Tag Archives: windows

HP business breakdown and why a PC spin-off could backfire

I had a look at HP’s latest financials, following last night’s triple blast of news from the computer giant. It is ceasing webOS operations, acquiring enterprise knowledge management company Autonomy, and considering (though only considering) a spin-off or other major change to its PC division, the Personal Systems Group. Here is what HP said:

As part of the transformation, HP announced that its board of directors has authorized the exploration of strategic alternatives for the company’s Personal Systems Group. HP will consider a broad range of options that may include, among others, a full or partial separation of PSG from HP through a spin-off or other transaction. (See accompanying press release.)

Looking at the results for the second quarter 2011, here is how the main pieces break down:

$millions

Segment Percentage of revenue Earnings Percentage of total earnings
Services 9,089 28.5% 1225 33.8%
Servers, storage and networking 5396 16.9% 699 19.3%
HP Software 780 2.4% 151 4.2%
Personal systems group 9,592 30.1% 567 15.7%
Imaging and printing 6,087 19.1% 892 24.6%
Financial Services 932 2.9% 88 2.4%

Note that “Earnings” is earnings from operations; HP actually made less money than that, because various other corporate costs have to be deducted. But it gives an idea of where HP’s profit comes from.

So what do these groups do? PSG is notebooks, desktops, workstations and other, where “other” I’d guess will include the webOS mobile devices. In PSG, notebooks accounts for 54% of the total, with desktops taking 38% of the rest. Virtually all of these run Windows.

In servers, storage and networking, 61% is from what HP calls “Industry standard servers”. This is code for Windows server.

Under services, the three big businesses are Infrastructure Technology Outsourcing (42%), Technology Services (30%) and Application Services (19%). The first of these is clear-cut (have HP run your infrastructure), but the second two are both consulting services and on a brief look seem to have some overlap.

Autonomy, by the way, reported revenue of $million 247 in the three months ending June 30 2011 – pretty tiny relative to HP.

A few comments then. It’s worth noting that PSG is the biggest single segment for revenue, but not so for profit, though it is still making a useful contribution.

Imaging and Printing contributes most earnings as a proportion of revenue. I do not know how much of that comes from absurdly overpriced ink cartridges!

If you take PSG together with Industry Standard Servers, you find that around 40% of HP’s revenue comes from boxes running Windows. If you then consider what its printers, network and storage systems attach to, and that a proportion of HP’s consulting business concerns Windows systems and applications, it is obvious that HP’s fortunes are deeply entwined with Microsoft.

If HP removes PSG that will still be true, though less so. But why would HP want do remove PSG? I would guess two main reasons. One is that it is unprofitable relative to the other segments, and the other is that HP foresees the business declining under the force of various well-documented pressures: Apple, mobile, cloud.

It still makes little sense to me. I can understand why HP might want to get out of consumer desktops and laptops, but it seems to me that to supply corporate PCs fits snugly with the rest of HP’s business and has beneficial side-effects. After all, PCs, printers and servers do all plug together both physically and conceptually. Getting rid of PSG might have a negative effect on other parts of HPs business.

In the SMB market, by the way, resellers like HP because unlike Dell it does not mainly sell direct. HP boxes generally work as advertised in my experience, though I rate the laptops less highly than the servers and desktops.

Reports of 19% decline in Western European PC market show structural change

As if we needed telling, a new Gartner report shows a steep decline in the PC market in Western Europe. A “PC” in this context includes Macs but excludes smartphones and what Gartner called “media tablets”, mostly Apple iPads. A few figures comparing shipments in the second quarter 2011 with the same period in 2010:

  • Total PC sales down 18.9%
  • Netbook sales down 53%
  • Desktop PCs down 15.4%
  • Apple up 0.5%
  • Consumer PC market down 27%

What interests me here is not so much the normal ebbing and flowing of the PC market, but structural change indicating a switch away from PCs and laptops to more lightweight mobile devices. I believe this is evidence of that, though the economy is weak and extending the life of existing PCs is an obvious saving both for businesses and consumers.

Still, the dramatic decline in netbook sales suggests that consumers really are buying the more expensive iPad in preference. If you believe that consumers are to some extent ahead of business in their technology choices, then we can expect more of the same in the corporate market too.

