setup. And if anyone tells you that every file server , or every DC , or every web Server or every hyper-v server MUST be on Core, or even should be on Core, then they are assuming that every use needs the same optimization. I’d question any advice I got from someone who made assumptions like that.


Unfortunately a few Microsoft folks are putting that message out into the world. So I guess we only have ourselves to blame when VMware post a blog item entitled “Hyper-V with Server Core — Too Dry and Crunchy for our Taste” . And they make the points that I’ve just made.  They say Microsoft execs are keen to say Hyper-V is just part of Windows. That’s certainly true, it’s covered by Windows support contracts, and support doesn’t increase in cost when you deploy on 4 processor boxes as it does with VMware. It doesn’t force any hardware choices on you. And it’s managed like any other part of Windows – including using the umbrella technologies of the system center family. It does cut training costs – if you know how to set up a high availability  file-server or exchange server you know how to set up a high availability hyper-v server. Next week at VM Expo I have a session where I’m going to install Hyper-v, install failover clustering, set up a VM and show it failing over in 30 minutes. It is easy. But I never run it on core.


The VMware post talks about the Microsoft execs using the phrase “it’s the Windows you know”  and then go on to say that actually for many people Core isn’t the Windows they know: true. “It’s the Windows you know, or an optimized Windows you might not know yet” isn’t a good sound bite. Of course those customers who will find core a good fit for Hyper-V probably do know it already; but there is no getting away from the fact that when you make your installation decision, core needs more expertise than full Windows. Of course that is a choice you get to make and one I don’t think VMware gives customers; so you’d expect ESX to be easier to set-up than core: something would be very wrong if it were harder.


Of course VMware ducks the issue of what happens after you’ve set up the server. Which is easier to manage once it is running ? Customers who are exclusively on VMware have been showing a lot of interest in System Center Virtual Machine Manager (the most expensive piece of our virtualization family). These are people who don’t see their strategy changing  to Hyper-v in the short term, and have spent a lot buying everything VMware has to offer (See Brett’s story about our VMwareCostsWayTooMuch site for VMworld). Yet they still feel that VMware’s management doesn’t cut it and  like the idea that with SCVMM they can make raise the management of VMware to the level we offer with Hyper-V.  Ironically by protesting that it takes longer to set-up Hyper-v in the worst case scenario, the VMware post emphasizes that the up-front investment needed if Hyper-V is to run in the optimized environment of core isn’t especially great when set against the savings Hyper-V offers.


This post originally appeared on my technet blog.

September 24, 2008

Active Directory User Group, first meeting with John Craddock and Sally Storey

Filed under: Events,Windows 2003 Server,Windows Server,Windows Server 2008 — jamesone111 @ 8:35 am

We now have a UK AD user group (ADUG) And they have their first meeting set for October 23rd at Microsoft’s offices in London, from 6 till 9 in the evening. They’ve managed to secure John and Sally – probably the best established AD double-act in Europe (their tech-ed sessions are always among the top scoring ones). When I first came to work at Microsoft Sally was in Microsoft consulting Services with me and I’ve got a lot time for her and for John. They’re going to focus on new features Server 2008 for this one. Their sessions tend to get full, so although it’s still a month away I wouldn’t waste any time booking if you want to go. I’m hoping the event will also be relayed by Live Meeting, but that has yet to be confirmed.

This post originally appeared on my technet blog.

September 20, 2008

Supercomputers and adverts.

Filed under: Exchange,Windows Server — jamesone111 @ 8:01 pm

One of those Saturday morning serendipitous link chains.

Sometime in 1988  I was talking to Mike, one of my colleagues at Research Machines. It’s impossible to keep using increasing computer for ever I said. He had 16Mhz 386 computer on his desk with a 387 maths co-processor and a special cache board in it and people were cooing over the power of the thing.  Mike said however much power the hardware folks gave us, the software folks would find a way to use it: way beyond his awesome 386 he said. But there has to be a limit, I said "Even if you could give everyone a Cray ?". Even then said Mike….