No doubt alarm bells have been ringing in Microsoft’s Redmond headquarters for some time. The company is betting on Windows 8 to rescue its operating system from permanent decline, which is why next month’s BUILD conference is so critical. Nevertheless, it will be a year or so before we get new-style tablets running Windows 8, so will it be too late? I tend to think not, just because of the strength of Microsoft in the business world and the importance of Windows for existing applications, but it is interesting to speculate.

One factor which you can argue either way, in terms of Microsoft’s prospects, is that non-iPad tablets seem to be struggling. HP’s TouchPad and RIM’s PlayBook seem to be selling poorly. Google Android looks more hopeful though overshadowed by legal concerns from multiple sources. In Australia and parts of Europe Apple has successfully barred or delayed sales of Samsung’s Galaxy Tab 10.1, though the latest news is that the ban has been lifted outside Germany.

See also: Fumbling tablet computing – Microsoft’s biggest mistake?

Building PasswordSafe for the Mac: Lion development hassles

I am doing some work on a Mac at the moment. On Windows I store passwords in PasswordSafe, an open source utility that works well, so I wondered if I could access my PasswordSafe database from the Mac.

image

I could have run the Windows version in Parallels, which I have just installed, but I figured a Mac version would be more convenient. I didn’t see a Mac build among the downloads, but PasswordSafe is cross-platform, so I downloaded the source to do a quick compile.

I was glad to find README.MAC.DEVELOPERS.txt in the PasswordSafe source and set to work. The first task is to download wxWidgets, a cross-platform GUI library, so I went off to download that. Ran the osx-build-wx script as instructed. Result: error message stating C compiler cannot create executables.

The problem seems to be that PasswordSafe expects GCC 4.0 but the latest Xcode installs GCC 4.2. The solution suggested here is to remove Xcode 4, install Xcode 3, and then reinstate Xcode 4. There are related issues concerning PPC fat binaries and older versions of the Mac SDK.

That solution seemed risky and ardous to me, and I remembered that I still had an old Mac Mini from which I was forced to upgrade in order to install Lion, the latest OS X. I hooked it up, removed Xcode 4, installed Xcode 3, and set to work again.

I get the impression not many people build PasswordSafe for the Mac. The first issue I discovered was that the steps in in the README.MAC.DEVELOPERS.txt don’t mention that after running osx-build-wx you also have to run make in order to build static libraries. That was easy though. The next thing is to load the supplied PasswordSafe project into Xcode and build.

I did that but got an error – the linker could not resolve SizeRestrictedPanel. The fix was to add SizeRestrictedPanel.cpp and SizeRestrictedPanel.h to the project. PasswordSafe then built and seems to work fine, on Lion as well as earlier versions of OS X, though there are a few cosmetic issues. You can see from the image that the caption for the New Database button is slightly awry.

If anyone wants my build, it is here. There is also a Java version, and some people have success with that on the Mac.

Delphi for Windows, Mac and iOS: screenshots and video of cross-platform development

Embarcardero is drip-feeding information about its forthcoming RAD Studio XE2 in an annoying manner; nevertheless the product does look interesting and promises cross-platform native code apps for Windows 64-bit, Windows 32-bit, Mac OS X and Apple iOS. I have grabbed some screens from a video recently posted by Embarcadero’s Andreano Lanusse; the video is also embedded below.

Here is Delphi XE2 showing a FireMonkey application in the designer. FireMonkey is a new cross-platform GUI framework.

image

Note the list of target platforms on the right. If you squint you can see 64-bit Windows, OSX, and 32-bit Windows.

image

How do you compile for the Mac? It is clear from the demo that Lanusse is running in a VMWare virtual machine on a Mac. He also has a Remote Profile option set to target the host Mac:

image

He then refers to a “Platform assistant” which you can see running in a terminal window on the Mac.  He is then able to compile and run from the Windows IDE:

image

Finally, he targets iOS, though this is a separate project, not just another target. The process exports the project to Xcode, Apple’s Mac and iOS IDE:

image

Next, we see the app running on the iPad simulator:

image

The ability to target the Mac is nice to have, but I suspect it is iOS that will attract more interest, given the importance of Apple’s mobile platform.

Here’s the complete video where you can perhaps puzzle out a few more details.

Update: there is also some Q&A in the comments here.

Graphics rendering is Direct2D or Direct3D on Windows, OpenGL on Mac. FireMonkey renders all components through the graphics API, it does not support use native OS components, though Embarcadero’s Michael Swindell says:

FireMonkey client area controls are rendered by OpenGL on Mac, but appear and work just like Cocoa controls – or however you want them to. There are many different Cocoa UI styles in OSX apps, and Firemonkey can render any of them – including iTunes, or Prokit which is an Apple UI style for Pro apps like Final Cut, not available to devs via Cocoa. Windows are Cocoa Windows and the client areas and all user controls are rendered by OpenGL in HD(2D) or 3D. Menus are std and rendered by Cocoa in the menu bar, and common dialogs are rendered by Cocoa. If the “true OSX” look isn’t for you, you’re welcome to use any included Style, download a custom style, or create your own custom style.

Swindell also addresses the matter of Linux and Android:

We do plan Linux and Android. But no eta yet until we get Win/OSX/iOS out. We would also like to provide language bindings for other languages.

Finally, a bit more about that Platform Assistant:

Developer requires a PC and a Mac (or Mac with VM running Windows). You will develop on Windows, and use the platform assistant (PA running on your Mac) to compile natively to your Mac and the PA handles debugging communication between the Mac and your IDE running on Windows. Delphi (or C++Builder) and Firemonkey create compiled stand alone OSX executables that you can sell/distribute to your users. They are native Mac apps. They “copy install” and run like any other Mac app, or you can use a Mac installer if you like.

What’s coming in Delphi RAD Studio XE2: more details of 64-bit and Mac announced, introducing FireMonkey

Embarcadero’s David Intersimone has posted more details of what is coming in the new version of Delphi and RAD Studio XE2, to tie in with an international publicity tour.

One intriguing comment is a reference to FireMonkey:

with FireMonkey, the GPU-powered next-generation application platform, you’ll be able to create visually stunning HD and 3D business applications

Here is the teaser list of features:

  • Create GPU-powered FireMonkey applications
  • Build 64-bit Delphi applications
  • Create a single application and target both Windows and OS X
  • Extend your multi-tier DataSnap applications with new mobile and cloud connectivity in RAD Cloud
  • Connect any visual element to any type of data using LiveBindings™
  • Modernize the look and feel of your Windows applications with VCL styles
  • Create mobile-optimized web applications and standalone apps for iOS and Android devices using RadPHP

Hmm, so RAD Studio XE2 is targeting iOS and Android not with Delphi, but with RadPHP. That suggests some sort of HTML and JavaScript approach rather than a true native executable.

I was not greatly impressed with Delphi for PHP when I first looked at it. That was four years ago though, and since then Embarcadero has acquired Qadram, the third-party developer behind Delphi for PHP, so I would expect something more worthwhile in the forthcoming new version.

Update: Embarcadero’s Andreano Lanusse has posted more details about FireMonkey.

Living in an App Store world: what are the implications?

A few recent events prompt some reflections on the rise of app stores and the implications for developers and for the IT industry.

One is Apple’s OS X Lion release, available only through the Mac App Store; and the removal of the optical drive on the Mac Mini, making it hard to install shrink-wrap software.

Another is Adobe’s closure of its InMarket service and AIR Marketplace app store. Some app stores are doing better than others.

A third is TechCrunch reporting that book apps such as Nook and Kindle are being hobbled or removed from the Apple iOS store. While I cannot verify this at the moment – I still see the Kindle app in the store, and it still has a link to the Kindle web store – it is in tune with Apple’s announcement in February:

… publishers may no longer provide links in their apps (to a web site, for example) which allow the customer to purchase content or subscriptions outside of the app.

Enforcing this on an app such as Kindle promotes Apple’s own iBooks app and store.

There are lots of app stores out there, though one fewer with the forthcoming closure of AIR Marketplace, but how many of them matter? Here is my pick of the top three:

  • Apple iOS and Mac App Store – arguably two different stores, but since you access them with the same account I bracket them together.
  • Google Android Market – not a lock-in like Apple’s store, but still the primary store for Android.
  • Windows vNext marketplace – how this will work is not yet public, but the existence of a new app store in Windows 8 is widely rumoured and might be expected to tie in with what is already in place for Windows Phone 7.

Perhaps I am overstating the importance of the Windows 8 marketplace, given the failure of the Windows Vista marketplace, but given that Apple has now shown the way I find it hard to see how Microsoft can fail with this one.