You might have heard that we’ve been running some new adverts. The first couple were to get people thing "Oh Microsoft have something to say do they"… I greeted these with some skepticism. Bill transitions out of this full time job at Microsoft and we start running adverts with Bill as the star. Even now I can hear someone pitching "Bill" to TV studios , a sitcom about a guy in Seattle, "Think Frasier, but instead of Psychiatrist he’s a computer geek. And the great thing is the real Bill is available. We can bring in guest comedians from all over…"  Reflexively my toes began to curl when I saw Bill in an ad because I learned a little rhyme when I was young.

When the client moans and sighs
make his logo twice the size
If he still should prove refractory
show a picture of his factory
Only in the gravest cases
should you show the clients’ faces

The first ad passed me by. Part of recent Microsoft folklore (if you can call blog posts that) was someone asking Steve Ballmer "When will act like an international company and not like an an American company which does business overseas". I didn’t think people in Britain would get it. It seems a great many Americans didn’t get it either: it’s an advert so what does Microsoft want us to buy ? Vista, office, Exchange, Mobile , the ideas of Live ? Or just open our minds enough to accept that Jerry Seinfeld who is famously from New York and Gates – famously from Seattle – could possibly be in the same shoe shop. The second one did raise a laugh but would a British Audience get it ? Was there something to get ? Or is just the most expensive clearing of the throat in recent corporate history ?

My smoke alarm went off a couple of days ago. It does it every so often – this time was because I left a door slightly open and steam from the shower was enough to trip the thing. Burnt toast or frying some things can also set it off if the kitchen door is open. The thing is a nuisance, so why don’t I just take its battery out ? You know why. I mention this because its actually quite a good metaphor with User Account Control in Windows Vista. Most of the time you forget it’s there. When you’re doing some things it’s a nuisance… But one day it might save you. Granted UAC won’t save your loved ones from death by smoke inhalation. And I’ll take rebuilding an infected computer over even minor fire damage any day.  Trust me on this: I’ve caused a kitchen fire at home which needed two fire engines and men in breathing apparatus.

I mention this because Apple’s ad campaigns got the skin of Microsoft folks so much, that they get into any internal talk about our own ads in first few seconds. Someone brought up this mac ad which is the most inaccurate representation of UAC I’ve seen to date.  Just like the ad agencies take on Viruses – Apple have more published vulnerabilities than Windows but – even virus writers don’t bother with it as a platform. This one seems to be saying "Apple OS-X :  like a house without a smoke alarm".

As a Microsoft shareholder there are times when I think the money we’ve spent with advertising agencies has been wasted – worse than that: I can think of cases where it would have done us less harm to spend the money buying air time for those Apple ads. And for pity’s sake why can I do I see ads like this one for the first time when trying to find something on youtube ? Watch it …. Notice it says 25 years ago a company was founded: that dates it to 2000. And the breaking down barriers bit … well we might see more of that in the "Without walls" campaign. I tell you this because I think the latest ad might turn out to be a master stroke. 

 http://silverlight.services.live.com/invoke/45594/Pride/iframe.html

They’ve taken Apples "PC" character and given him half a dozen words "I’ve been made into a stereotype" and shown a diverse set of people who are PCs. One of the comments I read  – and can’t find now – said Microsoft had missed that the "I’m a Mac" adverts asked "Who would you rather work with."  This ad tells people the PC is people like them, not just the guy in the Apple ad. It makes him look like that guy we’ve all met at some point in our working lives, the one with the ill-deserved superiority complex – the one who thinks the designer logo on the shirt he wears to work matters more than what he does when he’s there, the one who thinks constantly disagreeing with the consensus is proof of his creativity and superiority, when everyone else thinks its a sign of being a jerk. Long term readers might recall I linked to a Grauniad piece  which says the Adverts characterize  Macs as "Smug Preening Tossers". If people turn round and say "you know, the Mac guy isn’t someone I’d want to be around, much less someone I’d want to be " Apple can never run an advert with him in it again.