Note that an app store is not just a marketing ploy. It is a software deployment and update tool.

App Stores score well in terms of usability. Another advantage is that users have a centralised mechanism for software updates, managed by the operating system. That is good for security, because it is unlikely to be disabled, and good for usability as it should mean fewer third-party updaters like those from Adobe, Oracle Java, Symantec and others.

App Stores typically enforce certain conditions on developers. In essence they must be well-behaved. For examples, the Mac App Store prohibits apps that request escalation to root privileges. Apple also rejects apps that use “deprecated or optionally installed technologies”, including specifically Java and by implication Adobe Flash or other runtimes.

This is great for security. In principle, if you decide that you will only install apps from the App Store, you can be confident that all your apps are well-behaved. On the Mac this is interesting; on Windows it would be a revolution.

What are the business implications though?

  • First, it is a significant source of new revenue for the operating system vendor. It gets a cut of everything.
  • Second, it gives tremendous empowerment to user ratings and reviews. On iOS or Android, if you want an app, you automatically search the app store and take note of factors such as user ratings and popularity. Most of us can figure that if there are few ratings or reviews, the app is not popular.

If you are a software company, getting high ratings and good reviews on app stores is now a key challenge, even more so than it is already with the likes of Amazon.

  • Speaking of Amazon, the third point is that app stores will not be welcomed by software resellers. They are simply being bypassed. Amazon is addressing this with its own App Store for Android; but can it really win against the official Google Android Market? Its MP3 store is better value than Apple’s iTunes, but has smaller market share.

Amazon has other business to fall back on, but specialist software resellers will be watching the growth of app stores nervously. Apple resellers in general are already hurting and diversifying, thanks in part to Apple bypassing them with releases like OS X Lion.

The app store revolution is good for users in many ways, especially as prices seem to end up lower than before, but there are worrying aspects. In particular, the ability of the operating system vendor to tilt the store in its own favour is a concern, and we will hear more complaints about that.

Finally, it is interesting to speculate how this may impact enterprise software deployment. Will Microsoft aim to link its forthcoming Windows app store to other deployment mechanisms such as System Center Configuration Manager? What about volume licensing sales, will resellers be able to keep hold of those? Maybe we will learn more of Microsoft’s story on this at the Build conference in September.

Adobe closes AIR Marketplace, InMarket

Adobe is giving up its efforts to support developers deploying to multiple app stores. The idea of InMarket,  announced at the Adobe MAX Conference in October 2010, was to be a one-stop distribution point for developers seeking to target multiple platforms. Adobe handled distribution and billing. The reason given:

After reviewing our efforts and based on feedback from developers, we have decided that we will deliver the most value by helping developers author and publish their apps on multiple platforms. Given this focus, we have decided to discontinue development and support of Adobe InMarket. We are going to continue to provide support for publishing to different app stores through our tooling. The recent Flash Builder 4.5 and Flash Professional CS5.5 provide support for publishing to multiple mobile platforms including Android and Apple iOS devices.

Adobe is not giving developers much time to adjust. The InMarket URL will terminate on August 31. This is causing some consternation:

I don’t understand how you expect publishers will be able to push an update to all the markets they publish to with enough time to get their user base to update before they’re totally screwed. One month? You do realize that even updates pushed to AppUp can take up to 2-weeks for vetting? This is crazy

That said, the low traffic on the InMarket forum is a clue to what Adobe is closing it down.

InMarket only supported Intel AppUp and AIR Marketplace, which rather misses the point of targeting multiple platforms. Had Adobe been able to offer instant deployment to all the key app stores, including Android Market and Apple’s iOS App Store, it would have been more attractive. Given the complexities of the approval process, it is not surprising that this was hard to achieve. A further complication is that Adobe’s AIR runtime is not allowed on iOS. Apps for iOS have to be packaged as native iOS apps.

What about AIR Marketplace?

When we established Adobe AIR Marketplace three years ago, there were few distribution opportunities for AIR developers. There are now several app stores on desktops, mobile devices and tablets that service AIR developers including Apple App Store, Android Market, BlackBerry App World, Intel AppUp center, Samsung Apps, and Toshiba App Place. We encourage you to use these newer popular app stores to distribute your applications.