Whilst using Google’s (youtube’s) bandwidth to show the ad in flash form is all fine and good, the version above is the silverlight version which I got from my colleague Keith Combs. Keith has a couple of interesting links on the front page of his blog right now, he’s got the advance spec for Dell’s new monster laptop: with 3 times the CPU grunt and 4 times the memory of what I have now, in a package which can go in wheel on luggage – is could be the core of event delivery for the next couple of years. Since I’m going to be helping out with an HPC (nee Computer Cluster Server) event in October, his piece on the Cray CX1 caught my eye. With 8 nodes with 8 cores in a single box, and it would have made it onto the top 500 supercomputers in the world in 2004.  Which thinking about the power of those weird and wonderful beasts with the Stonehenge-inspired design that I was thinking of in that chat with Mike. Lordy, you have to love Moore’s law. According Wikipedia the Cray X-MP which was the top dog in 1988 was introduce in 1984 at a cost of $15M (call it £25M in today’s Money). It had 1/5 of the power of each core in my Laptop. And if you’ve got £50K to spend on the CX-1 it will buy you the equivalent of 500 X-MPs for 1/500th of the price (£100 a go, a 250,000 fold improvement. Based the traditional doubling every 18 months Moore would only give about 65,000 times improvement).  Watch Cray’s video which ends with the description "Unprecedented personal computing capabilities".  If a machine like that had a voice it would sound like James Earl Jones and it would be saying "AND I’M A PC"

 

This post originally appeared on my technet blog.

September 18, 2008

Brighton Pavillion Photo Synth

Filed under: Pages — jamesone111 @ 11:53 am

I wanted to create a page with a photosynth embedded in it


 

http://photosynth.net/embed.aspx?cid=c5cb9466-4048-4d89-a2ee-ed46899a400d&i=0:0:176&z=1314.4765759999995&g=14&p=0:0&m=false&c=-7.67073:-5.47273:0.349517&d=3.20103:-2.07398:-2.44516

So here is one.

This post originally appeared on my technet blog.

September 16, 2008

Need a product code name ? (Dunnington)

Filed under: General musings — jamesone111 @ 9:40 am

In past posts I’ve talked about a mythical Microsoft product named "Basingstoke". Microsoft’s legal folks recommend that if you have to come up with a codename for something Geographical is good – they’re fairly safe from a trademark point of view, and I recall in the mists of time one product team which used the names of places with Whiskey distilleries for the code names, but someone who wanted to have "Project Red-Bull" and who was told in no uncertain terms why he had to think again.

Basingstoke isn’t the kind of place which would have a product named after it; I’ve vowed that if I ever get to Redmond I’ll honour my family’s Humberside heritage and get a product named "Scunthorpe".

I was reading something internal with someone saying "Intel have announce the Dunnington processor – when will we clarify our support for it ?" (answer "Be patient, grasshopper" ) I tapped "Dunnington into IE’s search box- which I have defaulting to Live search – what I got was some news articles about the chip at the top of the page,  (good) but underneath I get Dunnington through the Ages, Driffield Online – The Digital Community for the Yorkshire Wolds. — what ? I still have Google as a backup search engine, and Google.com looks at my IP address and routes me to UK, no news and at top of the list is Dunnington Cricket Club followed by Dunnington through the ages again. [Update – news on the chip is now in 5th place]. It may cause the club secretary of Dunnington CC some head scratching when he gets the bill from his web site hoster and finds that a huge number of hits are coming from people clicking the "I feel lucky" button on Google.co.uk (on .com it takes you to a genealogy site).  

Dunnington is a village on the Yorkshire side of the river Humber not twenty miles from where my Grandfather was born. But oh the rich seam product code names on that Yorkshire Wolds page: whilst Scunthorpe has names for minor releases (Wrawby, Scawby,and Scotter  jump off the map) on the other side of the river the wolds can offer (among others) Barmston, Dringhoe, Emmotland , Fimber, Gembling,  Helperthorpe, Hempholme , Kiplingcotes, Lissett, Nafferton , Octon , Rotsea, Skipsea ,Thixendale, Ulrome and Wetwang  The last of these I have seen in Douglas Adams and John Lloyd’s The Meaning of Liff a book that set out to get "these words down off the signposts  [and] earning their keep". 

There’s got to be a whole new possibility for exploiting search engines here. If you are a pretty village which wants to attract tourists, get Intel , Microsoft and the like to use your village name as the code name for a new product. 

This post originally appeared on my technet blog.

September 14, 2008

Motor racing again.

Filed under: Uncategorized — jamesone111 @ 2:39 pm

I hope to get back to the Hyper-V , Windows and the rest shortly.

I read something Jeremy Clarkson wrote "It makes me so angry my teeth itch". That was how I felt after the what went on in Belgium with corrupt officials ruining things. Well today what happened in Italy has served as a reminder for why I feel passionate about formula one.  Doesn’t matter if your a fan of any team, or none , Just fantastic. (since I’m posting while the drivers are still on the podium I won’t spoil it for anyone who is watching later). Shame I broke my TV stick and its replacement is still in the post.

This post originally appeared on my technet blog.

September 10, 2008

Hyper-v and competitors /collaborators

When Ray Noorda ran Novell he coined the term Coopertition to describe their relationship with us. Microsoft’s Kevin Turner described this as "Shake hands but keep the other hand on your wallet".

We would love customers to buy ONLY Microsoft stuff (support would be SO much simpler), and competitors would love customers only to buy their stuff. A world where we go 100% of spending of x% of the customers would be so much neater than the real world where we get x% of the spending  (on average) of 100% of customers. Both we and competitors want customers to have a great experience of our respective technologies, and that doesn’t happen if we don’t cooperate on some things.

In the virtualization world it means two things ; being able to run competing OSes on our virtualization and being able to run our OSes on competing virtualization and give the customer clarity about support.

So first, if go to http://windowsservercatalog.com/ and click on the ‘certified servers’ link on the right side of the page under the Windows Server 2008 logo, you can check which Servers have been validated in the lab- there a sub-section for servers validated for hyper-v.

Second if you go to Server Virtualization Validation Program page and click on the  ‘products’  link on the left side of the page you can find out which products we support. As you can see VMware is on the list their entry says which version of Windows is supported on which version of VMware. Today it’s 32bit Windows 2008 only, on ESX 3.5 update 2 only.  That would tend to make people nervous about older versions of Windows until you the section which appears next to each catalogue entry "Products that have passed the SVVP requirements for Windows Server 2008 are considered supported on Windows 2000 Server SP4 and Windows Server 2003 SP2 and later.". It would be reasonable expect more products from more vendors to appear on the list, but it’s good to see that VMware was one of the first to pass the tests.

Third. Linux support. Mike Sterling has posted that Linux Integration Components are now posted , the actual link he provides to the connect web-site seems to be broken, but you can find the components in the Connections Directory

Steve and I are off to Edinburgh to do the 5th run of our Virtualization tech-ed event Seats are still available for tomorrow (Thursday 11th)

This post originally appeared on my technet blog.

September 8, 2008

Net maps – PowerShell and Excel

Filed under: Office,Powershell — jamesone111 @ 2:46 pm

Over the weekend I picked up a link from Jeffrey, which took me to this post by Doug. My post from Friday had already got me thinking about visualization stuff and so I went off on the link that Doug provided. A nice little snippet of PowerShell which piqued my interest and a link to another post of his which explains Microsoft Research released .NetMap – linking to the download on codeplex. “I’ve got to try this” I thought. So after a couple of hiccups with the installation I had it installed and was playing with the Excel test harness they built for it. To get an idea of what it can do it imports information about mails you’ve had with who links to who, and I had fun hour playing with that last night, Here’s the result. I’ve blurred the names except my own, but the items in centre are distribution lists I belong to. The huge star burst on the bottom left is one of the main activists on the Virtualization discussion lists: I’ll dream up some other uses for this in due course, but I’d love to see what other people come up with


Update I wrote a little bit of PowerShell to map out links between WMI objects

Function get-related
 {param($class, $maxlevels)
  if ($global:done -notContains $class) {$global:done +=  $class
      $classes=(([wmiclass]”\\.\root\cimv2:$class”).getRelatedClasses() | where {$_.name -like “Win32*”})
   $classes | select-object -property @{Name=”Source”; expression={$class}}, @{Name=”Target”;expression={$_.name}}
   if ($maxLevels -gt 0)   {$classes | foreach-object {get-related $_.name ($maxLevels -1 )}  }
 }}

PS > [string[]]$done=@()
PS > get-related “Win32_ComputerSystem” 2 | show-netmap

The result is attached below … the picture


Click for full size view

This post originally appeared on my technet blog.