This of course describes describes exactly the problem that InMarket was meant to address: the challenge of maintaining support for multiple app stores.

AIR Marketplace is still up and running at the time of writing, and seems to have more life than InMarket:

image

That said, why would any potential customer look specifically for AIR applications? It is a runtime and ideally should be invisible to the user. I was interested to see reference to AIR packagers for Windows, Mac and Android in a recent announcement, suggesting to me that Adobe is de-emphasising AIR as a runtime and making it into something more like a cross-platform development tool.

Microsoft financials: Office and server dominate as Windows falters

Microsoft has released its quarterly figures for January-March 2011. My at-a-glance summary is below.

Quarter ending June 30th 2011 vs quarter ending June 30th 2010, $millions

Segment Revenue Change Profit Change
Client (Windows + Live) 4740 -41 2943 -123
Server and Tools 4643 +494 1774 +214
Online 662 +94 -728 -40
Business (Office) 5777 +402 3618 +399
Entertainment and devices 1485 +341 32 +204

Business as usual? More or less, but there are a few points to note.

The figure that jumps out is the stunning performance of Office, which includes SharePoint and Exchange. Why is everyone buying Office 2010, when a document like the one I am typing now could be done just as well in Word 2.0 from 1991, or more plausibly the free OpenOffice?

The answer is the Microsoft has successfully transitioned many of its customers to using Office with SharePoint and Exchange, making it harder to stick with old versions and selling CALs (Client Access Licences) as well as the Office suite itself. This is highly profitable, though the aspect that puzzles me is that Office 365, which is cloud-hosted SharePoint and Exchange, is more cost-effective for the customer since it includes server software, CALs and in some cases the Office client for a commodity-priced subscription.

In other words, I find it hard to see how Microsoft can remain equally profitable if a significant proportion of its customers switch to Office 365. The company may be depending on its ability to upsell those customers to further online services; or perhaps it has not fully thought this through and has set Office 365 pricing at what it needs to be in order to compete with Google.

Fortunately for Microsoft, there is enough doubt concerning the safety of cloud services to sustain continued strong sales of on-premise solutions.

Second notable thing: Windows is in decline. The reason: it is losing market share to Apple and to Google Android. Netbook sales are down 41% according to the release, and I would guess that those sales have mostly gone to Apple iPad and Android tablets rather than to Windows notebooks.

Will Windows 8 reverse the decline? Speculation of course, but it will not repeat the success of Windows 7. In fact, my guess is that Windows 8 will be a hard sell to enterprises which have finally been persuaded to migrate from Windows XP. They are settling down for another five years of stability. Windows 7 was a consolidation release, just the sort of thing enterprises like. Windows 8 will be a revolution release, with most of the interest focused on what it can do in mobile and tablets. If it does succeed, it will do so slowly; there will be no rush to upgrade from 7 other than from the usual early adopters. It may improve sales in the consumer market, but neither Mac nor iPad nor Android is going away.

That leads on to mobile, the figures for which are buried under a pile of Xbox consoles. A good quarter for Xbox, though note how poor the margins are compared to those for Office or Windows.

Finally, the online money drain continues. Note that this is Bing and online advertising, not Azure or Office 365. Microsoft must feel that it the strategic value of these online services is worth the cost, particularly since they tie into mobile and the ecosystem which Nokia is depending on for a reversal of its fortunes. Given that the company has money to burn, there may actually be some sense in that; though for a segment to make such large and consistent losses over a long period has to be a concern.

Nokia results: demonstrating the Osborne effect?

Here’s Wikipedia:

The Osborne effect is a term referring to the unintended consequence of the announcement of a future product ahead of its availability and its impact upon the sales of the current product.

The reference is to Osborne Computer Corporation, a pioneer of early personal computers, which announced the next generation of its range long before it was available. Sales of the current model immediately dived, and the company went bankrupt.

In February this year, Nokia announced that it was abandoning its Linux-based MeeGo smartphone OS, then in development, and that Symbian would be reserved for low-end phones. Its future smartphone strategy will be based on Windows Phone.