Today’s "Get virtual now event" – official information published

Filed under: Events,Virtualization,Webcasts,Windows Server,Windows Server 2008 — jamesone111 @ 11:38 am

The link to this story is now on the press pass home page 

Today more than 20 partners announce plans for Microsoft virtualization solutions; Microsoft Hyper-V Server, Microsoft Application Virtualization 4.5 and System Center Virtual Machine Manager 2008 available within 30 days.

BELLEVUE, Wash. — Sept. 8, 2008 — Kicking off a global series of “Get Virtual Now” events that will reach more than 250,000 customers and partners, Microsoft Corp. today announced strong customer and partner adoption of Microsoft virtualization software, new tools and programs to drive partner success.

Microsoft also announced the upcoming availability of System Center Virtual Machine Manager 2008, Microsoft Application Virtualization 4.5 and the new Microsoft Hyper-V Server 2008 as a no-cost download during today’s keynote addresses streamed live at http://www.microsoft.com/presspass/events/virtualization/default.mspx.

The Hyper-V Server website is also live. And as I forecast over the weekend Patrick’s posted a summary 

This post originally appeared on my technet blog.

September 7, 2008

Monday’s "Get Virtual Now" event

Filed under: Events,Virtualization,Windows Server,Windows Server 2008 — jamesone111 @ 5:31 pm

You may have seen that there is a big Microsoft Virtualization event happening in the US on Monday.  I’m expecting it to be reported on the Virtualization team blog, and on the Sever and Tools Business "news bytes" blog, and possibly on the "Software enabled Earth"blog as well. 

I’ve had a short briefing ahead of the event, you can see from the published agenda that some of the corporate big-guns will be there,and I’m not going to pass any comment until after they have spoken. Our Press Pass site has a page for the event, including a live stream of it, so if like me you can’t make it to the event in person and want to keep up with what they say – well that seems like the place to be.

This post originally appeared on my technet blog.

September 6, 2008

Lies, Damn lies, and Statistics – a lesson in Benchmarking.

Filed under: Virtualization,Windows Server,Windows Server 2008 — jamesone111 @ 6:34 pm

Many years ago – before on-line meant "the internet" – I annoyed a journalist in an on-line discussion. I criticized the methodology used by his magazine to test file servers: Machines copied a large file making it a test of the cache effectiveness of the server.  As more machines and hence more files were added performance rose to a peak, then the total of files being coped exceeded the cache size, and it plummeted. This they explained as "ethernet collisions".

I mention this because there’s always a temptation to try to rip up a Benchmark someone else has done (I certainly didn’t use very diplomatic language back then). Single task tests can give you an idea how well a machine will carry out a similar task. What do file server tasks look like ? Realistic tests for file servers are hard. For virtualization it is close to impossible. If you a take a real world question like "I have 100 machines; when I multiply their CPU speed by their average CPU loading it they average out at 200MHz. How many Servers do I need ?" Obviously it’s more than 100×200 = 20GHz / (cores * Clockspeed) … but how much more ?" You need to answer questions like "What’s the overhead of running VMs ?" Would 5 servers running 20VMs have a bigger or smaller percentage overhead than 2 servers running 50 ? Assuming you could work this out and come up with an "available CPU" figure, it doesn’t answer questions about peaks of load i.e. "at any point in the last month would the instantaneous sum of CPU load totaled over a set of machines exceed the available CPU on the virtualization server ? And of course we haven’t mentioned disk and network I/O questions.

If that wasn’t enough to make people want to give up on Benchmarking,  Virtualization is a technology to put a lot of small loads into a single server. Benchmarks tend to load a system to the maximum and measure the throughput. Running benchmarks on virtualization is a bit like putting trucks on a race track. It might be fun and you will get a winner … but it doesn’t tell us anything much about the real world.  Still that doesn’t stop people doing it.  And just with motor racing even when the winner is declared people will argue afterwards.

Below are some numbers from Network world who benchmarked VMware and Hyper-V. They did the test on the two virtualization platforms, using Windows and Suse Linux as the guest OS. Using 4 procs and one proc per machine, they tested 1,3 and 6 instances and  I’m only showing the Windows-as-guest and I’ve multiplied the score per VM  by the number of VMs (they just showed the SPECjbb2005 bop/sec scores per VM, but you can get the raw numbers from pages 3 and 4 of their results (link below)

  One Instance Three Instances Six Instances
OS running natively on one CPU 18,153 n/a n/a
OS Running natively across four CPUs. 32,525 n/a n/a
Uni-proc VM on Hyper-V 17,403 49,089 87,186
Uni-proc VM on VMware ESX 17,963 53,205 83,784
4-proc VM on Hyper-V 31,037 101,022 87,528
4-proc VM on VMware ESX 31,155 81,429 96,816

Lets do some quick analysis of these numbers.

18,153 Vs 32,525 On the bare OS 1 Proc gives 18K bop/s and 4 procs gives 32K. So the application is not CPU bound – presumably I/O bound, and sensitive to disk configuration; however the "How we did it" page doesn’t say how the disks on each system were configured – for example did they use the OS defaults (slow, expanding  dynamic disks in Hyper-V) or do any optimization (faster, but more space consuming Fixed disks in Hyper-V). Given that other benchmarks show that Hyper-V can get better than 90% of the disk IOs of the native OS even under huge loading  – the separate disk benchmark they produced showing a very low number makes me suspect they followed the default. But the basic rule of lab work is record what it was you measured(a fixed, expanding or passthrough disk) otherwise your result lose their meaning. 

101,022 Another odd thing on disks is that the 3x Quad-Proc VM test, Hyper-V’s scores show each of the VMs getting 3.5% more throughput than the bare OS. Hyper-V does not cache VHD files, although it can coalesce disk operations which can occasionally throw tests, personally I don’t trust tests where this affects the result (which it plainly does in this case). 

87,186 Vs 83,784. The system had 16 cores and with uni-proc tests they stopped adding VMs at 6. With a single Uni-proc VM, VMware is a little faster, 3 such VMs and it’s faster still. Get to 6 and hyper-V is winning. Does efficiency favour Hyper-V as load increases who can tell ? Without tests at 12 or more VMs there’s no way of telling if this is an aberration or a trend.

Six instances. On uni-proc VMware and Hyper-V both get nearly 200% more throughput going from 1 to 3 VMs – that’s almost linear scalability, but it tails off to 57% and 78% going from 3 to 6. On 4 proc hyper performance actually goes down at 6 instances, and VMware only goes up 20% – this makes me think the system is I/O bound.

So who is the winner ?

In the uni proc test Hyper-V’s best score was 87,186 and VMware’s was 83,784. So Hyper-V is the winner !

In the quad-proc test Hyper-V’s best score was 101,022 and VMware’s was 96,816. Hyper-V is the Winner again !

Now, I’ve already suggested this is a field where there are more questions than answers. I’m left with just one: Since Microsoft clearly won, why is their article called VMware edges out Microsoft in virtualization performance test ?

Bonus link. This is NOT Simpson’s paradox. but if you’re every stuck for something to talk to statisticians about it’s worth knowing.

This post originally appeared on my technet blog.

September 4, 2008

It’s not about computation, it’s connection and visualization

Filed under: General musings,Photography,Working at Microsoft — jamesone111 @ 1:10 pm

Yesterday Microsoft UK had its company conference in Brighton, and I always have mixed feelings about these affairs. We transported 1000 people 100 miles and told and with no sense of irony whatsoever we told them about what the company was doing about the environment, showed them a video of the office they had led behind and showed them clips of Microsoft people telling them what it was like to work at Microsoft. I’ve been involved in enough events of my own to feel some sympathy for the organizers. What is informative to one person might seem insultingly obvious to someone else. What one person finds riveting makes someone else want to gnaw their own leg off. There are occasions where you need to gather the whole company together and yet when you make attendance mandatory, people think "They wouldn’t need to make it mandatory if people wanted to go".