Now here come the results:

The challenges we are facing during our strategic transformation manifested in a greater than expected way in Q2 2011.

says the release, which report an 11% decline in sales quarter-on-quarter and an operating loss:

In the period from January to June 2011, net financial expense was EUR 74 million (EUR 141 million). Loss before tax was EUR 141 million (profit before tax EUR 632 million). Loss was EUR 261 million (profit EUR 279 million), based on a loss of EUR 24 million (profit of EUR 576 million) attributable to equity holders of the parent and a loss of EUR 237 million (loss of EUR 297 million) attributable to non-controlling interests. Earnings per share was EUR -0.01 (basic) and EUR -0.01 (diluted), compared with EUR 0.16 (basic) and EUR 0.16 (diluted) in January-June 2010.

Would these results have been better, if Nokia had not bet its business on Microsoft’s mobile OS back in February? My guess is that they would. Nokia in effect announced the obsolescence of all its current Smartphone range. Smart device sales are down 32% year on year.

Still, even the Osborne effect does not account for the decline in its sales of feature phones, down 20% year on year and 25% quarter on quarter. This is what Nokia says:

The year-on-year and sequential declines in our Mobile Phones volumes were driven by distributors and operators purchasing fewer of our mobile phones during the second quarter 2011 as they reduced their inventories of those devices which were slightly above normal levels at the end of the first quarter 2011. In addition, our lack of Dual SIM phones, a growing part of the market, until late in the second quarter 2011 adversely impacted our Mobile Phones volumes during that quarter. Mobile Phones volumes were also adversely affected by continued pressure from a variety of price aggressive competitors.

Despite the grim figures though, it is too early to pronounce the failure of CEO Stephen Elop’s strategy. After all, no Nokia Windows Phones are yet on sale. The company must be hoping to hang on until it has a decent range of Windows Phones, and for Microsoft to grow its mobile market share dramatically above what it is currently.

Should Nokia have chosen Android rather than Windows Phone? Android’s extraordinary growth suggests that it should; yet there are signs of significant copyright and patent trouble for Android, and by attaching to Android Nokia would have been a me-too behind more established vendors such as Samsung, HTC, and nearly everyone else.

Should Nokia have persevered with MeeGo and Symbian? Although early MeeGo devices are winning praise, I doubt that the OS would have challenged Android and iOS; but that is open to speculation.

The problem for Nokia is that if it was going to make a radical platform shift, some degree of Osborne effect was inevitable. Adopting Windows Phone could not have been done in secret.

That said, perhaps the company could have been smarter. Rather than laying all its cards on the table, could it have announced a Windows Phone strategy alongside MeeGo and Symbian, adopting a more gradual approach to avoid shocking the market?

It is also worth noting that Nokia’s problems started long before the arrival of the new CEO. What we are seeing now is the playing out of old mistakes, not just the impact of what may be new ones.

However you spin it though, the new Nokia is a lesser thing than the Nokia of old, which commanded rather than followed the mobile market.

Microsoft partners with Joyent to bring node.js server-side JavaScript to Windows

Microsoft will port node.js to Windows in partnership with Joyent. This will work on Windows Azure as well as other versions of Windows back to Server 2003.

But can you not already run node.js on Windows? This is possible using Cygwin and instructions are here. Cygwin makes Windows more like Linux by providing familiar Linux tools and a Linux API layer. Cygwin is a great tool, though it can be an awkward dependency, but a true Windows port should be higher performance and more robust, particularly as the intention is to use the IOCP API. See here for an explanation of IOCP:

With IOCP, you don’t need to supply a completion function, wait on an event handle to signal, or poll the status of the overlapped operation. Once you create the IOCP and add your overlapped socket handle to the IOCP, you can start the overlapped operation by using any of the I/O APIs mentioned above (except recv, recvfrom, send, or sendto). You will have your worker thread block on GetQueuedCompletionStatus API waiting for an I/O completion packet. When an overlapped I/O completes, an I/O completion packet arrives at the IOCP and GetQueuedCompletionStatus returns.

IOCP is the Windows NT Operating System support for writing a scalable, high throughput server using very simple threading and blocking code on overlapped I/O operations. Thus there can be a significant performance advantage of using overlapped socket I/O with Windows NT IOCPs.

I was impressed by node.js when I saw it presented by author Ryan Dahl at a pre-Dreamforce event last year. Since then it has become better known. This is an interesting move, particularly in the context of an greater focus on JavaScript in the forthcoming version of Windows known as Windows 8. End to end JavaScript for your next-generation real time networking applications?