But at most of these events there is a segment where you turn to a colleague afterwards and say "that was worth coming for". For a lot of people where I was sitting the speaker who had that effect was an external one, Sir Mark Grundy. It’s an odd thing, still, to see a head teacher get a knighthood: Grundy got his for turning round a pair of schools in the Midlands; but I suspect it was more than just having a good plan and executing on it. The guy believes in something: "I’ve never met a pupil who came to school to fail" was one thing he came out with – he’s very articulate and his passion comes through when he speaks. But for us the big thing was the difference technology made to his business: he talked about teenagers and threw out the question "why is life at home so connected (messenger, Email, even Xbox live) but when kids come to school we expect them to be happy to be told to open a book and copy out a diagram", so his student portal is a way of linking students together (and shutting students out of it has become one of his best disciplinary tools) . Since he gets how important it is to have parents on side – he’s got a portal to show them information too – from PerformancePoint among other sources. And other heads ask him why the providers who deliver their services can’t do that. But this idea of taking a lot of … stuff and using the power of software to make sense of it carried on he had a love of Deep-zoom (and got a good laugh when he said he liked the old name, Sea-Dragon "Did someone on work experience choose the new name ?"), and Photosynth. And another Deep zoom-like product (I’m not sure the base IS deep zoom) and that’s PPTPlex (watch the videos there to get some idea). It was the first time I’d seen someone use it on stage, and it was enough of a spur to get me to install it when I got home, and if you’re attending any of my events in the near future I may experiment with it.
I’d planned to do a photosynth of the area round the royal pavilion so I whizzed off at lunch time and shot that. I have the details that I can find in there. Which came back to one of Sir Mark’s points : this is the kind of technology kids should be getting exposed to and if not, ask "why not ?"

Indeed.

 

Technorati Tags: ,,

This post originally appeared on my technet blog.

IE 8 testing – part 2

Filed under: Beta Products,Internet Explorer,Working at Microsoft — jamesone111 @ 11:15 am

image One of the things about a new browser is that various web sites expect particular ID strings to identify browsers they know and get awkward if they don’t recognize the browser ID. So one of the nice things in IE 8 beta 2 is that you can select compatibility mode for particular sites and send the string which says "I am IE 7."

I’ve mentioned a few times that Microsoft has quite a sophisticated set of benefits; employees get to check their payslip on line, and wouldn’t you know it… the payroll system is outsourced and is one of those sites which keels over if confronted with a new browser (actually that makes sense – you don’t want to assume a previously unknown browser will render all your columns correctly when it’s showing people what’s been added to or removed from parts of their pay).

So a quick tip-of-the-hat is due to our finance people.Within a couple of days of beta 2 going live they warned there was a problem. And in some companies that would be it. Want to see your payslip, don’t run beta software. Not here: they went and tested it and came back with the conclusion that, yes IE8 works just fine in IE7 mode, and with instructions on how to configure it. And after my notes on people’s bad mail habits they even sent it out with an opening sentence which lets you triage it correctly: bin it (I don’t use on-line payroll or IE 8) , Act on it now (I’ve just installed IE8 and I use the on-line payroll) or file it for the future (I use on-line payroll, and I’m going to try IE8 sometime)

I said before that letting users put the latest versions of software on their machines before IT have tested everything is the Microsoft way, but it wont’ work for many companies. Notice also that by empowering people in this way, finance deal with issues involving a finance application; I’ve been to companies where this would get stuck between finance and IT for weeks. Getting the testing done up front, involving internal customers (pilot users) and recording the problems and fixes, IS something that everyone can and should do, starting with the sites where you either don’t control the client (that’s outward facing sites where your customers go) or the server (that’s partner/extranet sites where your people go).  This is also one of those cases where using sharepoint to create a Wiki works, because the same people -pilot users -  have both questions and have answers to share. (See Raymond for why Wikis and the like fail )

 

.

This post originally appeared on my technet blog.

Blog at WordPress.com.